Dopamine modulates the retinal clock through melanopsin-dependent regulation of cholinergic waves during development

As we previously reported a shortening of the endogenous period of the retinal clock and a loss of clock gene rhythms in the photoreceptor layer in the adult Opn4^−/− mouse [21, 27, 43], we analyzed the potential impact of the absence of melanopsin on the onset and the maturation of the retinal clock. Similar to the wild-type mouse, PER2::Luc oscillations started at P5 in the Opn4^−/−::Per2^Luc mouse (Fig. 1B). However, the amplitude was lower compared to the wild-type mouse, and did not allow reliable analysis of the clock properties at this developmental stage. PER2::Luc oscillations then became more robust, with mostly similar amplitudes between genotypes in the different developmental stages. The main difference concerned the significant shortening of the period observed in all developmental stages in Opn4^−/−::Per2^Luc retinal explants (P8: 25.32 ± 0,24 h, p = 0.0011; P15: 24.55 ± 0.19 h, p = 0.01; P30: 24.01 ± 0.04 h, p = 0.0034; adult: 24.24 ± 0.16 h; p = 0.0011, Fig. 1C) suggesting that the absence of melanopsin has a permanent consequence on the period length of the retinal clock. A significant advance in the phase of PER2::Luc oscillations (CT 13.51 ± 0.28; p = 0.0011) was observed only at P15 in retinal explants from Opn4^−/−::Per2^Luc mice compared to wild-type explants (CT 16.54 ± 0.39; p = 0.0011). To analyse the relationship between the clock features (period, phase and amplitude) during development, we fitted a linear model function to wildtype and Opn4^−/−::Per2^Luc data from P8 to P60 (Additional file 2: Fig. S2). For both genotypes, we noticed a negative correlation between the phase and the period during development. This means that as the period become shorter over developmental time, the phase gets delayed (p ≤ 0.0001). There was no significant change in the slope between genotypes and across different developmental stages (p = 0.25). Additionally, we did not observe a significant correlation between the amplitude of the rhythm and the period for both genotypes (p = 0.16).

Although the role of DA in the regulation of daily and circadian retinal functions in adult is well known, its role during the ontogeny of the retinal clock has not been investigated. We first examined whether the dopaminergic system analyzed in vitro faithfully reflects in vivo conditions. Immunostaining with an antibody directed against tyrosine hydroxylase (TH), the rate-limiting enzyme of DA synthesis, confirmed that 2-DIV explants retained dopaminergic neurons at P8 and P30 (Fig. 3A). In addition, the relative expression of Th mRNA quantified at P8 showed unaltered levels (Fig. 3B: p ≥ 0.05) in both ex vivo (CT0 and CT12) and in vitro retinas after 2 days of culture (trough and peak of PER2::Luc oscillations). We then examined whether DA is required for the generation of PER2::Luc rhythms. Retinal explants were treated with both alpha methyl-L-tyrosine (L-AMPT) and reserpine (Res), inhibitors respectively of DA synthesis and vesicular DA uptake (Fig. 3C, red bars) and compared to untreated retinas at P8, P15, and P30 (Fig. 3C, grey bars). At the 3 developmental stages, PER2::Luc rhythms persisted after DA depletion, with no changes in the endogenous period at all stages (P8: 26.45 ± 0.23 h; P15: 25.50 ± 0.25 h; P30: 24.94 ± 0.11 h; p ≥ 0.05) by comparison to control retinas of the same stage. Additionally, DA depletion did not affect the phase of PER2::Luc oscillations at P8 (controls: CT 11.73 ± 0.53; depleted-retinas: CT 11.51 ± 0.87; p ≥ 0.05) and P30 (controls: CT 20.40 ± 0.65; depleted-retinas: CT18.17 ± 0.50; p = 0.053) and induced an advance at P15 (controls: CT 16.54 ± 0.39; depleted retinas: CT 14.73 ± 0.27; p = 0.01). The amplitude of the rhythm was slightly increased at both P8 (controls: 30.51 ± 2.04 cps; depleted-retinas 74.72 ± 32.05 cps; p = 0.045) and P15 (controls: 267.61 ± 83.42 cps; depleted-retinas: 580.14 ± 30.50 cps; p = 0.048). These results suggest that DA is not required for the generation of PER2::Luc rhythm during the retinal clock development. We then investigated whether DA supplementation affects the onset of spontaneous oscillations of PER2::Luc in vitro. In our culture system, the drugs were applied as a bolus and were active for at least 4 circadian cycles, as evidenced by the altered period of PER2::Luc rhythms of supplemented explants compared to untreated retinas. Additional file 3: Fig. S3 shows that P1 explants did not express circadian oscillations in the presence of DA for the first 4–5 days in culture (5-DIV P1 + DA). Furthermore, as observed in the non-supplemented 5-DIV P1 retinas, the oscillations began to emerge after 5 days in vitro with very low amplitudes. No significant difference was observed in the period and the phase between 5-DIV P1 explants supplemented with DA and P5 explants. We then treated explants from the same developmental stages (P8, P15, and P30) with DA (50 µM; light green bars), or apomorphine (Apo), a non-selective DA agonist (Fig. 3C; dark green bars). Interestingly, supplementation with DA or Apo at P8 significantly lengthened the endogenous period (DA: 27.16 ± 0.13 h, p = 0.002; Apo: 27.55 ± 0.15 h, p = 0.005) and delayed the phase (DA: CT 15.69 ± 1.01, p = 0.016; Apo: CT 16.89 ± 0.50, p = 0.005) of PER2::Luc oscillations by comparison to non-supplemented explants. Both drugs also lengthened the period at P30 (DA: 25.53 ± 0.22 h, p = 0.0008; Apo: 26.76 ± 0.73 h, p = 0.004). The effect of DA supplementation on the amplitude of the PER2::Luc retinal rhythm was variable, with an increase at P8 and P15 (respectively p = 0.0007 and p = 0.033) and a decrease at P30 (p = 0.006). We did not observe any effect of DA supplementation in the period or the phase of PER2::Luc oscillation in adults (respectively: 25.40 ± 0.65 h and CT 18,69 ± 1,21; p ≤ 0.05; data not shown).

Male and female C57BL/6 J Per2^Luc [83] and Opn4^−/−::Per2^Luc mice were housed in a temperature-controlled room (23 ± 1 °C), under 12 h light/12 h dark cycle (12L/12D, light intensity around 200–300 lx) with food and water ad libitum. All animal procedures were in strict accordance with current national and international regulations on animal care, housing, breeding, and experimentation and approved by the French Ministry of Higher Education, Research and Innovation (APAFIS #19,976–2,019,032,610,179,225). All efforts were made to minimize suffering. Animals were used at E18, P1, P5, P8, P11, P15, P30, and in 2 month-old adults. The day of birth corresponded to P1.

For E18, Per2^Luc pregnant mice were killed by cervical dislocation, and embryos were isolated. Animals were sacrificed by decapitation between P1 and P8 and cervical dislocation from P11. All animals were sacrificed one hour before light offset at zeitgeber time 11 (ZT11). Eyes were enucleated and placed on ice in Hank's balanced salt medium (HBSS; Invitrogen). Retinas were gently isolated from the rest of the eyecup and flattened, ganglion cell layer up, on a semi-permeable (Millicell) membrane in a 35 mm culture dishes (Nunclon) containing 1.2 mL Neurobasal-A (Life Technologies) with 2% B27 (Gibco), 2 mM L-Glutamine (Life Technologies) and 25 U/mL antibiotics (Penicillin/Streptomycin, Sigma), incubated at 37 °C in 5% CO2 for 24 h. From this step onwards, all manipulations of explants were performed under dim red light. After 24 h, at the projected ZT12, retinas were transferred to 1.2 ml of 199 medium (Sigma), supplemented by 4 mM sodium bicarbonate (Sigma), 20 mM D-glucose (Sigma), 2% B27, 0.7 mM L-Glutamine, 25 U/mL antibiotics (Penicillin/Streptomycin, Sigma) and 0.1 mM Luciferin (Perkin). Culture dishes were sealed and placed in a Lumicycle (Actimetrics, Wilmette, IL, USA) to record the global emitted bioluminescence. All medium changes were performed under dim red light.

When applicable, 199 medium was supplemented with different molecules that remained in the medium throughout the entire recording without further medium changes: DA (50 and 100 µM), alpha methyl-L-tyrosine, an inhibitor of DA synthesis (L-AMPT, 100 µM), reserpine, an inhibitor of vesicular DA uptake (Res, 10 µM), apomorphine, a non-selective DA agonist (Apo, 50 µM), a DA D1R antagonist SCH39166 (50 µM), a DA D2R antagonist L741626 (25 µM), carbenoxolone, a non-selective gap junction blocker (CBX, 100 µM), mecamylamine, an acetylcholine receptor antagonist (MMA, 100 µM) and 6,7-Dinitroquinoxaline-2,3-dione disodium salt (DNQX, 20 µM), a selective ionotropic glutamate receptor antagonist.

P8 retinal explants from Per2^Luc and Opn4^−/−::Per2^Luc mice were exposed to a 465 nm monochromatic light pulse (30 min, 10¹⁴ photons/cm²/s, CT16) using bright LED light sources (Superbrightleds) with or without MMA supplementation. To avoid putative bias linked to retina development in culture, light stimulation was applied on the first PER2::Luc oscillation at CT16. Briefly, CT16 was determined based on the time of occurrence of the trough of the first PER2::Luc oscillation using the beginning of culture (CT12) as a reference as previously described [27, 43]. Retinas from the same animal were either light-stimulated or used as a dark control. Phase shifts following light stimulation or MMA application were calculated as the difference between the trough time of the cycle after treatment compared to the trough time of the control retina from the same animal, as determined by Lumicycle Analysis software.

To quantify DA receptors, 2 retinas of wild-type and Opn4^−/− mice at P8 and P30 were dissected at ZT11, pooled, and stored at -80 °C until RNA extraction and quantification. For Th quantification, retinas were collected at CT0 and CT12 or after 2 days of culture at the trough or the peak of PER2::Luc oscillations. Total RNAs were extracted using Trizol reagent (Invitrogen) and reverse transcribed using random primers and MMLV Reverse Transcriptase (Invitrogen). Real-time RT-PCR was then performed on a LightCycler™ system (Roche Diagnostics) using the light Cycler-DNA Master SYBR Green I mix. Hypoxanthine ribosyl-transferase (Hprt) was used for the internal standardization of target gene expression. The efficiency and the specificity of the amplification were controlled by generating standard curves and carrying out melting curves. Relative transcript levels of each gene were calculated using the second derivative maximum values from the linear regression of cycle number versus log concentration of the amplified gene. Primer sequences were: Hprt sens ATCAGTCAACGGGGGACATA and reverse AGAGGTCCTTTTCACCAGCA; D1R sens CAGCCTTCATCCTGATTAGCGTAGGCG and reverse CTTATGAGGGAGGATGAAATGGCG; D2R sens CAGTGAACAGGCGGAGAATG and reverse CAGGACTGTCAGGGTTGCTA; D4R sens CGTCTCTGTGACACGCTCATTA and reverse CACTGACCCTGCTGGTTGTA; D5R sens CATCCATCAAGAAGGAGACCAAGG and reverse CAGAAGGGAACCATACAGTTCAGG, Th sens GAAGGGCCTCTATGCTACCC and reverse GGCATGACGGATGTACTGTG.

Retinal explants from Per2^Luc mice were cultured as previously described and fixed for 24 h in 4% paraformaldehyde (PFA) after 2 days in vitro (2-DIV). Retinas were blocked with 1% bovine serum albumin (BSA) and 0.3% Triton X-100 in 0.1 M phosphate buffered saline (PBS) for 2 h and then incubated in the sheep anti-TH (1/200, Millipore) antibody at 4 °C for 24 h. After rinsing, a secondary incubation was performed for 2 h with the secondary anti-sheep biotinylated IgG (1/200, Vector Laboratories, Burlingame, USA), rinsed twice and incubated with Cy3-conjugated streptavidin (1/500, Jackson Immunoresearch laboratories). Samples were then mounted and coverslipped using Vectashield (Vector Laboratories). For BMAL1 staining, pups at P3, P5 and P8 (n = 3 for each genotype) were rapidly anesthetized and perfused intracardially with warm heparinized saline followed by PFA fixative at 4 °C. The eyes were removed and post-fixed overnight in the same fixative at 4 °C, then rinsed in 0.1 M phosphate buffer (PB, pH 7.4). Eyes were cryoprotected in 30% sucrose in PB for 24 h and sections of the retina were made at 20 µm on a cryostat. Retinal sections were incubated overnight with the rabbit antibody BMAL1 (1/500, Novus Biological), followed by a biotinylated goat anti-rabbit for 2 h (1/200; Vector Laboratories) and a Cy3-conjugated streptavidin (1/500, Jackson Immunoresearch laboratories). Retinal sections were then counsterstained with DAPI (1/10000, Invitrogen) for 5 min. Micrographs were obtained using Leica confocal microscopy.

Rhythm onset was defined when the first complete bioluminescence oscillation was observed (first trough and peak clearly distinguishables). The period and the amplitude of PER2::Luc oscillations were determined using SigmaPlot 12.5 software by fitting a linearly detrended sinusoidal curve oscillating around a polynomial baseline to the first three complete oscillations from each sample as previously described [23, 27, 43]. The equation used was:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$y\left(x\right)=\left(a-b*x\right)*\mathrm{sin}\left(2*\pi *\frac{x+\varphi }{\tau }\right)+(c+d*x+e*{x}^{2}+g*{x}^{3})$$\end{document}yx=a-b∗x∗sin2∗π∗x+φτ+(c+d∗x+e∗x2+g∗x3)where a is the amplitude (difference between the peak and the mesor values in the detrented time series data), τ is the period (duration of a complete circadian cycle, time peak to peak), and φ the phase. The mesor represents the circadian rhythm-adjusted mean based on the parameters of the sinusoidal function fitted to the raw data. The phase referred to the circadian time of the fitted peak of the first oscillation in the current manuscript, and was calculated based on a time reference (CT12), which corresponds to the time of medium change corrected by the endogenous period [27, 43].

Statistical analyses (Statistica Software) were performed using non parametric Kruskal Wallis one-way ANOVA followed when significant by the Mann–Whitney post-hoc test in order to compare the period, the phase and the amplitude of PER2::Luc signals, pharmacological treatments, and the relative expression of DRs and Th mRNA during development in Per2^Luc mice. Non parametric Kruskal Wallis two-ways ANOVA, followed by the Mann–Whitney post-hoc test, were used to compare the clock parameters between Per2^Luc and Opn4^−/−::Per2^Luc mice during development. Data are represented as mean ± SEM, with consideration of p ≤ 0.05 as the threshold of statistical significance.

Additional file 1: Fig. S1. Representative bioluminescence recording of PER2::Luc retinal explants at embryonic day 18. No oscillations of PER2::Luc were detected in cultured retinal explants even after 6 days in culture.Additional file 2: Fig. S2. Scatter plots showing the relationship between retinal clock parameters in Per2^lucand Opn4^-/-::Per2^lucretinas during development. A. The phase versus the endogenous period of the retinal circadian clock shows a negative, linear correlation in both genotypes with no significant change in the slope between genotypes. B. The period versus the amplitude did not show a significant correlation between the period and the amplitude for both genotypes and across the different developmental stages.Additional file 3: Fig. S3. Dopamine did not influence the onset of the spontaneous oscillations of PER2::Luc in vitro. A. Representative bioluminescence recording of PER2::Luc 5 days in vitro P1 cultured explants supplemented with DA. The dotted black rectangles correspond to the analysis windows including the 3 first complete oscillations. B. Comparison of the means of the endogenous period, the phase, and the amplitude of 5-DIV P1, 5-DIV + DA and P5 retinal explants. The total numbers of retinas used were: 5-DIV P1, n=5; 5-DIV + DA, n=4; P5, n=6. Bars represent mean ± SEM.Additional file 4: Fig. S4. The relative expression of D1- and D2-like dopaminergic receptors are similar between wild-type and Opn4^-/–retinas at P8. Means of the relative expression of D1-like and D2-like receptors in the retinas of wild-type and Opn4^-/-mice. The total numbers of retinas used were: WT, n=6 ; Opn4^-/-, n=7. Bars represent mean ± SEM.Additional file 5: Fig. S5. The blockade of glutamatergic waves has no effect on the period and the phase of the retinal clock. A. Means of the period and the phase in controland retinas supplemented with 6,7-Dinitroquinoxaline-2,3-dione disodium saltin P8 explants from wild-type Per2^luc mice. The total numbers of retinas used were: C, n=4; DNQX, n=5. Bars represent mean ± SEM.Additional file 6: Table S1. Raw data.