Rhythmic Diurnal Synthesis and Signaling of Retinoic Acid in the Rat Pineal Gland and Its Action to Rapidly Downregulate ERK Phosphorylation

Vitamin A is important for the circadian timing system; deficiency disrupts daily rhythms in activity and clock gene expression, and reduces the nocturnal peak in melatonin in the pineal gland. However, it is currently unknown how these effects are mediated. Vitamin A primarily acts via the active metabolite, retinoic acid (RA), a transcriptional regulator with emerging non-genomic activities. We investigated whether RA is subject to diurnal variation in synthesis and signaling in the rat pineal gland. Its involvement in two key molecular rhythms in this gland was also examined: kinase activation and induction of Aanat, which encodes the rhythm-generating melatonin synthetic enzyme. We found diurnal changes in expression of several genes required for RA signaling, including a RA receptor and synthetic enzymes. The RA-responsive gene Cyp26a1 was found to change between day and night, suggesting diurnal changes in RA activity. This corresponded to changes in RA synthesis, suggesting rhythmic production of RA. Long-term RA treatment in vitro upregulated Aanat transcription, while short-term treatment had no effect. RA was also found to rapidly downregulate extracellular signal-regulated kinase (ERK) 1/2 phosphorylation, suggesting a rapid non-genomic action which may be involved in driving the molecular rhythm in ERK1/2 activation in this gland. These results demonstrate that there are diurnal changes in RA synthesis and activity in the rat pineal gland which are partially under circadian control. These may be key to the effects of vitamin A on circadian rhythms, therefore providing insight into the molecular link between this nutrient and the circadian system.


Introduction
Vitamin A is an essential dietary component that is delivered to tissues in the form of retinol. It primarily acts through two active metabolites, synthesized in a two-step oxidative pathway. The first, retinaldehyde, is used in the visual cycle for the functioning of rhodopsin in phototransduction. Retinoic acid (RA), the metabolite of retinaldehyde, is a potent transcriptional regulator in the central nervous system (CNS) through activation of ligandgated transcription factors, retinoic acid receptors (RARs) [1]. In addition to its canonical genomic activities, RA also has nongenomic effects including regulation of protein translation [2][3][4] and kinase phosphorylation; studies have demonstrated that RA modulates activation of a number of kinases including extracellular signal-regulated kinase (ERK) 1/2 and Akt [5][6][7].
The role of RA in embryonic development is well established [8] and it remains active in the adult CNS where it has roles in neurogenesis [9,10] and synaptic plasticity [11,12]. There is increasing evidence emerging for a new role of RA in the regulation of circadian rhythms (reviewed in Ransom et al. [13]). Vitamin A deficiency (VAD) has been found to disrupt oscillations in clock gene expression [14,15] and daily rhythms of locomotor activity, a direct output of the central circadian clock [16]. Currently, it is not clear how these effects are mediated, but previous studies have alluded to a role for RA. The three isomers of RA, all-trans-RA, 9-cis RA and 13-cis RA, were identified as potential circadian entrainment factors in a screen of 299 peptides and bioactive lipids, due to their ability to entrain rhythmic expression of the clock gene, Per2 [17]. RA also influences other components of the circadian clock, including inhibition of CLOCK:BMAL [18,19], and upregulation of Per1 [19]. Furthermore, the RAR-related orphan receptor (ROR) β, for which RA is a high-affinity ligand [20], is highly expressed in the circadian system [21,22], and its deletion leads to an extended free-running period in mice [23,24]. In addition, studies suggest vitamin A is necessary for pineal gland function, whose primary purpose is to regulate physiological rhythms by relaying the signal of circadian time to the body via melatonin secretion. The suprachiasmatic nucleus (SCN), the site of the master clock, drives the nocturnal increase in melatonin production by inducing activation of arylalkylamine N-acetyltransferase (AANAT) at night, the penultimate melatonin synthetic enzyme. In a number of vertebrates, including rodents, the AANAT rhythm is predominantly under transcriptional control with an increase in Aanat mRNA of more than a 150-fold at night in rats [25]. Studies have found that VAD in rats leads to a significant reduction in peak nocturnal AANAT activity and melatonin levels, as well as the disappearance of the daily rhythm in mitogenactivated protein kinase (MAPK) activation [26,27]. However, it is currently unknown how these effects are mediated.
Retinol has roles in the pineal gland of non-mammalian vertebrates; in reptiles it is required for photoreceptor integrity [28] and in birds it is involved in phototransduction [29,30]. However, its role in the mammalian pineal gland has scarcely been investigated, despite it containing high levels of retinol and retinol binding proteins [31][32][33]. Retinaldehyde is undetectable in the mammalian pineal gland indicating that it is not likely to serve a phototransduction role [32,34]. Whereas genes associated with RA signaling have been reported to be highly expressed in the rodent pineal gland [35], suggesting that RA is active in the mammalian pineal gland. These include genes encoding the retinoid X receptor (RXR) γ, which binds to the 9-cis isomer of RA [36], and RORβ, which is rhythmically expressed in the rat pineal gland with a peak during the night [22,37].
In order for RA to be a regulator of circadian rhythms, it should itself be subject to diurnal variations in activity. In this study, focussing on the pineal gland component of the circadian system, we determined that RA synthesis and signaling do indeed exhibit diurnal changes. The effect of RA on Aanat transcription and kinase activation, which comprise two key molecular rhythms in this gland, was also investigated. Several of the genes involved in RA signaling displayed diurnal changes in expression and production of RA by the pineal gland was found to increase during the night, suggesting there is a diurnal rhythm in RA synthesis and signaling in this gland. Short-term treatment of RA did not influence Aanat transcription in cultured pineal glands indicative that RA would not control Aanat rhythmicity, although long-term treatment induced a twofold upregulation in expression, suggesting that RA may influence the amplitude of AANAT's rhythmical expression. In addition, RA induced a rapid decrease in ERK1/2 phosphorylation, demonstrating that it has the capacity to influence kinase activity and therefore molecular rhythms in the rat pineal gland. ERK1/2 signaling is subject to diurnal oscillations in activity in the pineal gland with an increase in phosphorylated levels during the night [26,38], which likely enables its role in the regulation of the circadian pacemaker in this gland [39,40]. In addition, it is also thought to be a key mediator of photoentrainment in the SCN [41][42][43]. Therefore, the phosphorylated level of ERK1/2 has an important role in biological timekeeping, and the present study supports the growing evidence that RA should also be considered as an important component of the circadian system.

Rats
Male Sprague Dawley (SD) rats at 6-7 weeks of age were used for all experiments unless otherwise stated, maintained at 20-24°C with unrestricted access to rodent chow and water. They were housed under a 12 h light:12 h dark (LD) cycle (lights on 07:00-19:00), or in constant darkness for the last 2 days (DD). Animals were sacrificed by rising CO 2 and cervical dislocation unless otherwise stated. For polymerase chain reaction (PCR), quantitative PCR (qPCR) and western blotting, pineal glands were dissected and rapidly frozen on dry ice. For PCR analysis, adult male SD rats were sacrificed at zeitgeber time (ZT) 2, where ZT0 corresponds to 07:00. For western blotting analysis, adult male or female SD rats were sacrificed at ZT2 or 7. For immunohistochemistry, rats were sacrificed by intraperitoneal (IP) injection of pentobarbital at ZT5 and transcardially perfused with 4% paraformaldehyde in phosphate buffer, the pineal glands removed and immersed in the same fixative overnight at 4°C. To determine whether RA signaling genes exhibit diurnal changes in expression, pineal glands were collected at 6 h intervals over 24 h (h), ZT0, 6, 12 and 18, and gene expression determined by qPCR. At ZT12 and 18, which occurred during the dark period, dissections were performed under dim red light. To determine whether changes in RA signaling genes are driven by the light/ dark cycle, rats were housed in constant darkness for 2 days and pineal glands were collected under dim red light at 6 h intervals over 24 h (circadian times (CT) 0, 6, 12 and 18; CT0 and CT12 correspond to subjective light on and light off, respectively). All animal procedures were carried out in accordance with Home Office regulations and local ethics committee guidelines.

Immunohistochemistry
Perfusion-fixed pineal glands were processed into paraffin wax blocks, sectioned at 7 μm onto polysine-coated slides and dried overnight at 37°C. Sections were dewaxed in Histo-Clear (National Diagnostics) and rehydrated in decreasing ethanol concentrations. Antigen retrieval was performed by boiling the sections for 10 min in sodium citrate buffer, pH 6. Sections were labeled with antibodies against RARα, RALDH1, S-antigen (SAG) and glial fibrillary acidic protein (GFAP; Table 2), and appropriate fluorescent secondary antibodies. The slides were mounted using mounting medium containing bisbenzimide

Quantitative PCR
Total RNA was extracted from individual pineal glands using an RNeasy mini kit (Qiagen) with on-column DNase digestion (Qiagen). RNA was quantified using a NanoDrop spectrophotometer (Thermo Scientific) and precipitated in 100% ethanol, linear acrylamide and ammonium acetate. cDNA was synthesized from 500 ng total RNA using a High Capacity RNA-to-cDNA kit (Applied Biosystems). qPCR analysis was performed using SensiMix SYBR mastermix (Bioline) using primers designed for Aanat, Bmal1, Rarg, Raldh3, Rdh10, Rdh12, Cyp26a1, Cyp26b1 and Gapdh (Table 1a) or using PrecisionPLUS MasterMix (PrimerDesign) and probes designed by PrimerDesign Ltd for Raldh1, Raldh2, Rara, Rarb and Gapdh (Table 1b). Samples were run on a LightCycler 480 (Roche) and data analyzed using LightCycler 480 Software 1.5. Target gene expression was normalized to Gapdh expression. The specificity of all primer sets towards the intended target was confirmed by sequencing of the PCR products by GATC Biotech, followed by checking the sequences against the rat RefSeq RNA database on NCBI BLAST.

Pineal Gland Culture
Pineal glands were obtained for tissue culture at ZT 4-5 and were cultured based on the method used by Bailey et al. [35].
They were rapidly dissected and placed immediately into icecold culture medium consisting of BGJb medium (Fitton-Jackson modification; Gibco) containing 0.1% BSA, 25 mM HEPES buffer (Sigma-Aldrich), 2 mM GlutaMAX supplement (Gibco), 0.1 mg/ml ascorbic acid (Stem Cell Technologies) and 100 U/ml penicillin-streptomycin (Gibco). Meninges were removed under a dissection microscope and pineal glands were transferred onto Millicell culture plate inserts (Millipore) in a 6-well plate, one pineal gland per well. Pineal glands were incubated in 1 ml culture medium at 37°C, 5% CO 2 ; media was changed after 24 and 48 h. The pineal glands were treated after 48 h in culture. For gene expression experiments, treatments consisted of all-trans-RA (1 μM; Sigma-Aldrich), norepinephrine (NE; 1 μM; Sigma-Aldrich), or vehicle control for 4 h. For kinase activation experiments, treatments were all-trans-RA (1 μM), epidermal growth factor (EGF; 0.6 μg/ml; R & D Systems), or vehicle control for 10 min. RA and NE were dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich), therefore control and EGF treatments received an equivalent concentration of DMSO (0.01%). Following treatment, pineal glands on the membrane inserts were rapidly frozen on dry ice and stored at − 70°C until qPCR or western blotting analysis. For the culture of pineal glands obtained from postnatal day (P) 10-12 rats, male SD rat pups were euthanized by IP injection of pentobarbital at ZT7 and the pineal glands were rapidly dissected and placed immediately into ice-cold culture medium consisting of 50% minimum essential medium (MEM; Sigma-Aldrich), 25% heat-inactivated horse serum (Sigma-Aldrich), 25% Hanks' buffered salt solution (HBSS; Gibco), supplemented with 100 U/ml penicillin-streptomycin, 2 mM GlutaMAX supplement, 5 mg/ml additional glucose and buffered with 25 mM HEPES. Meninges were removed under a dissection microscope and pineal glands were transferred onto Millicell culture plate inserts in a 6-well plate, 2 to 3 pineal glands per well. The pineal glands were incubated in 1 ml

Retinoic Acid Measurement
A RA reporter cell line (Sil-15) was used to measure RA production in rat pineal glands [44,45], as the lipid nature of RA and nanomolar concentrations endogenously present in tissues make this compound difficult to detect by other means. The cell line is derived from F9 teratocarcinoma cells transfected with a transgene consisting of LacZ driven by multiple tandem copies of the RA response element (RARE) from the mouse Rarb promoter [46]. Sil-15 cells were grown in Dulbecco's Modified Eagle Medium (Gibco) containing 10% fetal calf serum (Gibco) and 0.8 mg/ml G418 (Sigma-Aldrich), at 37°C, 5% CO 2 . Cells were plated onto gelatin-coated 96-well plates and allowed to reach confluence before use. Pineal glands were collected from 6week-old male SD rats at ZT0 and 12, sliced into three parts and transferred to 96-well plates (one pineal gland per well) in culture medium consisting of 50% MEM, 25% heat-inactivated horse serum, 25% HBSS, supplemented with 1 μM retinol (Sigma-Aldrich) to match circulating levels of retinol, 100 U/ ml penicillin-streptomycin, 2 mM GlutaMAX supplement, 5 mg/ml additional glucose, and buffered with 25 mM HEPES. Pineal glands were incubated at 35°C, 5% CO 2 for 2 h, protected from light. The pineal gland-conditioned media and blank medium were diluted 1:3 in the culture medium before adding to the plated Sil-15 cells in triplicate under dim yellow light. A serial dilution of all-trans-RAwas prepared and added to Sil-15 cells for generation of a standard curve. Sil-15 cells were incubated at 37°C, 5% CO 2 for 22 h. Cells were washed in PBS, fixed with 1% glutaraldehyde (Sigma-Aldrich) for 15 min, followed by additional PBS washes. LacZ expression was detected using X-gal solution (2 mg/ml 5-bromo-4-chloro-3indolyl-β-D-galactopyranoside (X-gal; Promega), 3.3 mM potassium ferrocyanide, 3.3 mM potassium ferricyanide, 1 mM magnesium chloride) incubated at 37°C, 5% CO 2 for 24 h. The absorbance was measured on an Emax precision microplate reader (Molecular Devices). Mean absorbance readings were corrected for the mean blank medium absorbance and RA concentration interpolated from the cubic spline-fitted standard curve generated using GraphPad Prism version 5.04.

Statistics
Data were analyzed by unpaired Student's t test or one-way ANOVA with Tukey's multiple comparison test as appropriate. In cases where data were not normally distributed, Mann-Whitney U or Kruskal-Wallis tests were performed as appropriate. Non-parametric data is plotted as median values with interquartile range.

The Components Required for RA Signaling Are Present in the Rat Pineal Gland
For RA signaling to occur in a given tissue, the enzymes that synthesize RA from retinol and RARs must be present. An initial PCR screen was performed on the adult rat pineal gland to determine the expression of the genes encoding the required components of the RA signaling pathway. Transcripts encoding the retinol dehydrogenase (RDH) and retinaldehyde dehydrogenase (RALDH) enzymes that catalyze the first and second steps, respectively, of the conversion of retinol to RA were detected by PCR (Fig. 1a). Transcript encoding another member of the RDH family, Rdh12, was also detected, which primarily acts as a retinaldehyde reductase enzyme, converting retinaldehyde to retinol [47]. Genes encoding each of the three subtypes of RARs, Rara, Rarb and Rarg, were also found to be expressed. Transcripts encoding Stra6, the retinol transporter, and Cyp26a1 and Cyp26b1, encoding the RA catabolic enzymes, were also detected. Western blotting confirmed the presence of RALDH1, RALDH2, RARα and RARβ protein in the adult rat pineal gland (Fig. 1b). This demonstrates that cells in the pineal gland have the potential to synthesize, respond to and degrade retinoic acid.

Immunohistochemical Detection of RALDH1 and RARα in the Pineal Gland
The components required for RA synthesis and signaling were found to be present in the pineal gland, therefore their cellular localization was determined for those of which the antibodies worked by immunohistochemistry; this was the case for antibodies recognizing the enzyme RALDH1 and the receptor RARα. The pinealocyte is the principal cell type in the pineal gland, comprising more than 95% of cells [48]. Interstitial cells resembling astrocytes are also present, several of which express GFAP [49]. To determine which cell type express RARα and RALDH1, pineal gland sections were double-labeled with antibodies against SAG, a pinealocyte marker [48], and GFAP.
RARα-immunoreactivity was detected in the cytoplasm of cells throughout the pineal parenchyma, with an average of 14% of cells labeled (Fig. 2). None of these cells were found to coexpress SAG, however a small subset of 1% of RARαpositive cells co-expressed GFAP (Fig. 2c). RALDH1immunoreactivity was also detected in the cytoplasm of cells, with a uniform distribution throughout the pineal gland sections and an average of 6% of cells labeled (Fig. 3). There was no double-labeling of RALDH1 with SAG or GFAP detected. These results suggest that RARα and RALDH1 are predominantly present in the cytoplasm of SAG-negative pinealocytes, given the number of immunoreactive cells present and the lack of co-expression with GFAP, with a small proportion of RARα also present in GFAP-positive cells.

Diurnal Changes in the Expression of RA Signaling Genes
The expression of RA signaling genes over the 24 h light/dark cycle were investigated to determine if RA signaling is subject to diurnal changes. Expression of the genes encoding the RA receptors (Rara, Rarb and Rarg) and the enzymes for the first (Rdh10 and Rdh12) and second (Raldh1, Raldh2, and Raldh3) steps of RA synthesis were measured by qPCR at ZT0, 6, 12 and 18. Aanat, the gene encoding the melatonin synthetic enzyme AANAT, was used as a positive control and its mRNA expression showed a peak at the expected time of ZT18 (Fig. 4a) [50]. A circadian clock gene was also examined as a second positive control, Bmal1, which peaked at ZT0-6 ( Fig. 4b), as previously reported for the rat pineal gland [51]. The genes required for RA signaling all showed some level of expression throughout the light/dark cycle (Fig. 4c-j) suggesting that a background level of RA signaling is present at all times. Several of these genes showed significant diurnal changes in expression. The expression of Rdh10 was lowest at the start of the light period at ZT0 and peaked at ZT12, the start of the dark period (Fig. 4c); this gene encodes the enzyme required for the first step of the conversion of retinol to RA. At the following time point sampled, ZT18 the mid-point of the dark period, the Rarg gene peaked, increasing in expression following its lowest expression during the day at ZT0-6 ( Fig.  4i). This gene encodes one of the RA receptors necessary to transduce the RA signal. Another member of the RDH family, Rdh12, which reduces and so removes retinaldehyde available for RA synthesis [47], peaked at ZT0 at the end of the dark period (Fig. 4j), then decreased at ZT6 and remained low at ZT12 and ZT18. These results show there are diurnal changes in the expression of RA signaling genes at each stage of the signaling pathway that may collectively complement each other to promote diurnal changes in RA production and activity.

Diurnal Change in RA Production
The components required for RA production are present in the rat pineal gland at the mRNA and protein level. RA was measured at two time-points to confirm that the pineal gland synthesizes RA and to determine whether RA production is subject to diurnal changes. RA release was measured from pineal glands collected at the start and end of the light period, ZT0 and ZT12, respectively. Pineal glands were cultured for 2 h immediately following dissection to allow release of RA into the medium, followed by analysis of RA concentration in the spent media by a RA reporter cell line. RA was detected at both time-points and was found to be significantly higher at ZT0 compared to ZT12, with a sevenfold difference in concentration (Fig. 5). These results demonstrate that the pineal gland is capable of producing RA and this is subject to diurnal changes, with higher levels present at the start of the light period compared to the end.

Diurnal Changes in RA Activity
Significant diurnal changes in the expression of several genes encoding RA signaling proteins and in RA production suggest that RA signaling is subject to diurnal changes. RA is a potent regulator of gene transcription, therefore to investigate changes in RA activity, diurnal changes in two target genes, Cyp26a1 and Cyp26b1, were determined. These genes encode Fig. 4 Retinoic acid signaling genes exhibit diurnal changes in expression in the rat pineal gland. qPCR analysis of expression of Aanat, Bmal1 and retinoic acid signaling genes in Sprague Dawley rat pineal glands collected at different zeitgeber times (ZT) throughout a 24 h cycle; the horizontal bar represents the light/dark cycle, white bar = light phase, black bar = dark phase; grey shading indicates the dark period. Values represent mean mRNA expression relative to Gapdh, ± SEM; with the exception of Raldh2 which represent median mRNA expression relative to Gapdh, with interquartile range, as these data are not normally distributed. N ≥ 6 glands per time-point. *P < 0.05; **P < 0.01; ***P < 0.001 the RA catabolic enzymes and have been shown to be potently induced by RA [52]. Firstly, Cyp26a1 and Cyp26b1 were shown to be rapidly upregulated by RA in cultured ex vivo rat pineal glands (Fig. 6a). A 4-h RA treatment induced a 26-and 37-fold increase in expression of Cyp26a1 and Cyp26b1, respectively. This demonstrates that these genes are responsive to RA in the rat pineal gland and are therefore indicators of RA activity. qPCR analysis of diurnal changes in expression of these genes found that overall there are significant differences in Cyp26a1 (P = 0.0446; Fig. 6b), and it appears that the lowest expression is at ZT12, the start of the dark period. Cyp26b1 expression exhibited a similar trend though this was not statistically significant (P = 0.2809). This suggests a diurnal change in RA activity occurs endogenously in the rat pineal gland.

Changes in Expression in Constant Darkness
The RA signaling genes Rdh10, Rarg, Rdh12 and Cyp26a1 were shown to be differentially expressed between day and night. To determine if these changes persist in the absence of external light cues, expression of these genes over 24 h in the pineal glands of rats maintained in constant darkness were analyzed (Fig. 7). In constant darkness, Rarg and Rdh12 maintained the same patterns of expression as in a normal LD cycle. In contrast, Rdh10 expression, which displayed significant variation under LD conditions (Fig. 4), did not vary under conditions of constant darkness. Cyp26a1 expression was very low at all time-points sampled, suggesting that expression is reduced under constant darkness compared to LD. The finding that the diurnal changes in Rarg and Rdh12 persist in constant darkness suggest that they are driven by the endogenous circadian pacemaker, whereas the differences observed in the patterns of Rdh10 and Cyp26a1 expression between the LD and DD conditions suggest that the diurnal changes in these genes require input from the external light/dark cycle.

NE Regulation of RA Signaling Genes
The majority of genes that cycle between day and night in the rat pineal gland are regulated by NE [35], therefore it was investigated whether the RA signaling genes that showed diurnal changes in expression in vivo, (Rdh10, Rdh12, Rarg, and Cyp26a1) are regulated by NE in vitro. Cultured pineal glands were treated with 1 μM NE for 4 h and gene expression determined by qPCR (Fig. 8). Rdh12 and Rarg were responsive to NE and demonstrated a decrease in expression of 3.7and 1.7-fold, respectively. There were no changes in Rdh10 or Cyp26a1 expression in response to NE (P = 0.3906 and 0.1692, respectively).

Effect of RA on Aanat Induction
VAD has previously been shown to reduce the peaks of the rhythms in AANAT activity and melatonin in the rat pineal gland [26,27]. In rodents, the AANAT rhythm is predominantly under transcriptional control with an increase in Aanat mRNA of more than a 150-fold at night in rats [25]. This study has demonstrated that RA, the active metabolite of vitamin A, is produced by the pineal gland on a diurnal basis and the components necessary for RA signaling are present, therefore the effect of RA on Aanat transcription was determined. Cultured pineal glands were treated with vehicle control, 1 μM RA or 1 μM NE for 4 h and gene expression analyzed by qPCR. As expected, Aanat responded robustly to treatment with NE, with expression increasing by 93-fold (Fig. 9), but there was no effect of RA on Aanat. As NE is the endogenous mediator of Aanat induction, the effect of RA combined with Fig. 7 Diurnal changes in retinoic acid signaling genes, Rarg and Rdh12, persist in constant darkness in the rat pineal gland. qPCR analysis of expression of retinoic acid signaling genes in pineal glands of Sprague Dawley rats maintained in constant darkness for 2 days. Pineal glands were collected at different circadian times (CT) throughout a 24 h cycle; CT0 and CT12 correspond to subjective light on and light off, respectively; horizontal black bar and grey shading indicate the dark period. Values represent mean mRNA expression relative to Gapdh, ± SEM. N = 5 glands per time-point. **P < 0.01; ***P < 0.001 Fig. 8 Norepinephrine represses Rarg and Rdh12 gene expression. qPCR analysis of cultured rat pineal glands following 4-h treatment with vehicle control or norepinephrine (NE). Values represent the fold change in mean mRNA expression compared to control, ± SEM. N ≥ 3 glands per treatment. *P < 0.05; **P < 0.01 NE was also determined. Treatment of pineal glands with NE and RA induced an increase in Aanat comparable to treatment with NE alone, suggesting that RA does not influence NEmediated induction of Aanat. Expression of the RAresponsive gene, Rarb, was also determined as a positive control. This increased in response to treatments of both RA and NE combined with RA, with comparable expression observed between these two treatments, indicating that RA is active. Taken together, these results suggest that RA does not have a short-term effect on Aanat transcription in the pineal gland.
RA can also regulate gene expression indirectly, through upregulation of the expression of transcription factors [53]. RA may act through such a mechanism to influence Aanat transcription in the pineal gland, however this would require a longer treatment period for an effect to be observed. This was investigated using pineal glands obtained from P10-12 rats. Firstly, it was confirmed that a short-term RA treatment does not influence Aanat induction in this culture model, as was previously observed in cultured pineal glands from older rats. Cultured pineal glands were treated with vehicle control or 1 μM RA for 4 h and gene expression analyzed by qPCR. As observed previously, there was no effect of RA on Aanat transcription following the 4-h treatment period (Fig. 10a). The effect of long-term RA treatment on Aanat was then tested by treating pineal glands with 1 μM RA for 48 h followed by qPCR analysis. In response to RA, Aanat was significantly increased by twofold relative to vehicle control (Fig. 10b). These results suggest that long-term RA treatment can influence Aanat transcription.

RA Regulation of Kinase Activation
In the pineal gland, there is a daily rhythm in ERK1/2 phosphorylation with a peak during the night [26,38] and vitamin A is required for this rhythm [26]. RA can modulate ERK1/2 phosphorylation in a number of cell lines and primary neuronal cultures [6,[54][55][56] and has been shown here to exhibit diurnal changes in signaling in the pineal gland, suggesting it may have a role in driving the daily rhythm in ERK1/2 activation. The effect of RA on kinase phosphorylation in the pineal gland was therefore determined. Cultured pineal glands were treated for 10 min with 1 μM RA, vehicle control or EGF as a positive control, and phosphorylated and total protein levels of ERK1/2 and Akt determined by western blotting. As expected, EGF induced activation of both ERK1/2 and Akt as previously reported [57]. RA treatment resulted in a significant decrease in ERK1/2 phosphorylation of nearly 50% (Fig. 11a), with both ERK1 and ERK2 individually exhibiting a comparable change (data not shown). There was no change in Akt phosphorylation in response to RA (Fig. 11b). These results demonstrate that RA can rapidly downregulate ERK1/2 phosphorylation in the rat pineal gland.

Discussion
Previous studies have suggested an important role for vitamin A in circadian rhythms; its deficiency leads to disruption of daily rhythms in locomotor activity and clock gene expression [14][15][16], as well as reduction of the nocturnal peak in melatonin in the pineal gland and loss of rhythm in MAPK activation [26,27]. The results presented here show that RA, the potent active metabolite of vitamin A, is subject to diurnal changes in production and activity, and it can induce rapid changes in kinase activation in the rat pineal gland.
There are several lines of evidence presented here for rhythmic diurnal RA synthesis and signaling in the pineal gland (summarized in Fig. 12). Analysis of diurnal changes in expression of the RA signaling genes demonstrated significant changes at every stage of the signaling pathway, including RA synthesis from retinol, signal transduction and RA degradation. Furthermore different components of the pathway exhibit complementary patterns of gene expression which appear to act together to produce significant changes in RA activity. Rdh10, which encodes the rate-limiting enzyme required for RA synthesis from retinol [58], was found to be lowest at ZT0 and to peak at ZT12, the start of the dark period. This would be expected to lead to an increase in RA synthesis from retinol. At the next time point sampled, ZT18 the midpoint of the dark period, the gene encoding one of the RARs, Rarg, was found to peak, which may act to increase RA signaling to coincide with a rise in RA. At the end of the dark period, ZT0, there was a peak in Rdh12 expression, this encodes a retinal reductase enzyme that converts retinaldehyde to retinol [47], therefore reducing the amount of retinaldehyde available for conversion to RA, which may act to bring down RA levels at the end of the night. This coincided with a decrease in both Rdh10 and Rarg expression at ZT0. Significant diurnal changes were also observed in Cyp26a1 expression, a RA-responsive gene and therefore an indicator of RA activity, which suggest an increase in RA signaling during the night. This gene encodes one of the RA catabolic enzymes and diurnal changes in its expression are also likely to lead to changes in RA degradation, assuming they are translated to changes at the protein level.
The combined effect of the changes in different components of the RA signaling pathway at complementary timepoints appear to have a significant effect on RA synthesis, despite being relatively small changes in gene expression; measurement of RA production detected a sevenfold higher concentration at the end of the night compared to the start. This could be due to multiple components of the same signaling system acting in a concerted way leading to an additive effect. It is also possible that other compounds required for RA synthesis that were not measured here are changing diurnally to contribute to the diurnal change in synthesis. Although there is a transient postprandial peak in serum retinol, limited evidence suggests that there is no diurnal variation in circulating levels [59,60], however diurnal changes in retinoid binding proteins have been reported. Plasma levels of the carrier for retinol in the circulation, retinol binding protein 4 (RBP4), have been shown to oscillate across the light/dark cycle in mice [61]. Furthermore, expression of the gene encoding the retinol transport protein, cellular retinol binding protein 1 (CRBP1), is under circadian control and peaks during the subjective night in liver [62]. Also expression of Ttr, which encodes another retinol carrier protein through its association with RBP, has been shown to increase significantly during the night in the rat pineal gland [35], which may act to increase retinol uptake during the night. Fig. 11 Retinoic acid rapidly downregulates ERK phosphorylation. Activation of ERK1/2 (a) and Akt (b) was determined by western blotting following 10 min treatment of cultured rat pineal glands with vehicle control or retinoic acid (RA). EGF was tested as a positive control (n = 1). Phosphorylated levels were normalized to total ERK or Akt; values represent the fold change in mean compared to control, ± SEM. N ≥ 3 glands per treatment. ***P < 0.001, compared to control This is the first study to investigate diurnal changes in RA synthesis in the pineal gland or any other tissue. Bailey et al. [35] previously identified T3/RA signaling as one of four specialized functional gene groups in the rat pineal gland following a large-scale transcriptomic analysis, highlighting the importance of RA synthesis in this gland. However, the study did not pick up the relatively small changes in expression of the RA signaling genes that were observed here, most likely because they were below the threshold used to keep the false discovery rate low. Rdh12 was highlighted however, for being highly expressed in the pineal gland relative to other tissues. Similarly, changes were seen in Rorb, which encodes a nuclear receptor that binds RA with high affinity [20]. This was found to increase in expression during the night at ZT18, in line with previous studies [22,23]. This could act to increase RA activity during the night, consistent with the finding presented here of a peak in Rarg expression also at ZT18. An increase in the expression of retinoid receptors during the night has also been reported in the rat hippocampus; Rara, Rarb and Rxrb display circadian rhythms in expression which may be driven by direct transcriptional control by clock proteins via clock-responsive Ebox sites in their regulatory regions [15].
The day/night variations in Rdh12 and Rarg expression persisted under constant darkness, indicating that they are circadian in nature and driven by the endogenous circadian clock. This finding was supported by the result that NE regulates their expression in vitro. The nightly release of NE to the pineal gland from neuronal input from the superior cervical ganglia (SCG) is controlled by indirect innervation from the SCN, the site of the master clock, and is the primary regulator of circadian changes in gene expression in the pineal gland [35,63]. Therefore, it is likely that the circadian changes in Rdh12 and Rarg are driven by this input. In contrast, the diurnal changes in Rdh10 and Cyp26a1 were abolished under constant darkness at the time-points sampled, and these genes were found to be unresponsive to NE in vitro. This suggests that an alternative mechanism regulates the diurnal changes in Rdh10 and Cyp26a1, independent of the SCN-driven NE input, but which requires input from the external light/dark cycle. In the case of Cyp26a1 this is presumably RA itself and that RA synthesis is reduced under constant darkness and no longer cycles diurnally. Guillaumond et al. [26] reported that vitamin A is necessary for a diurnal rhythm in MAPK activation in the rat pineal gland by a mechanism that is independent of the SCG but that does require an intact SCN, demonstrating that an NE-independent route is important for the effects of vitamin A on pineal gland rhythms. This may also be important for generating the rhythm in RA synthesis. These findings suggest that the rhythmic RA signaling in the pineal gland is under the control of at least two modes of regulation-one that is driven by the circadian clock and another that is reliant on input from the external light/dark cycle. The RA rhythm may therefore constitute a unique rhythmic signaling system in the pineal gland, as other rhythmic events are almost exclusively driven by the rhythmic release of NE [35].
The diurnal rhythm in RA synthesis observed here makes it well-suited for a role in the rhythmic functioning of the pineal gland. Activation of ERK1/2 is subject to diurnal oscillations in the rat pineal gland, with an increase in levels of the phosphorylated forms of ERK1/2 shortly following onset of darkness [26,38]. Vitamin A is required for this molecular rhythm [26], but it is currently unknown how its effects are mediated. RA was shown here to downregulate ERK1/2 phosphorylation in cultured pineal glands, therefore it may be involved in driving the rhythmic activation in vivo. There is a decrease in ERK1/2 activation towards the end of the night [26] when RA levels were found to be high, while lower RA levels at the start of night may permit the upregulation in ERK1/2 phosphorylation. ERK1/2 is involved in the regulation of the circadian clock in the pineal gland [39,40] and there is also evidence that it serves a modulatory role in the nightly increase in AANAT activity [64], therefore its rhythmic activation may be an important signaling event for the generation of the rhythm in melatonin production. Fig. 12 Summary schematic of the proposed retinoic acid rhythm in the rat pineal gland. At the start of the light period, zeitgeber time (ZT) 0, measured retinoic acid (RA) concentration is at its highest. At this time, the expression of the rate-limiting enzyme for the conversion of retinol to retinoic acid, Rdh10, is lowest. While there is a peak in the expression of Rdh12, which encodes an enzyme that converts retinaldehyde to retinol, therefore reducing the amount of retinaldehyde available for conversion to RA. These changes in expression of the RA synthetic enzymes are likely to lead to the sevenfold reduction in RA synthesis during the day, resulting in the low RA concentration measured at ZT12. At this time, Rdh10 expression rises to its peak, which is expected to increase RA synthesis again during the night, returning RA concentration to the high levels detected at the end of night, at ZT0. This may be accompanied by an increase in RA signaling during the night, with a peak in expression of Rarg at ZT18, which encodes one of the RA receptors, and increase in expression of the RA-responsive gene, Cyp26a1