In the rat, exogenous melatonin increases the ... - Wiley Online Library

European Journal of Neuroscience, Vol. 16, pp. 1090±1098, 2002

ã Federation of European Neuroscience Societies

In the rat, exogenous melatonin increases the amplitude of pineal melatonin secretion by a direct action on the circadian clock BeÂatrice Bothorel, SteÂphane Barassin, Michel Saboureau, SteÂphanie Perreau, Berthe Vivien-Roels, AndreÂ Malan and Paul PeÂvet Laboratoire de Neurobiologie des Rythmes, UMR 7518 CNRS/UniversiteÂ Louis Pasteur, Strasbourg, France Keywords: circadian rhythms, microdialysis, nonphotic cues, phase-shift, suprachiasmatic nuclei

Abstract The effect of exogenous melatonin on pineal melatonin synthesis was studied in the rat in vivo. Daily melatonin pro®les were measured by transpineal microdialysis over 4 consecutive days in rats maintained on a 12-h light : 12-h dark schedule (LD 12 : 12). Curve-®tting was used to determine the amplitude of the peak of melatonin production, and the times of its onset (IT50) and offset (DT50). A subcutaneous injection of melatonin (1 mg/kg) at the onset of darkness (ZT12) induced an advance of IT50 on the second day after the treatment, in 50% of the animals kept in LD. When the animals were switched to constant darkness, the treatment caused no detectable advance of IT50, while 70% of individuals showed a signi®cant delay in DT50 2 days after the injection. Locally infusing the drug by reverse microdialysis into the suprachiasmatic nuclei (SCN) failed to enhance the shift in melatonin onset. Following subcutaneous melatonin injection, a signi®cant increase (» 100%) in melatonin peak amplitude was observed. This increase persisted over 2 days and occurred only when the melatonin was applied at ZT12, but not at ZT6, 17 or 22. The effect was also observed when the drug was infused directly into the SCN, but not into the pineal. Thus, the SCN are the target site for the effect of exogenous melatonin on the amplitude of the endogenous melatonin rhythm, with a similar window of sensitivity as its phase-shifting effect on the pacemaker.

Introduction In mammals, circadian rhythms are driven by an internal pacemaker located within the suprachiasmatic nuclei (SCN). These rhythms are primarily entrained to the external light/dark cycle via visual projections from the retina to the SCN (Morin, 1994). In addition to this strong photic synchroniser, several nonphotic cues, such as restricted hypocaloric feeding (Challet et al., 1996), physical activity (Turek, 1989), and melatonin administration (Redman et al., 1983) are potent synchronisers of the circadian clock. In rats in vivo, the melatonin-induced entrainment of circadian rhythms is generally restricted to a speci®c time window, i.e. when the daily administration of the drug coincides with the onset of the subjective night (Redman et al., 1983; Cassone et al., 1986; Pitrosky et al., 1999). Furthermore, the entraining effect of melatonin is related to its phaseadvancing properties (Warren et al., 1993). Numerous studies, both in vitro or in vivo, have shown that an acute application of melatonin, near the onset of subjective night, causes a stable phase-advance in the rhythms. This was observed on the electrical activity rhythm of SCN neurons (McArthur et al., 1991), on the pineal N-acetyltransferase rhythm (Humlova & Illnerova, 1990) and on the wheel-running activity rhythm (Armstrong & Chesworth, 1987) in rodents. It is generally believed that the hormone mediates these effects through high af®nity receptors that are localized in the rat SCN (Vanecek et al., 1987; Gauer et al., 1993). This interpretation is Correspondence: Dr B. Bothorel, as above. E-mail : [email protected] Received 5 March 2002, revised 24 June 2002, accepted 16 July 2002 doi:10.1046/j.1460-9568.2002.02176.x

strengthened by the observation that the melatonin receptor antagonists block the phase-advancing effects of melatonin (Hunt et al., 2001; Weibel et al., 1999). However, uncertainty exists as to whether these receptors are crucial for the hormone's phase-shifting action in vivo. Indeed, if in vitro physiological doses (nM) of melatonin are suf®cient, in vivo phase-shifts occur only when supraphysiological doses (mM) are applied. No simple direct experimental evidence of SCN involvement in this effect has been available so far. Another potential effect of exogenous melatonin consists in the modulation of endogenous melatonin synthesis. Findings from Zaidan et al. (1994) in humans suggest that a single dose of melatonin is able to increase the amplitude of the endogenous melatonin peak. On the other hand, in the rat, chronic administration of melatonin in constant darkness decreases melatonin production by the pineal (Drijfhout et al., 1996). Further studies are required to con®rm these results, and to determine the site and mechanism of melatonin action. The aim of the present study is to determine, in vivo, whether the SCN are involved in the melatonin-induced phase-shifts, and to describe the effect of exogenous melatonin on pineal melatonin synthesis. In previous experiments, we have reported that in any individual the secretion pro®le of melatonin measured by transpineal microdialysis remains stable over at least 5 days (Barassin et al., 1999). Therefore, the daily melatonin rhythm can be used as a marker of the output of the circadian clock, each animal serving as its own control. The nocturnal melatonin pro®les recorded by transpineal microdialysis after exogenous melatonin administration, either peripherally or locally into the SCN, were analysed under various lighting conditions.

SCN as a melatonin target 1091 TABLE 1. Summary of the experiments to investigate the effect of exogenous melatonin on the nocturnal melatonin secretion measured by in vivo microdialysis in the rat pineal gland. Experiment

Group

N

Photoperiodic conditions

Amount of melatonin

Route of administration

Time of administration (ZT)

A

I II III IV V VI VII VIII IX X XI XII

5 8 7 8 4 6 5 3 4 6 6 4

LD LD LD DD DD LD LD LD LD LD LD LD

0.5 mg/kg 1 mg/kg Ringer 1 mg/kg Ringer 1 mg/kg 1 mg/kg 1 mg/kg 5 mM 2.5 mM 25 mM Ringer

s.c. s.c. s.c. s.c. s.c s.c. s.c. s.c. Probe Probe Probe Probe

12 12 12 12 12 6 17 22 11.75±12.25 11.75±12.25 11.75±12.25 11.75±12.25

B C D

Materials and methods Animals Male Wistar rats were used (300±350 g; DeÂpreÂ, St Douchard, France). The animals were housed in a temperature-controlled room (21 6 2 °C) and maintained on a reverse 12-h light : 12-h dark lighting regimen (lights off from 10:00 until 22:00 h) during at least 3 weeks before surgery. Dark onset (10 : 00 h) corresponded to Zeitgeber Time 12 (ZT12). Water and food were available ad libitum. The DeÂpreÂ rats were used for their particular nocturnal melatonin pro®le, which peaks late in the dark period of the LD cycle (Barassin et al., 1999). This choice was made to facilitate the observation of an advance in melatonin onset, if any, and to eliminate any confusion between the exogenous and endogenous melatonin peaks. All experiments with animals were performed in accordance with `Principles of laboratory animal care' (NIH pub. no. 86±23, revised 1985) as well as in accordance with the French national laws. Surgery and dialysis Animals were operated the week before the microdialysis experiments. They were anaesthetized with equithesine (0.4 mL/100 g body weight, i.p.). After surgery, they were allowed to recover for 1 week in individual cages. Pineal dialysis Implantation of the microdialysis probe was performed as previously described (Drijfhout et al., 1993; Barassin et al., 1999). The dialysis membrane of saponi®ed cellulose ester (0.22 mm inside diameter; 0.27 mm outside diameter, 10 000 molecular mass cut-off), born by a tungsten wire with a sharpened extremity, was fastened in a horizontal position to a holder mounted on a stereotaxic apparatus (David Kopf Instruments). A hole was drilled on each side of the temporal bone and the probe was inserted transversally through the pineal gland 1.6 mm ventral to the skull and 0.7 mm posterior to lambda according to the atlas of Paxinos & Watson (1982). At the other extremity, a blunted stainless-steel needle was glued to the membrane with epoxy resin. The tungsten wire was removed, after which both the inlet and outlet of the tube were ®xed to the skull in a vertical position with dental cement. During the experiments, the inlet of the probe was connected with polyethylene tubing to a microinjection pump (Pump 22, Harvard, Biosciences, Les Ulis, France) via a one- or two-way ¯uid swivel

(Pineal) (SCN) (SCN) (SCN)

(375/22, Instech Laboratories, Plymouth Meeting, PA, USA). The swivel was attached to a counterbalance beam allowing the animal to move freely. The probe was perfused with Ringer's solution at a ¯ow rate of 3 mL/min during experimental periods and 1 mL/min during rest periods. The outlet connection consisted of peek polymer tubing (0.13 mm inside diameter, 0.51 mm outside diameter) connected to a 1.5-mL polypropylene microvial. Collected samples were stored at ±20 °C until assayed by radioimmunoassay (RIA). At the end of the experiment, rats were decapitated, and the brain (with the pineal gland attached) was frozen at ±20 °C. Cryostat sections (25 mm) of the brain/pineal were stained with cresyl violet in order to determine the location of the probe. SCN dialysis Reverse dialysis was used to infuse exogenous melatonin into the SCN through a U-shape probe. This hand-made probe consisted of two stainless tubes (0.3 mm inside diameter; 0.4 mm outside diameter; 20 mm length) soldered together, with one of the tubes bent on its extremity. A stainless wire was pushed through 5 mm of the dialysis membrane tube (with the same properties as that used for pineal probe). The membrane tube, with wire enclosed, was bent to form a U-shape, after which it was glued to the double tubing with epoxy resin leaving membrane free over 3 mm. The U-shape probe was implanted into the lateral region of the SCN with an angle of 3°, 1.6 mm anterior to bregma, +0.5 mm lateral to midline and 8.5 mm ventral to dura (incisor bar at +5 mm). The probe was ®xed to the skull with dental cement. The day of exogenous melatonin infusion, the inlet of the probe was connected to a microinjection pump via a two-way swivel (Instech 375/22). The SCN was perfused with Ringer's solution or melatonin solution at 1 mL/min during 30 min around ZT12. Chemicals Melatonin was obtained from Sigma (France). It was dissolved in ethanol and diluted with Ringer, to a ®nal ethanol concentration (v/v) of 5%. Experimental procedure In all experiments, pineal melatonin release was measured during four consecutive nights. The ®rst, third and fourth days of sampling were designated as control days CTR1, CTR2 and CTR3, respectively. Administration of melatonin (Mel) or Ringer (Ring) by

ã 2002 Federation of European Neuroscience Societies, European Journal of Neuroscience, 16, 1090±1098

1092 B. Bothorel et al. subcutaneous injection or by reverse dialysis was performed at a de®ned ZT on the second experimental day, designated as MelZT or RingZT. Samples were collected hourly from ZT11 to ZT1, except during the 2 h following injection of melatonin where dialysates were sampled every 30 min. All experiments are summarized in Table 1.

1 : 100 000 and [125I]-2-iodomelatonin labelling. The limit of sensitivity of the assay was 0.5 pg per tube. The direct melatonin assay was validated for dialysates as already reported (Barassin et al., 1999).

Experiment A

For data analysis, pineal samples that had undetectable concentrations of melatonin were assigned the values of the limit of sensitivity of the assay (0.5 pg per tube). All data of melatonin concentrations were expressed in absolute values (nM). Each control melatonin pro®le was characterized by its onset time (IT50), its offset time (DT50) and the peak amplitude (Yampl). These parameters were determined by ®tting a logistic peak with the following equation:

In order to investigate the effect of exogenous melatonin on the endogenous rhythm of melatonin production, rats were divided into three groups. Group I (n = 5) and group II (n = 8) received a subcutaneous injection of melatonin on the second day at ZT12, 0.5 and 1.0 mg/kg b.w., respectively (100 mL ethanolic solution/100 g b.w.). Rats of the control group (group III, n = 7) were injected with an equivalent volume of Ringer's solution. Experiment B

Data analysis and statistics

Y = Y0 + Yampl/{[1 + e2.9´(IT50

± x)

] 3 [1 + e2.8´(x

± DT50)

]}

th

To study the phase-advancing effect of melatonin as a nonphotic synchroniser, we investigated the daily melatonin rhythm under constant darkness conditions (DD) following melatonin administration. After the measurement of melatonin secretion in LD 12 : 12 (CTR1), animals (group IV, n = 8) were injected with melatonin (1.0 mg/kg) and switched to DD until the end of the experiment (CTR2 and CTR3). Pineal dialysates were collected hourly on both days in dim red light (< 2 lux). Rats of the control group (group V, n = 4) were injected with an equivalent volume of Ringer's solution. Experiment C In this experiment, we investigated the effect of the injection of a single dose of melatonin (1.0 mg/kg, s.c.) under LD conditions at different Zeitgeber times. Rats of group VI (n = 6) were injected at ZT6, i.e. in the middle of the light period. In group VII (n = 5) melatonin was administered around the middle of the dark period, at ZT17. Rats of group VIII (n = 3) were injected at ZT22, around the end of the night. Group II from experiment A represented the rats injected at ZT12, the day/night transition. Experiment D This experiment aimed at determining the effect of local application of exogenous melatonin on pineal metabolism. In a ®rst study, a 5-mM melatonin solution was infused into the pineal gland through the transverse probe at 3 mL/min during 30 min around ZT12 (group IX, n = 4) on the second experimental day. The probes were perfused with Ringer for the remainder of the time. To determine whether the SCN is the site of action of exogenous melatonin on the endogenous rhythm of melatonin synthesis, a second study was performed using reverse dialysis to infuse the drug locally. On the second experimental day, melatonin was perfused in the SCN through the U-shape probe at 1 mL/min during half an hour around ZT12. Experimental groups infused either with 2.5 or 25 mM melatonin were designated as group X (n = 6) and group XI (n = 6), respectively. Following melatonin infusion of group XI, dialysates were also collected during 4 h from the SCN probe outlet. We therefore observed that exogenous melatonin was rapidly eliminated within 2 h. A control group was infused into the SCN with Ringer's solution at 1 mL/min around ZT12 (group XII, n = 4). In each group, dialysates were collected hourly from the pineal probe outlet. Radioimmunoassay Melatonin concentrations in dialysates were determined in duplicate in 25 mL samples by RIA using a speci®c rabbit antiserum (R19540) provided by INRA (Nouzilly, France) at a ®nal dilution of

where Y was the n data point, x represents the time point of the nth point, Y0 the basal level during daytime, and Yampl the amplitude of the nocturnal peak over the basal level. IT50 was the time point at which 50% of the increase in melatonin level was reached and DT50 the time at which 50% of the decrease occurred. The multiplicative factors in the exponentials (2.9 and 2.8) were chosen arbitrarily to ®t the observed slopes (Barassin et al., 1999). The nonlinear regression analysis was performed with SIGMAPLOT software (SPSS ASC GmbH, Erkrath, Germany) and ®tted to the data points of each experiment. Compared with the intuitive methods used in the past, the logistic function has a major advantage: the onset and offset times calculated present little dependency, if any, on the peak amplitude. This was systematically veri®ed in the present study. In the absence of standard statistical methods for variancecovariance analysis involving nonlinear regression, we assessed the effect of treatment in two different ways. In both, the data were estimates of the parameters as derived from the nonlinear regression analysis (asymptotic estimates). First, an ANOVA for repeated measures was carried out for each group of rats on IT50, DT50 and Yampl with animals and the control days (CTR1, CTR2 and CTR3) as factors. In this approach, we assumed that the dispersion of each parameter estimate (as calculated from the regression) was small enough compared with the other sources of dispersion to be ignored. The result of this analysis was considered as representing a common effect on a group of animals. Secondly, individual comparisons of CTR2 and CTR3 vs. CTR1 were performed for each parameter with t-tests using the errors (SE) given as the asymptotic estimates of each parameter by the regression analysis. The level of statistical signi®cance was set to 0.05 throughout. All data are expressed as mean 6 SEM, except when otherwise speci®ed.

Results Experiment A: subcutaneous injection of melatonin in LD Averaged pro®les of melatonin Figure 1A presents a typical example of melatonin pro®les observed during 4 successive days in LD 12 : 12 on one rat from group II. On the second experimental day (MelZT12), melatonin injection (1 mg/ kg) at ZT12 immediately produced a high peak of exogenous melatonin, but 2 h later, melatonin concentration had returned to basal level. This ensures that further pro®les re¯ected the endogenous melatonin pro®les. On days CTR2 and CTR3 following melatonin injection (MelZT12), the rat showed a signi®cant advance in IT50 of


SCN as a melatonin target 1093

FIG. 2. Averaged nocturnal pro®les of pineal melatonin concentrations measured during the ®rst control day (CTR1) and during the 2 days (CTR2 and CTR3) following a single subcutaneous injection of melatonin 1 mg/kg at ZT12 under LD conditions. Data are plotted as the mean 6 SEM (Group II, n = 8). The black bar represents the dark period of the LD 12 : 12 cycle.

group II (Fig. 2). The mean melatonin peak amplitude was found to increase by » 100% the day after melatonin injection, as compared to the ®rst control day, CTR1. This was a long-term effect as it was still observed 2 days after the injection. In the following, individual data observed after melatonin treatment (CTR2, CTR3) will be presented as deviations from control day level CTR1, and expressed in percentages of the control for amplitude (CTR1 = 100%) and in min for IT50 and DT50 (positive for an advance and negative for a delay). Effect of melatonin dose

FIG. 1. Examples of individual pro®les of pineal melatonin concentrations during the ®rst control day (CTR1) and during the 2 days (CTR2 and CTR3) following: (A) a single injection of melatonin 1 mg/kg at ZT12 (MelZT12) under LD conditions; (B) an injection of Ringer's solution 1 mL/kg at ZT12 (RingZT12) under LD conditions. The black bar represents the dark period of the LD 12 : 12 cycle.

» 80 min vs. CTR1. Unexpectedly, this was associated with a signi®cant increase in the amplitude of melatonin rhythm. The averaged day/night rhythms of melatonin (mean 6 SEM) observed in group II (n = 8) during the 3 control days, before (CTR1) and after (CTR2, CTR3) melatonin injection at the dose of 1 mg/kg, are shown in Fig. 2. After melatonin treatment, a slight phaseadvance in melatonin onset, 22.7 6 12.5 min, was observed on CTR3 compared to CTR1 under LD conditions, with a high interindividual variability. The effect of melatonin on the peak amplitude shown above for one rat was consistently observed in the whole

The mean values of amplitude observed on the ®rst control day CTR1 were 3.3 6 0.6 and 3.0 6 0.5 nM for group I (n = 5) and group II (n = 8), respectively. The variations in amplitude of pineal melatonin peak after a single subcutaneous injection of melatonin at ZT12 in LD 12 : 12 are shown in Fig. 3 for the doses of 0.5 mg/kg (group I, n = 5) and 1 mg/kg (group II, n = 8). In group II, the ANOVA revealed a signi®cant increase (P = 0.009) of » 100% in nocturnal melatonin amplitude during the 2 days (CTR2, CTR3) following melatonin injection, compared to the ®rst control day. On CTR2 and CTR3, the mean melatonin amplitude increases were 3.2 6 0.9 and 3.0 6 0.9 nM, respectively, as compared to CTR1. Individual analyses with t-tests, using the error given by the regression analysis, showed that all the eight animals of group II responded with an increase in melatonin peak amplitude during 2 days after melatonin treatment. With the dose of 0.5 mg/kg (group I), the effect was not signi®cant (ANOVA; P = 0.27). Mean amplitude values observed on CTR2 and CTR3 were 4.4 6 0.2 and 3.8 6 0.8 nM, respectively. Variations in IT50 and DT50 after melatonin injection compared with the ®rst day are shown in Fig. 4 for group II (1 mg/kg). Mean IT50 values were 18.7 6 0.6, 18.4 6 0.7, and 18.3 6 0.6 h on CTR1, CTR2, and CTR3, respectively. Deviations from the ®rst day


1094 B. Bothorel et al.

FIG. 3. Effect of exogenous melatonin on the amplitude of the nocturnal melatonin peak determined by in vivo microdialysis. Melatonin (0.5 or 1.0 mg/kg) was injected subcutaneously at ZT12 under LD conditions and pineal dialysis samples were collected afterwards during 2 days (CTR2 and CTR3). The amplitudes of the melatonin pro®les measured on CTR2 and CTR3 are expressed as percentages of the ®rst control day CTR1. Open symbols: individual values following melatonin treatment. Closed symbols: means 6 SEM.

were not signi®cant (ANOVA; P = 0.172). However, IT50 variation tended toward a phase-advance, as 50% of the animals presented a signi®cant advance in IT50 of at least 15 min, on CTR3 compared to CTR1, as shown by individual t-tests. Mean DT50 values observed on CTR1, CTR2 and CTR3 were 23.4 6 0.1, 23.2 6 0.2 and 23.2 6 0.2 h, respectively, and did not present any signi®cant difference (ANOVA; P = 0.235). In group I, melatonin injection at 0.5 mg/kg affected neither melatonin onset nor offset in any animal (data not shown). The control group III (n = 7) showed no signi®cant difference after the subcutaneous injection of Ringer at ZT12 on the second day, neither in amplitude nor in IT50 or DT50. The typical melatonin pro®les of a rat from this group are represented on Fig. 1B. Mean values (6 SD) over the 3 control days were 10.6 6 4.1 nM for amplitude, 17.9 6 1.1 h for IT50 and 23.2 6 0.4 h for DT50. Experiment B: subcutaneous injection of melatonin in DD In group IV (n = 8), rats were subcutaneously injected with a 1-mg/ kg melatonin solution at ZT12 and kept under DD conditions until the end of the experiment. A typical example of individual pro®les of one rat from group IV is represented in Fig. 5. This rat showed a slight increase in the amplitude of its melatonin peak on CTR2 and CTR3, and a marked delay in DT50 on CTR3. The amplitude of melatonin peak was 3.6 6 0.8 nM on CTR1. Under DD conditions, after melatonin treatment the amplitude of melatonin peak signi®cantly increased (ANOVA; P = 0.025). Five out of eight animals presented a signi®cant increase in melatonin amplitude on CTR2 vs. CTR1, and this was maintained on CTR3 in four animals, as revealed with t-tests. The mean increases in melatonin amplitude were 0.7 6 0.3 and 0.8 6 0.2 nM on CTR2 and CTR3, respectively, compared to CTR1. The increase in melatonin peak amplitude was only » 30% under DD conditions (Fig. 6A). In the control group V (n = 4), which received Ringer at ZT12 on the second day, and were then kept under DD, three out of four animals

FIG. 4. Variations in the onset IT50 (A) and the offset DT550 (B) of the melatonin pro®les observed on CTR2 and CTR3 vs. CTR1 following a single injection of melatonin (1 mg/kg) at ZT12 under LD conditions. Open symbols: individual values. Closed symbols: means 6 SEM.

presented a signi®cant decrease in melatonin amplitude on CTR3 vs. CTR1, as revealed with t-tests. The mean decrease in melatonin amplitude was 3.6 6 1.5 nM on CTR3 compared to CTR1, i.e. » 30% (amplitude of melatonin peak at CTR1 was 10.3 6 4.3 nM). Concerning the phase parameters, ANOVA (P = 0.076) did not reveal any signi®cant effect of melatonin treatment on IT50. However, the analysis of variance revealed a highly signi®cant variation in DT50 (P < 0.001) due to DD conditions. If only two out of seven animals presented a signi®cant DT50 delay on CTR2, ®ve out of seven animals had a signi®cant delay in DT50, higher than 40 min, on CTR3 vs. CTR1. The mean DT50 was 23.3 6 0.1 h on CTR1. A signi®cant delay of 52 6 6 min took place on CTR3 (Fig. 6B). The control group showed no signi®cant difference neither


SCN as a melatonin target 1095

FIG. 5. An example of individual pro®les of pineal melatonin concentrations during the ®rst control day (CTR1) and during the 2 days (CTR2 and CTR3) following a single injection of melatonin 1 mg/kg at ZT12 (MelZT12, only values over 2 h after the injection have been reported). CTR1 and MelZT12 were studied under LD conditions and, CTR2 and CTR3 under DD conditions. The black bars represent the dark periods.

in IT50 nor in DT50. Mean values (6 SD) over the 3 control days were 15.5 6 1.3 h for IT50 and 22.9 6 0.4 h for DT50. Experiment C: subcutaneous injection of melatonin at different ZTs Melatonin peak amplitudes were compared to CTR1 after a subcutaneous injection of melatonin (1 mg/kg) at different Zeitgeber times: ZT6 (group VI), ZT12 (group II), ZT17 (group VII) and ZT22 (group VII) (Fig. 7). No signi®cant effect was detected by ANOVA, neither at ZT6 (P = 0.085), ZT17 (P = 0.35) nor ZT22 (P = 0.95). The mean values (6 SD) for melatonin amplitude over the 3 control days were 4.8 6 3.9, 5.9 6 2 and 8.2 6 6.7 nM under ZT6, ZT17 and ZT22 conditions, respectively. The only signi®cant effect (ANOVA; P = 0.009) of melatonin on the peak amplitude was observed when melatonin was applied at ZT12, the day/night transition, as described in experiment A. As expected, the ANOVA detected no treatment effect neither on IT50 nor on DT50 when melatonin was applied at ZT6, ZT17 or ZT22. Experiment D: local infusion of melatonin by reverse microdialysis Effect of melatonin infusion into the pineal When perfused locally within the pineal gland through the dialysis probe for 30 min around ZT12 (data not shown), exogenous melatonin (5 mM) had no signi®cant effect on the parameters of nocturnal melatonin pro®le (Yampl, IT50, DT50), compared with CTR1, for the rats from group IX (n = 4). Effect of melatonin infusion into the SCN Rats of group X (n = 6) and group XI (n = 6) were infused locally within the SCN through the U-shape dialysis probe, during 30 min

FIG. 6. Variations in the amplitude (A) and of the offset DT50 (B) of the melatonin pro®les observed on CTR2 and CTR3 vs. CTR1 following a single injection of Ringer or melatonin (1 mg/kg) at ZT12 and studied under DD conditions. Open symbols: individual values. Closed symbols: means 6 SEM.

around ZT12, with 2.5 and 25 mM melatonin solutions, respectively (Fig. 8). After melatonin infusion at the dose of 2.5 mM, the nocturnal pro®le of melatonin showed a slight, nonsigni®cant, increase in amplitude. In contrast, infusion of 25 mM melatonin into the SCN induced a signi®cant increase in endogenous melatonin amplitude (ANOVA; P = 0.041). The mean amplitude on CTR1 for group XI was 4.2 6 1.6 nM and from this the mean increase in amplitude was 2.2 6 0.8 and 2.9 6 1.3 nM on CTR2 and CTR3, respectively, i.e. » 50% increase. When tested individually, ®ve animals out of six presented a signi®cant increase in melatonin peak amplitude on CTR2, which was maintained on CTR3 in four animals.


1096 B. Bothorel et al.

FIG. 7. Effect of exogenous melatonin on the amplitude of the nocturnal endogenous melatonin peak determined by in vivo microdialysis. Melatonin (1.0 mg/kg) was injected subcutaneously at different Zeitgeber times ZT6, ZT12, ZT17 and ZT22 under LD conditions. Pineal dialysis samples were collected afterwards during 2 days (CTR2 and CTR3). The amplitudes of the melatonin pro®les on CTR2 and CTR3 are expressed as percentages of the ®rst control day CTR1. Open symbols: individual values. Closed symbols: means 6 SEM.

FIG. 8. Effect of exogenous melatonin infused locally within the SCN by reverse microdialysis on the amplitude of the endogenous nocturnal melatonin peak. Melatonin (2.5 or 25 mM) was applied within the SCN between ZT11.75 and ZT12.25 and pineal dialysis samples were collected over 2 days (CTR2 and CTR3) after melatonin perfusion. The amplitudes of the melatonin pro®le measured on CTR2 and CTR3 are expressed as percentages of the ®rst control day CTR1 designated as 100%. Open symbols: individual values. Closed symbols: means 6 SEM.

No signi®cant change in the phase parameters IT50 and DT50 took place after melatonin infusion in the SCN, neither at 2.5 mM nor at 25 mM. In the control group XII (n = 4), Ringer was infused within the SCN during 30 min around ZT12. No signi®cant variation in the parameters of the nocturnal peak of melatonin in the pineal gland was found in that condition.

melatonin in the absence of the photic signal. To test this, we exposed animals to DD conditions for 2 days after a single peripheral melatonin administration. Under these experimental conditions, the averaged IT50 of nocturnal melatonin secretion showed no signi®cant phase-advance, but a highly signi®cant delay of the decline of melatonin secretion could be observed on the second subjective day under DD conditions. When rats are transferred from LD to DD, a delay of both onset and offset has been shown for NAT or melatonin nocturnal peaks, as well as a lengthening of the peak duration due to a more rapid phase-delay in the offset than in the onset (Illnerova & Vanecek, 1982; Drijfhout et al., 1996). These changes in the characteristics of the free-run have been observed, however, after several days/weeks of exposure to DD conditions. After only 2 days under DD conditions, such free-runinduced changes are probably not large enough to be detected in most of the animals. This could explain why in the present experiment, no change, neither in the onset nor in the offset of melatonin peak, was observed after saline injection. The delay observed in the decline of melatonin peak after a single melatonin injection might therefore be the consequence of an acceleration of the free-run expression in the animals of this group. This conclusion however, remains to be experimentally demonstrated.

Discussion Modi®cation of melatonin onset The present study showed that under LD 12 : 12 conditions, a single administration of exogenous melatonin can phase advance the onset of pineal melatonin peak. This advance, if any, occurred only when melatonin was administered at the day/night transition. This is in accordance with the `narrow window of sensitivity' of melatonin to phase-shift as described by Armstrong & Chesworth (1987), and with the time-domain of the clock sensitivity, as reviewed by Gillette & Mitchell (2002). However, the phase-advancing effect of melatonin was not consistent as it was observed in only 50% of the animals. Most probably, such variability would result from the presence of the light/dark cycle, a strong synchroniser. In contrast, Humlova & Illnerova (1990) described a clearcut phase-advance of the evening N-acetyltransferase (NAT) rise in the rat pineal after a single melatonin injection, under LD conditions. This, however, is only an apparent discrepancy. Indeed, this NAT phase-advance was only observed under short photoperiodic conditions (LD 10 : 14 and LD 8 : 16), suggesting that the resetting effect of melatonin is photoperiod-dependent (Humlova & Illnerova, 1992). When applied under DD conditions, either in vivo or in vitro, exogenous melatonin induces a stable phase-advance in the rhythms (Armstrong & Chesworth, 1987; McArthur et al., 1991; Pitrosky et al., 1999). This suggests that the clock would be more sensitive to

Amplitude variation after melatonin treatment Our study reveals that a single application of melatonin increases the amplitude of the nocturnal peak of endogenous melatonin and that this effect is still present 2 days after the treatment. Even though previous work in humans (Zaidan et al., 1994) suggests that an ampli®cation of the nocturnal melatonin rhythm after exogenous melatonin administration occurs, this is the ®rst experimental demonstration of such an effect. Moreover, we have demonstrated that the effect of melatonin on the ampli®cation of the endogenous nocturnal peak occurs only when the drug is applied at the day/night transition. This particular effect of melatonin is therefore time-


SCN as a melatonin target 1097 dependent. The time-window in which melatonin induces an increase in endogenous melatonin peak amplitude is similar to the timewindow of clock sensitivity in which rhythm entrainment or phase shifts are generally observed (Armstrong & Chesworth, 1987; Humlova & Illnerova, 1990; Pitrosky et al., 1999). This result raises questions about the target area(s) of melatonin that enhance the production of endogenous melatonin. As exogenous melatonin is known to modulate melatonin biosynthesis in the pineal through an inhibition of the serotonin reuptake (Miguez et al., 1996), a possible direct effect on the pinealocytes has to be considered. As melatonin rhythm was unaffected after melatonin infusion, by reverse microdialysis within the pineal gland itself, such a direct action can be ruled out. The results obtained, i.e. a sustained increase in the nocturnal melatonin peak for at least 2 days, when melatonin was applied locally within the SCN (double microdialysis), clearly demonstrate the implication of the SCN. Which could be the mechanisms involved? Kalsbeek et al. (1999) have shown that melatonin biosynthesis in the pineal gland is controlled by GABA release from SCN nerve terminals in the paraventricular nucleus (PVN). It is now well known that in addition to its phase-shifting effect, exogenous melatonin inhibits the ®ring frequency of SCN neurons (Shibata et al., 1989; Stehle et al., 1989; Liu et al., 1997; McArthur et al., 1997), and reduces their excitability (van den Top et al., 2001). Such melatonininduced inhibition of neuronal activity of SCN cells would decrease GABA release from SCN nerve terminals in the PVN/DMH area and thereby stimulate melatonin production in the pineal. This mechanism by itself, however, could not explain why the increased melatonin peak amplitude persists after 2 days unless, under such conditions, the melatonin-induced decrease in SCN neurons excitability provokes a greater amplitude of circadian oscillations. The ability of melatonin to suppress SCN neuronal ®ring in vitro was abolished in Mel1a receptor-de®cient mice (Liu et al., 1997). This suggests that melatonin receptors are involved in the phenomena we have observed. The results from Drijfhout et al. (1999) showing an increase of the amplitude of melatonin secretion following treatment with various melatonin receptor agonists, strongly support this concept. In rodents, melatonin receptors have been found not only in the SCN but also in several other hypothalamic structures in rodents (Masson-PeÂvet et al., 1994). An effect of melatonin on these others target areas (such as the PVN) cannot be completely excluded but is unlikely as the diffusion of melatonin out of the SCN probe is known to be limited to » 1 mm around the membrane tip (Benveniste, 1989). The effect of melatonin on the endogenous peak amplitude was found only when applied at a speci®c time ± the day/night transition. This particular effect of melatonin is therefore time-dependent. The time-window in which melatonin induces an increase in endogenous melatonin peak amplitude is similar to the time-window of clock sensitivity in which rhythm entrainment or phase shifts are generally observed (Armstrong & Chesworth, 1987; Humlova & Illnerova, 1990; Pitrosky et al., 1999). Experiments are thus needed to check whether similar mechanisms are involved. Interestingly, the effect of melatonin on the amplitude magnitude was found to be of less importance under DD conditions (30% increase) than under LD conditions (100% increase). This suggests a direct effect of light on the circadian organization of SCN neuronal ®ring patterns (Meijer et al., 1998). However, as a decrease of 30% in amplitude was observed in the control group (saline in DD), this effect might simply be the consequence of the addition of the positive (melatonin) and negative (DD) effects.

In conclusion, using the transpineal microdialysis technique, we have demonstrated for the ®rst time that exogenous melatonin administered at the LD transition increases the peak amplitude of the endogenous melatonin rhythm. This effect was observed when the drug was perfused within the SCN, but not when infused within the pineal gland, indicating a direct action of the pineal hormone on the SCN by an effect on the amplitude of the clock oscillations. Our experimental set-up allowed us to demonstrate a long-lasting effect of the drug, i.e. the melatonin peak amplitude was increased for at least 2 days. Our data also show that this action of melatonin is timedependent and modulated by light/dark conditions. From a therapeutic point of view, this long-term effect on endogenous melatonin of a single melatonin application might be of interest. It is wellknown that in elderly people, circadian rhythms, including that of endogenous melatonin, are dampened. Therefore, our experimental set-up could be used in animal models to test the ef®cacy of melatonin to restore the amplitude of its endogenous rhythm.

Acknowledgements The authors are specially grateful to M. Franco for experimental help. This work was ®nancially supported by the Fondation pour la Recherche MeÂdicale, ReÂgion Alsace and by IRIS.

Abbreviations LD, light : dark cycle (normally 12 h of light : 12 h of dark); ZT, Zeitgeber Time; s.c., subcutaneous injection; SCN, suprachiasmatic nuclei.

References Armstrong, S.M. & Chesworth, M.J. (1987) Melatonin phase-shifts a mammalian circadian clock. In Trentini, G.P., De Gachini, C. & PeÂvet, P. (eds), Fundamentals and Clinics in Pineal Research. Raven Press, New York, pp. 195±198. Barassin, S., Saboureau, M., Kalsbeek, A., Bothorel, B., Vivien-Roels, B., Malan, A., Buijs, R.M., Guardiola-Lemaitre, B. & PeÂvet, P. (1999) Interindividual differences in the pattern of melatonin secretion of the Wistar rat. J. Pineal Res., 27, 193±201. Benveniste, H. (1989) Brain microdialysis. J. Neurochem., 52, 1667±1679. Cassone, V.M., Chesworth, M.J. & Armstrong, S.M. (1986) Dose-dependent entrainment of rat circadian rhythms by daily injection of melatonin. J. Biol. Rhythms, 1, 219±229. Challet, E., Malan, A. & PeÂvet, P. (1996) Daily hypocaloric feeding entrains circadian rhythms of wheel-running and body temperature in rats kept in constant darkness. Neurosci. Lett., 211, 1±4. Drijfhout, W.J., de Vries, J.B., Homan, E.J., Brons, H.F., Copinga, S., Gruppen, G., Beresford, I.J., Hagan, R.M., Grol, C.J. & Westerink, B.H. (1999) Novel non-indolic melatonin receptor agonists differentially entrain endogenous melatonin rhythm and increase its amplitude. Eur. J. Pharmacol., 382, 157±166. Drijfhout, W.J., Grol, C.J. & Westerink, B.H. (1993) Microdialysis of melatonin in the rat pineal gland: methodology and pharmacological applications. J. Neurochem., 61, 936±942. Drijfhout, W.J., Homan, E.J., Brons, H.F., Oakley, N.R., Skingle, M., Grol, C.J. & Westerink, B.H. (1996) Exogenous melatonin entrains rhythm and reduces amplitude of endogenous melatonin: an in vivo microdialysis study. J. Pineal Res., 20, 24±32. Gauer, F., Masson-PeÂvet, M., Skene, D.J., Vivien-Roels, B. & PeÂvet, P. (1993) Daily rhythms of melatonin binding sites in the rat pars tuberalis and suprachiasmatic nuclei; evidence for a regulation of melatonin receptors by melatonin itself. Neuroendocrinology, 57, 120±126. Gillette, M.U. & Mitchell, J.W. (2002) Signaling in the suprachiasmatic nucleus: selectively responsive and integrative. Cell Tissue Res. 309, 99± 107. Humlova, M. & Illnerova, H. (1990) Melatonin entrains the circadian rhythm in the rat pineal N-acetyltransferase activity. Neuroendocrinology, 52, 196± 199.


1098 B. Bothorel et al. Humlova, M. & Illnerova, H. (1992) Entrainment of the circadian rhythm in the rat pineal N-acetyltransferase activity by melatonin is photoperiod dependent. J. Pineal Res., 13, 151±157. Hunt, A.E., Al Ghoul, W.M., Gillette, M.U. & Dubocovich, M.L. (2001) Activation of MT(2) melatonin receptors in rat suprachiasmatic nucleus phase advances the circadian clock. Am. J. Physiol., 280, C110±C118. Illnerova, H. & Vanecek, J. (1982) Two-oscillator structure of the pacemaker controlling the circadian rhythm of N-acetyltransferase in the rat pineal gland. J. Comp. Physiol., 145, 539±641. Kalsbeek, A., Cutrera, R.A., Van Heerikhuize, J.J., van der Vliet, J. & Buijs, R.M. (1999) GABA release from suprachiasmatic nucleus terminals is necessary for the light-induced inhibition of nocturnal melatonin release in the rat. Neuroscience, 91, 453±461. Liu, C., Weaver, D.R., Jin, X., Shearman, L.P., Pieschl, R.L., Gribkoff, V.K. & Reppert, S.M. (1997) Molecular dissection of two distinct actions of melatonin on the suprachiasmatic circadian clock. Neuron, 19, 91±102. Masson-PeÂvet, M., George, D., Kalsbeek, A., Saboureau, M., LakhdarGhazal, N. & PeÂvet, P. (1994) An attempt to correlate brain areas containing melatonin-binding sites with rhythmic functions: a study in ®ve hibernator species. Cell Tissue Res., 278, 97±106. McArthur, A.J., Gillette, M.U. & Prosser, R.A. (1991) Melatonin directly resets the rat suprachiasmatic circadian clock in vitro. Brain Res., 565, 158± 161. McArthur, A.J., Hunt, A.E. & Gillette, M.U. (1997) Melatonin action and signal transduction in the rat suprachiasmatic circadian clock: activation of protein kinase C at dusk and dawn. Endocrinology, 138, 627±634. Meijer, J.H., Watanabe, K., Schaap, J., Albus, H. & Detari, L. (1998) Light responsiveness of the suprachiasmatic nucleus: long-term multiunit and single-unit recordings in freely moving rats. J. Neurosci., 18, 9078±9087. Miguez, J.M., Simonneaux, V. & PeÂvet, P. (1996) Changes in pineal indoleamines in rats after single melatonin injections: evidence for a diurnal sensitivity to melatonin. J. Neuroendocrinol., 8, 611±616.

Morin, L.P. (1994) The circadian visual system. Brain Res., 19, 102±127. Paxinos, G. & Watson, C. (1982) The Rat Brain in Stereotaxic Coordinates. Academic press, New York. Pitrosky, B., Kirsch, R., Malan, A., Mocaer, E. & PeÂvet, P. (1999) Organization of rat circadian rhythms during daily infusion of melatonin or S20098, a melatonin agonist. Am. J. Physiol., 277, R812±R828. Redman, J., Armstrong, S. & Ng, K.T. (1983) Free-running activity rhythms in the rat: entrainment by melatonin. Science, 219, 1089±1091. Shibata, S., Cassone, V.M. & Moore, R.Y. (1989) Effects of melatonin on neuronal activity in the rat suprachiasmatic nucleus in vitro. Neurosci. Lett., 97, 140±144. Stehle, J., Vanecek, J. & Vollrath, L. (1989) Effects of melatonin on spontaneous electrical activity of neurons in rat suprachiasmatic nuclei: an in vitro iontophoretic study. J. Neural Transm., 78, 173±177. van den Top, M., Buijs, R.M., Ruijter, M., Delagrange, P., Spanswick, D. & Hermes, M.L.H.J. (2001) Melatonin generates an outward potassium current in rat suprachiasmatic nucleus neurones in vitro independent of their circadian rhythm. Neuroscience, 107, 99±108. Turek, F.W. (1989) Effects of stimulated physical activity on the circadian pacemaker of vertebrates. J. Biol. Rhythms, 4, 135±147. Vanecek, J., Pavlik, A. & Illnerova, H. (1987) Hypothalamic melatonin receptor sites revealed by autoradiography. Brain Res., 435, 359±362. Warren, W.S., Hodges, D.B. & Cassone, V.M. (1993) Pinealectomized rats entrain and phase-shift to melatonin injections in a dose-dependent manner. J. Biol. Rhythms, 8, 233±245. Weibel, L., Rettori, M.C., Lesieur, D., Delagrange, P., Renard, P. & Van Reeth, O. (1999) A single oral dose of S 22153, a melatonin antagonist, blocks the phase advancing effects of melatonin in C3H mice. Brain Res., 829, 160±166. Zaidan, R., Geoffriau, M., Brun, J., Taillard, J., Bureau, C., Chazot, G. & Claustrat, B. (1994) Melatonin is able to in¯uence its secretion in humans: description of a phase-response curve. Neuroendocrinology, 60, 105±112.
