INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

ENTRAINMENT TO LIGHT-DARK (LD) cycles is accomplished via photic resetting of an endogenous circadian pacemaker that drives daily rhythms in behavior. Photic input to the pacemaker around dusk and dawn produces phase shifts of appropriate magnitude that permit the pacemaker to maintain a stable phase relationship to the photocycle (22). When photic input to the circadian system of entrained rodents is interrupted by transection of their retinohypothalamic tract, animals will free run despite the presence of an LD cycle (19). Animals housed in an LD cycle will also free run if they have been bilaterally enucleated or when light intensity is insufficient for the zeitgeber to gain phase control over the pacemaker (2, 28). With the exception of these special conditions, laboratory animals invariably reentrain to the LD cycle within several days, even when abrupt phase shifts of up to 12 h in duration are used (2).

Recent data from our laboratory demonstrate that Siberian hamsters (Phodopus sungorus sungorus) stably entrain to a 16:8 LD cycle (i.e., 16 h of light/24 h) but fail to reentrain after the photocycle is phase delayed by 5 h (27). Circadian body temperature (T_b) and locomotor activity rhythms never reentrained in >90% of the animals, and >80% of the hamsters free ran with stable periods [i.e., tau ()] between 24.3 and 26.3 h for several weeks or months, despite the presence of an LD cycle. Nearly 30% of these free-running animals became arrhythmic several weeks after the phase shift, and several others became arrhythmic within a few days after the phase shift (27). Only two hamsters reentrained, but they required up to 3 wk to do so. It appears that, despite stable entrainment before the phase shift, a 5-h phase delay of the LD cycle either attenuated the capacity of the circadian system to phase shift in response to photic input or disrupted photic input to the pacemaker.

The inability of these animals to reentrain to the LD cycle does not, however, preclude entrainment to nonphotic stimuli. Nonphotic cues such as food, access to running wheels, and melatonin can serve as zeitgebers in rodents (1, 6, 7, 15). Daily injections of the pineal gland hormone melatonin can entrain circadian activity rhythms of rats and juvenile golden hamsters in constant conditions (1, 6, 10). Melatonin injections most likely entrain rodents by phase shifting the hypothalamic suprachiasmatic nucleus (SCN). The SCN is a circadian pacemaker that controls the phase of behavioral rhythms via its response to photic input (22, 29). Entrainment of circadian rhythms to daily melatonin injections occurs in species that have a high density of melatonin receptors in their SCN. For example, rats readily entrain to daily melatonin injections and have high-affinity melatonin receptors in their SCN (1, 33). In contrast, activity rhythms of mink are not entrained by melatonin, and few melatonin receptors are found in their SCN (3). Melatonin injections also do not entrain circadian rhythms of pregnant golden hamsters but do synchronize rhythms of their fetuses (10); high-affinity melatonin receptors are abundant in the SCN of juvenile hamsters but absent from the adult SCN (13).

Melatonin has multiple effects on the circadian organization of Siberian hamsters. Exogenous melatonin increases the phase angle of entrainment (PAE; i.e., advances activity onset closer to dark onset) (21), whereas extirpation of the pineal gland both removes the endogenous source of plasma melatonin and decreases PAE in short day lengths; pinealectomy also slows the rate of reentrainment of activity rhythms (9). In this species, a significant proportion of animals are termed "nonresponders" because they fail to undergo gonadal regression and body mass loss normally associated with prolonged exposure to short day lengths (21, 23). In short days, nonresponders have a highly negative PAE with activity onsets occurring >5 h after dark onset (21). Daily melatonin injections administered shortly after dark onset and over the course of several weeks increased their PAE so that their activity onsets gradually advanced and became similar to those of "responders." These animals also subsequently underwent gonadal regression and body mass loss (21). It is unclear, however, whether melatonin can entrain free-running rhythms in this species as it does in rats. In the only study that has addressed this issue (9), melatonin infusions of a 10-h duration administered every 24.6 h appeared to entrain locomotor activity rhythms, although it was not clear whether true entrainment was attained or  was transiently altered by the treatment (B. D. Goldman, personal communication). Nevertheless, melatonin has pronounced effects on circadian organization, and its efficacy as a zeitgeber in Siberian hamsters may depend on the specific conditions (e.g., prior photoperiod, illumination, or mode of drug administration) under which entrainment to melatonin is tested.

The effects of melatonin on circadian organization in Siberian hamsters and the abundance of high-affinity melatonin receptors in their SCN (12) suggest that melatonin may be able to facilitate reentrainment in animals that fail to reentrain to a phase delay of the LD cycle. Alternatively, the circadian system of those animals that continued to free run in the presence of the photocycle may have a diminished sensitivity to nonphotic as well as photic zeitgebers. Of the nonphotic zeitgebers that have been extensively studied in rodents (e.g., food availability, melatonin, access to running wheels), only melatonin has been investigated in this species. Thus the goal of the present study was to determine 1) whether hamsters that failed to reentrain to a 5-h phase delay of the photocycle would entrain to daily melatonin injections, 2) whether appropriately timed melatonin injections would facilitate reentrainment to the LD cycle, and 3) whether exogenous melatonin could prevent failure to reentrain if administered before the phase shift. We report here that melatonin attenuated loss of entrainment but failed to facilitate reentrainment to the LD cycle in free-running hamsters despite the presence of two zeitgebers that were timed to provide animals with the strongest possible cues for entrainment.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Housing Conditions

Body Temperature

Melatonin Injections

Data Analysis

Presence of T_b rhythms was determined by a standard deviation-based periodogram analysis on 10-day intervals of data (11). This analysis was performed on data from both experiments that were obtained immediately before, during the last 10 days of, and beginning 5 days after termination of the injection regimen. Peaks in the periodogram were deemed statistically significant if they exceeded the the 99.9% confidence interval limit. Because Q values (i.e., power) from individual periodogram analyses are normalized to the overall variance in the data set, mean Q values ± SE can be calculated for statistically meaningful group comparisons (25). Mean T_b and T_b rhythm amplitude were determined from the same 10-day intervals of data used in the periodogram analyses. T_b amplitude was determined for individual animals by calculating the difference between the highest and lowest T_b values that were measured within each 10-day interval.

T_b rhythms were considered entrained or free running on the basis of the periodogram analysis. Because a 10-min sampling interval was used for data acquisition, periodogram analysis occasionally estimated rhythm period to be 23.83 or 24.17 h (i.e., 24.00 h ± 1 sampling interval) in entrained animals. T_b rhythms were considered entrained if the period estimate was 23.83, 24.00, or 24.17 h and if the rhythms appeared to maintain a stable phase relationship to the LD cycle. Rhythms were considered free running if periodogram analysis estimated  to be either <23.83 or >24.17 h. We refer to the time of day when an animal's T_b first increases above the daily mean T_b and remains 0.25°C above the mean for 30 min as T_b onset. T_b onset indicates the initiation of the nightly sustained elevation of T_b and coincides closely with activity onset (27).

An acute elevation in T_b after each daily injection was evident in data of individual animals. The duration and maximum T_b attained during this elevation were determined for individual animals in experiment 1 in the following manner. A mean 24-h waveform of T_b was constructed from the 20-day interval of data that began 4 days after initiation of the injection regimen. Thus each animal's waveform represents its mean T_b in 10-min intervals. Means ± SE of duration and peak were calculated for each group on the basis of waveforms of individual animals. This analysis was not performed in experiment 2 because the reentrainment process alone alters rhythm amplitude and this effect could not be distinguished from the effects of melatonin on rhythm amplitude.

Analysis of variance (ANOVA) or t-tests were used to determine differences among groups and changes in dependent variables over time (repeated measures). Dunnett's correction was applied for all independent post hoc pairwise comparisons. Differences in the percentage of animals that reentrained were tested using the ² distribution for goodness of fit, with Yates' correction for continuity for single degree of freedom comparisons. Group differences were deemed statistically significant if P < 0.05.

Experimental Protocol

Experiment 1: Melatonin and recovery of entrainment. Transmitters were implanted in 12 male hamsters that were 2-3 mo of age. Three weeks later, the LD cycle was phase shifted by extending the light phase 5 h for one photocycle. Animals remained on the 16:8 LD cycle (lights on at 0700, PST) for the remainder of the experiment. Tau was estimated for free-running animals on days 40-50 after the phase shift. Animals with similar  were paired and separated into the melatonin or vehicle treatment groups; one unpaired animal and both arrhythmic animals were added to the melatonin group. Each pair of animals was injected with melatonin or vehicle solution daily for 28 consecutive days beginning 52-63 days after phase delay of the LD cycle. This protocol permitted the injection regimen to be initiated for each hamster 2 days before the time when its T_b onset coincided with dark onset, because it maintained the phase relationship between T_b onset, dark onset, and time of injection for every animal regardless of individual differences in . T_b was recorded for 25 days after termination of the injection regimen. Data collected before the injection schedule have been reported previously (27). A portion of those data is illustrated here for comparison to data obtained during the injection regimen.

Experiment 2: Melatonin and facilitation of reentrainment. Sixteen male Siberian hamsters had transmitters implanted at 2-3 mo of age. Three weeks later, the LD cycle was phase shifted as described in experiment 1. Hamsters were injected daily with melatonin or the vehicle solution for 14 consecutive days, beginning on the day of the phase shift. The phase relationships among T_b onset, the LD cycle, and time of injection were the same as in experiment 1. T_b was recorded for 50 additional days.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Experiment 1: Melatonin and Recovery of Entrainment

T_b. Mean T_b was not significantly different among hamsters that were administered melatonin (n = 7) or vehicle solution (n = 5) [F(1,10) = 0.17, P > 0.05] and did not change significantly during the experiment [F(4,40) = 0.36, P > 0.05; Fig. 1]. T_b rhythm amplitude did not differ significantly among these two groups [F(1,10) = 0.29, P > 0.05] but was significantly lower after the phase shift during both the preinjection (t = 3.59, P < 0.05) and injection (t = 2.47, P < 0.05) intervals (Fig. 1).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Means ± SE of body temperature (Tb) (A) and Tb rhythm amplitude (B) of free-running hamsters injected with melatonin (Mel) or vehicle solution (Veh). Values were obtained before phase shift (Pre-PS) and before (Pre-Inj), during (Inj), and after (Post-Inj) injection regimen. * Both groups significantly different from Pre-PS values, P < 0.05.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Time course of acute Tb elevation (A) after injection (time 0) in hamsters administered Mel or Veh. Duration (B) is interval during which Tb remained above mean Tb, and peak (C) is magnitude of Tb increase after injection with Mel or Veh. Data are normalized to each animal's mean Tb (mean Tb = 0°C). * P < 0.05, Mel vs. Veh.

Circadian organization. All animals were stably entrained before the phase shift. Beginning several weeks after the phase shift, free-running hamsters were injected with melatonin (n = 5) or vehicle solution (n = 4; e.g., Fig. 3). Tau values did not differ significantly among these groups before initiation of the injection schedule (25.68 ± 0.34 vs. 26.00 ± 0.17 h, melatonin vs. vehicle, respectively; t = 1.84, P > 0.05). The acute T_b elevation that occurred after each injection produced a significant 24-h rhythm in all animals. During this time, 24-h periodicity was accompanied by significant non-24-h periods in all vehicle-treated hamsters and in four of five melatonin-treated hamsters that were free running;  did not differ between these two groups (25.39 ± 0.66 vs. 25.30 ± 0.25 h, melatonin vs. vehicle, respectively; t = 0.26, P > 0.05), and there was no evidence of entrainment in any of these animals. Melatonin had no effect on the temporal organization of T_b in two additional animals that were arrhythmic, but it produced acute T_b elevations after each injection similar to those observed in free-running hamsters (e.g., Fig. 3).

View larger version (57K): [in this window] [in a new window]   Fig. 3.   Double-plotted Tb from the only free-running hamster that appeared to entrain to Mel injected daily from day 73 (left-pointing triangle) to day 100 (right-pointing triangle). Tau () is indicated in periodogram plots and was estimated from bracketed intervals of 20 days. Q(p) indicates relative power of individual peaks. Only Tb values greater than animal's overall mean Tb are plotted. Light-dark (LD) cycle indicated by bars at top; PS was on day 20. Each animal's daily minimum, mean, and maximum Tb are plotted to left of its Tb, in actogram format; daily mean Tb indicated by gap in horizontal bars.

Experiment 2: Melatonin and Facilitation of Reentrainment

T_b. Mean T_b was not significantly different between hamsters injected with melatonin (n = 9) and those injected with vehicle solution (n = 7) [F(1,14) = 0.33, P > 0.05], and it did not change during the experiment [F(2,28) = 1.45, P > 0.05; Fig. 4A]. T_b rhythm amplitude was significantly reduced in both groups during the injection regimen compared with prephase shift values [F(1,13) = 24.59, P < 0.05; Fig. 4B] and was significantly different between both groups of animals only during the postinjection interval (t = 6.30, P < 0.05; Fig. 4B). On termination of the injection schedule, T_b rhythm amplitude remained significantly below prephase shift values in animals administered the vehicle solution (t = 3.33, P < 0.05) but was restored to prephase shift values in animals injected with melatonin (t = 1.17, P > 0.05; Fig. 4B).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Mean Tb ± SE (A) and Tb rhythm amplitude (B) of hamsters injected with Mel or Veh during Pre-PS, Inj, and Post-Inj intervals. Mean Tb ± SE (C) and Tb rhythm amplitude (D) for Mel-treated hamsters that reentrained or failed to reentrain to LD cycle. Note that Tb rhythm amplitude was restored to Pre-PS values only in animals that reentrained. * P < 0.05 compared with Pre-PS values applies to both groups during Inj. Absence of error bars on some symbols indicates SE smaller than symbol size.

Circadian organization. All animals were stably entrained before phase shift of the LD cycle. After the phase shift, seven of nine hamsters that were administered melatonin, but none of the animals treated with vehicle solution, appeared to entrain to the injection schedule. Two additional melatonin-treated hamsters became arrhythmic after the phase shift. When injections were terminated, a significantly greater proportion of animals that were administered melatonin (5 of 9) than of those administered vehicle solution (1 of 7) reentrained to the LD cycle (² = 16.84, P < 0.001; Fig. 5 and Fig. 6, A and B). One of these melatonin-treated hamsters initially free ran on termination of the injection regimen and reentrained to the LD cycle several weeks later when its T_b onset coincided with dark onset (Fig. 6B). Because reentrainment in this animal might not have been attributable to the melatonin treatment, a second ² test was conducted without this hamster, and a similar result was obtained (² = 7.75, P < 0.01).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Percentage of animals entrained to LD cycle Pre-PS and Post-Inj. No. of animals indicated in parentheses. Stippled bar represents total no. of animals that reentrained to LD cycle (Post-Inj); filled bar represents same group minus 1 animal that required several weeks to reentrain (i.e., hamster in Fig. 6B). * P < 0.001 (stippled bar) and P < 0.01 (filled bar) compared with Veh-treated animals.

View larger version (90K): [in this window] [in a new window]   Fig. 6.   Representative Tb data from 4 Mel-treated hamsters in experiment 2 that were injected daily from day 10 (left-pointing triangle) to day 24 (right-pointing triangle). Note that free-running rhythms (B and C) begin at termination of injections (day 24) and compression of  (A-C) during injection regimen (days 10-24);  for 1 hamster (D) was not similarly compressed. LD bars and data plotted as in Fig. 3.

View larger version (93K): [in this window] [in a new window]   Fig. 7.   Representative Tb data from 4 Veh-treated hamsters in experiment 2 that were injected daily from day 10 (left-pointing triangle) to day 24 (right-pointing triangle). Note that free-running rhythms (A and B) began on day of PS (day 10). Tb became arrhythmic in a minority of animals (e.g., C) immediately after PS, and 1 animal spontaneously reentrained (D). Missing data in B are due to temporary transmitter failure. LD bars and data plotted as in Fig. 3.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Melatonin facilitated reentrainment to the LD cycle when daily injections began on the same day as the phase shift. In contrast, melatonin failed to promote reentrainment to the LD cycle in animals that had been free running for several weeks. A similar pattern of results was obtained for entrainment to melatonin; T_b rhythms entrained to daily melatonin injections if the injections began on the day of the phase shift but not if they were initiated several weeks later. We interpret experiment 2 as evidence that T_b rhythms entrained to melatonin injections, because those rhythms began to free run only when the injection regimen was terminated, whereas T_b rhythms in vehicle-treated animals started to free run on the day of the phase shift (Figs. 6 and 7). Because hamsters were injected near the time of dark onset, however, we cannot rule out the possibility that animals administered melatonin were actually entrained to the LD cycle and that entrainment could not be maintained once the injections were terminated. Entrainment in one animal from the first experiment could not be unambiguously determined, because the free-running rhythm of this animal may have been masked by the injection regimen (Fig. 3). Nevertheless, the lack of clear entrainment in that experiment is remarkable, considering that both zeitgebers, melatonin and the LD cycle, were temporally coordinated to provide the animals with the greatest possible information for entrainment, and the lack of clear entrainment likely results from diminished sensitivity to both entraining agents; the potency of either zeitgeber presented alone under other conditions has been well documented (1, 2, 6, 10, 15, 28).

One might suggest that free-running animals in experiment 1 failed to entrain to daily melatonin injections because their  were beyond the range of entrainment for melatonin to serve as an effective zeitgeber. Most of those animals had  values close to 25.5 h, which are substantially greater than  reported for hamsters released into constant darkness (DD) from a 16:8 LD cycle (i.e., <24 h). Melatonin may have been unable to produce a phase advance sufficient to permit stable entrainment of these animals. A similar finding and explanation have been described for rats free running in constant light (LL) (6). Rats readily entrain to daily melatonin injections administered in DD, but, in LL, where  are much greater, melatonin injections failed to entrain free-running locomotor and feeding rhythms (6). This explanation does not, however, completely explain data from the present study. Some animals in experiment 2 entrained to melatonin injections that were initiated on the day of the phase shift, and they free ran with  of 25.0-25.5 h once the injection schedule was terminated (e.g., Fig. 6C). The mean  for these animals was not different from several of the animals in experiment 1 that never entrained to melatonin. Because T_b rhythms with  of 25.0-25.5 h were entrained by melatonin in experiment 2 but not in experiment 1, a range-of-entrainment argument is inadequate to explain why melatonin failed to entrain free-running animals. These data suggest that SCN melatonin receptors or signal transduction mechanisms downstream from receptors became less sensitive to melatonin after the phase shift.

Daily melatonin injections produced an acute elevation in T_b but had no effect on daily mean T_b. The elevation was likely a result of some nonspecific effect of handling the animals, because the magnitude of the T_b increase did not differ between hamsters administered melatonin or vehicle solution. The duration of hyperthermia in animals injected with melatonin was, however, more than twice that observed in control animals. This result was unexpected and is contrary to the consistent hypothermic effects of melatonin in humans (4, 5, 31). The dose of melatonin used in the present studies produces plasma melatonin concentrations well above the physiological range for this species (30); similar results might not be obtained with physiological concentrations of melatonin. Alternatively, melatonin may affect T_b differently in nocturnal rodents compared with humans. Nevertheless, prolonged hyperthermia indicates that melatonin receptors, presumably including those on SCN neurons, were activated by melatonin.

The amplitude and waveform of the T_b rhythm in animals administered melatonin were also influenced by whether animals reentrained or free ran after the phase shift. Rhythm amplitude was reduced after the phase shift but was restored to prephase shift values in melatonin-treated animals only if they reentrained to the LD cycle. The duration of T_b above the daily mean (i.e., ) was strikingly compressed during melatonin injections in experiment 2. In general,  compressed gradually during the first few days after the phase shift, was confined to the early hours of dark onset, and then decompressed once injections were terminated (Fig. 6, A and C). One animal retained a compressed  and free ran on termination of the injection regimen until the second time its T_b onset coincided with dark onset (Fig. 6B);  then decompressed, and the animal reentrained. Reentrainment of this animal to the LD cycle cannot be definitively attributed to melatonin because of the long interval between the phase shift and reentrainment. We have not, however, previously observed this phenomenon in vehicle-injected or untreated hamsters (27). In free-running golden hamsters, a transient compression of , as determined for wheel running activity and duration of melatonin secretion, is associated with photic-induced phase advances but not delays (16). Melatonin injections that phase advance circadian rhythms may compress  in a manner similar to light pulses. Daily melatonin injections would have to produce entrainment via daily phase advances, because  of animals in the present study were >24 h. Thus the effects of melatonin on  may be associated with phase advances produced by melatonin.

Some animals from both treatment groups in both experiments became arrhythmic immediately after the phase shift. Arrhythmicity in the presence of two zeitgebers that were coordinated to maximize entrainment is difficult to explain, particularly when these animals expressed robust entrainment before the phase shift. A complete loss of rhythm coherence also has been observed in Siberian hamsters that experienced the same phase shift but were not injected daily (27). Similarly, melatonin injections had no observable effect on timing of behavior in rats that were rendered arrhythmic by exposure to constant bright light (6). It appears that melatonin is not capable of restoring rhythmicity or modulating behavior patterns of arrhythmic animals. We interpret these data to mean that arrhythmicity was induced by the phase shift. The data do not support the alternative interpretation that these animals were arrhythmic before the phase shift and that their T_b rhythms were simply a passive response to the LD cycle (i.e., diurnal rhythms). If these rhythms were diurnal rather than circadian, T_b rhythms would have been present after as well as before the phase shift.

Perspectives

The suggestion that melatonin might have a functional role in entrainment is tempered by observations that pinealectomized rats and golden hamsters reentrain to phase shifts of the LD cycle and that pinealectomized Siberian hamsters do not spontaneously desynchronize from the photocycle as one would expect if melatonin was essential for entrainment (17, 24). A more conservative hypothesis suggests that disrupted melatonin secretion may be a consequence, rather than a cause, of failure to reentrain to the LD cycle. The same light exposure that can suppress melatonin may also disrupt SCN function, which could in turn disrupt melatonin secretion, because pineal melatonin production is largely controlled by the circadian system and hence the SCN (8). The fact that melatonin promotes reentrainment is not sufficient evidence to infer that it serves a functional role in the entrainment process. It is important to note, however, that this issue has been addressed in a very limited number of mammalian species, all of which reentrain to phase shifts of the LD cycle. Siberian hamsters have so far proven unique among animals in their failure to reentrain to the photocycle. Thus it remains to be determined whether the pineal gland contributes to entrainment differently in this species from other rodent species.