Melatonin's role in the timing of sleep onset is conserved in nocturnal mice

Abstract

Melatonin supplementation strengthens non-restorative sleep rhythms and its temporal alignment in both humans and night-active rodents. Of note, although the sleep cycle is reversed in day-active and night-active (nocturnal) mammals, both, produce melatonin at night under the control of the circadian clock. The effects of exogenous melatonin on sleep and sleepiness are relatively clear, but its endogenous role in sleep, particularly, in timing sleep onset (SO), remains poorly understood. We show in nocturnal mice that the increases in mid-nighttime sleep episodes, and the mid-nighttime decline in activity, are coupled to nighttime melatonin signaling. Furthermore, we show that endogenous melatonin modulates SO by reducing the threshold for wake-to-sleep transitioning. Such link between melatonin and SO timing may explain phenomena such as increased sleep propensity in circadian rhythm sleep disorders and chronic insomnia in patients with severely reduced nocturnal melatonin levels. Our findings demonstrate that melatonin's role in sleep is evolutionarily conserved, effectively challenging the argument that melatonin cannot play a major role in sleep regulation in nocturnal mammals, where the main activity phase coincides with high melatonin levels.

Keywords: Circadian rhythms and sleep; Melatonin.

Fig. 1. Endogenous melatonin modulates nighttime activity and the timing of nocturnal sleep-onset.

a Average running wheel activity profiles for MT_1/2^+/+ (n = 5) and MT_1/2^−/− mice (n = 7). The vertical line divides the nighttime into 6-h bins, and Φ1 and Φ2 indicate the 1st and the 2nd half of the nighttime, respectively. b Comparison of the mean difference in activity between Φ1 and Φ2 across both genotypes (MT_1/2^+/+; p < 0.0001, and MT_1/2^−/− p > 0.05, Student’s t-Test). c Representative hypnogram showing NREM and REM sleep stages for MT_1/2^+/+ mice. Below are high-resolution hypnograms corresponding to the early dark phase (ZT13-ZT15 of Φ1) and the late dark phase (ZT21-ZT23 of Φ2). d Quantification of total sleep (), NREM sleep () and REM sleep () during the early and late night (ZT13-ZT15 and ZT21-ZT23, respectively) in MT_1/2^+/+ mice (n = 6). e Distribution of activity (), NREM (), and REM sleep () during the dark phase in MT_1/2^+/+ mice (n = 6). f Average running wheel activity profile for C57BL/6 mice (n = 6), treated with vehicle (veh.) on day 1 and melatonin (mel.) on days 2–6. Yellow and gray sections show light and dark phases, respectively. The actograms show wheel revolutions/10 min bins. g Comparison of the mean percent change in nocturnal activity between Φ1 and Φ2 between veh. and mel. treated C57BL/6 mice (C57BL/6 mice during Φ2 veh. (day 1) and mel. (day 6); p < 0.005, Student’s t-Test, n = 5). The data are presented as means ± S.E.M. * and *** indicate p < 0.05 and p < 0.0001, respectively. ns stands for p > 0.05.

Fig. 2. Comparison of immobility-defined sleep parameters between melatonin receptor proficient and deficient mice.

a₁ Circular bar plot showing the 1-h averages of the amount of infrared-based immobility-defined sleep distributed across a 12:12 h LD cycle in MT_1/2^+/+ mice (n = 7). ZT stands for Zeitgeber Time, and the blue line indicates the phase of peak endogenous melatonin levels. a₂ A box-and-whisker blot for the total amount of sleep for MT_1/2^+/+ mice (one-way ANOVA followed by pairwise Student’s t-Test, p < 0.005). a₃ A box-and-whisker blot for the number of sleep bouts (one-way ANOVA followed by pairwise Student’s t-Test, p < 0.05) for MT_1/2^+/+ mice. b₁ Circular bar plot showing 1-h averages of the amount of sleep distributed across a 12:12 h LD cycle in melatonin receptors deficient mice (MT_1/2^−/−, n = 8). b₂ A box-and-whisker blot for the total amount of sleep for MT_1/2^−/− mice (one-way ANOVA followed by pairwise Student’s t-Test, p > 0.05). The total amount of sleep during the 2nd half of night (Φ2) is comparable with the amount of sleep during the light phase. b₃ A box-and-whisker blot for the number of sleep bouts for MT_1/2^−/− mice (one-way ANOVA followed by pairwise Student’s t-Test, p < 0.05). c₁, d₁ Horizontal (home-cage) locomotor activity profiles for mice (MT_1/2^+/+ and MT_1/2^−/−) treated with vehicle or melatonin in drinking water. Comparison of day and night activity for MT_1/2^+/+ (c₂, n = 7) and MT_1/2^−/− (d₂, n = 8) mice between treatment groups (Student’s t-Test, p < 0.05). The data are presented as means ± S.E.M. * and *** indicate p < 0.05 and p < 0.0001, respectively. ns stands for p > 0.05.

Fig. 3. Melatonin potentiates light-induced suppression of locomotor activity and sleep induction.

a, b (Top) MT_1/2^+/+ mice (n = 8) pre-treated (i.p. injections) with vehicle (veh.) or melatonin (mel.) were exposed to a 3-h light pulse (LP) of different illuminance during the night. (Left) Representative actograms for vehicle and melatonin injected mice exposed to a LP every 3rd day. (Right) Quantification of the activity data. One-way ANOVA for illuminance of the vehicle treated mice, F_{(5, 30)} = 16.85, p < 0.0001. One-way ANOVA for illuminance of the melatonin treated mice, F_{(5, 18)} = 4.843, p _<0.005. Two-way ANOVA for interaction (treatment vs. illuminance), F_{(5, 48)} = 5.184 p < 0.00005, illuminance, F_{(5, 48)} = 13.48 p < 0.0001, and for treatment, F_{(1, 48)} = 86.09, p < 0.0001. The blue horizontal line corresponds to the baseline activity (BL). c (Left) Experimental protocol for three conditions (cond.); mice injected with veh. (n = 6), mel. (n = 6), or luzindole (luz) and mel. followed by a 3-h LP (65 lux). (Right) Quantification of photosomnolence in MT_1/2^+/+ mice. Melatonin enhanced light-induced immobility-defined sleep, and which was blocked when mice were pre-treated with luz. (Student’s t-Test, p < 0.05). d Quantification of video-based SO (left) and sleep duration (right) in MT_1/2^+/+ mice entrained to an 18:6 h LD cycle for mice injected (i.p.) with either veh. (n = 9) or mel. (n = 9) followed by a 1-h LP of 65 lux (Student’s t-Test, p < 0.05). * Indicates p < 0.05. The data are presented as means ± S.E.M.

Fig. 4. The effect of melatonin on photosomnolence is dependent on melatonin receptor signaling.

a Illustration showing the placement of EEG and EMG recording electrodes. The image was created with BioRender.com. b Representative hypnogram showing darkness control (left) and light-induced (right) NREM and REM sleep during the 1-h light pulse (LP; 65 lux) for a MT_1/2^+/+ mouse maintained under a 12:12 h LD cycle. c Quantification of total sleep (gray), NREM sleep () and REM sleep () under darkness and LP conditions. d MT_1/2^+/+ mice (n = 6) exposed to a 1-h LP at ZT19 () versus control mice exposed to darkness only (). Light-pulsed mice showed a significant effect of light on infrared-based locomotor activity (two-way ANOVA, p < 0.05). e MT_1/2^−/− mice (n = 6) exposed to a 1-h LP at ZT19 () versus control mice exposed to darkness only (). Light-pulsed mice showed no significant effect of light on infrared-based locomotor activity (two-way ANOVA, p > 0.05). f Comparison (Student’s t-Test, p < 0.05) of sleep onset latency and total immobility following a LP between MT_1/2^+/+ and MT_1/2^−/−. MT_1/2^+/+ mice exposed to darkness were active for the duration of the recording (ZT19-ZT20). MT_1/2^+/+ mice exposed to a LP fell asleep within 22 ± 5 min compared to 48 ± 6 min in MT_1/2^−/− mice (Student’s t-Test, p < 0.05). Sleep onset latency was identified as the time from ZT19 until the first sleep episode (60 s or longer of immobility). MT_1/2^+/+ mice exposed to a LP accumulated 25 ± 8 min of sleep compared to 8 ± 4 min of sleep in MT_1/2^−/− mice (Student’s t-Test, p < 0.05). The data are presented as means ± S.E.M. and * indicates p < 0.05.

Fig. 5. Melatonin differentially inhibits light induced SCN activation.

a Representative confocal images of cFos-protein expression in the SCN, and vSPVZ of MT_1/2^+/+ mice injected with vehicle (veh.; n = 4) or melatonin (mel.; n = 4), 30 min before a 1-h light pulse (LP; 65 lux). (Bottom) quantification of the cFos signal in the SCN, vSPVZ, and VLPO. Melatonin treatment inhibits light-induced cFos protein expression in the SCN (one-way ANOVA for the effect of melatonin on light-induced cFos expression in the SCN, F_{(3, 18)} = 12.45, p < 0.0001) with no significant change in cFos protein expression in the vSPVZ (one-way ANOVA, F_{(3, 20)} = 0.91, p_>0.8) and VLPO (one-way ANOVA, F_{(3, 20)} = 0.166, p > 0.7). b (Left) Representative SCN neuronal firing during a 60 min. LP for veh. or mel. injected mouse. (Right) Comparison of the 60 min. average percent change in SCN neuronal spike amplitude between veh. and mel. treated MT_1/2^+/+ mice (n = 3). The data are presented as means ± S.E.M. * and ** indicate p < 0.05 and p < 0.05, respectively. ns stands for p > 0.05.

Fig. 6. Melatonin receptor signaling sets the phase of the hippocampal pCREB “rhythm”.

a The diurnal pCREB rhythm is phase advanced in MT_1/2^−/− compared to MT_1/2^+/+ mice. For better visualization of the rhythms and their phase relationship, the 24-h data points are double plotted. b representative time series of pCREB signal (green) in the dorsal hippocampus of MT_1/2^−/− and MT_1/2^+/+ mice. Cosinor analysis was applied to the 24-h dataset.

Fig. 7. Modeling the evolutionarily conserved role of melatonin on sleep in diurnal and nocturnal species.

a₁ In diurnal species, the activity of the vSPVZ is predominantly regulated by the SCN. During the day, when the SCN is active (prevalent [green] neurons), the vSPVZ (predominantly inhibitory) is strongly inhibited, decreasing the inhibitory tone of the vSPVZ, resulting in increased alertness. a₂ During the night, when the SCN output is relatively neutral (equal number of silent [red] and active [green] neurons), the vSPVZ is less inhibited, increasing the inhibitory tone of the vSPVZ, resulting in reduced alertness. a₃ melatonin inhibits the SCN further (predominantly silent [red] neurons), further reducing the inhibitory tone of the vSPVZ, resulting in additionally reduced alertness (facilitating sleep). b₁ In nocturnal species, the activity of the vSPVZ is regulated by the SCN and the retina. During the day, light activates the SCN (prevalent [green] neurons) and the vSPVZ (predominantly excitatory). As a result, the excitation of the vSPVZ overcomes the inhibition by the SCN, resulting in a strong positive excitatory tone and inhibiting alertness. b₂ During the nighttime, the excitatory tone of the vSPVZ is regulated by the SCN only. A reduced nighttime SCN inhibitory tone and the absence of light-induced vSPVZ activation results in a net decrease in vSPVZ excitatory tone, promoting wakefulness. b₃ Similar to diurnal mammals at night, melatonin further inhibits the SCN’s inhibitory tone, resulting in the vSPVZ’s excitatory to inhibitory tone ratio increasing and enhancing the inhibition of alertness. The bold lines indicate strong signal strength, and hashed lines refer to a weak or absent signal. Modified from ref. .