INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

OCTODON DEGUS, commonly called degu, a South American hystricomorph rodent, has become an increasingly popular experimental animal in recent years. Degus live in central Chile and are found from sea level to ~2,000 meters (for more data, see Refs. 3 and 8). Its way of life is terrestrial and fossorial, and it displays a very elaborate social behavior (5, 6).

From a chronobiological point of view, the degu is of great interest. Field and laboratory studies have shown that they are active during the day throughout the year, with their activity pattern characterized by two main peaks of activity in the morning and late afternoon (3, 15). Degus were more active immediately before lights-off and the 2 h before lights-on, anticipating the illumination changes. Sleep periods were mainly confined to the nocturnal period, with a minimal amount of sleep during lights-on (4). In accordance with its diurnal behavior, its temperature acrophase occurred at the end of the light period, in close association to the second activity peak (4, 14).

Most animal chronobiological studies on the chronopharmacology, behavior, and physiology of the circadian system have been developed in nocturnal rodents such as rat, hamster, or mouse. The degu's diurnalism, a rare characteristic among rodents, as well as some features of its circadian rhythms, such as the variability of individual rhythms, the presence of different morning-evening chronotypes, and the dramatic bimodal pattern in its circadian locomotor rhythm, make the degu a model of special interest in chronobiology (10). Although the multioscillatory nature of the circadian system is an old issue, the assumption that the mammalian circadian system could be based on the degree of intercommunication of several neural oscillators with different intrinsic frequencies and varying capacities for light synchronization remains open (2). The bimodality of the degu's locomotor rhythms is a useful model for obtaining further evidence to support the above-mentioned hypothesis.

To date, only circadian rhythms of locomotor activity, body temperature, and sleep have been studied (4-6, 9, 10, 12, 14, 15), our study being the first attempt to study feeding and wheel running activity rhythms (FA and WR, respectively) simultaneously in degus kept under laboratory conditions.

The aim of the present experiment was to describe the synchronization of FA and WR rhythms to different light-dark (LD) cycles to characterize the endogenous nature of the two main peaks of activity, described in the above studies. For this, the degu's FA and WR, two exclusive variables, were continuously recorded under complete and skeleton photoperiods and constant darkness (DD).

	MATERIAL AND METHODS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Animals and housing. Ten male Octodon degus (10 mo old at the beginning of the experiments) reared in a laboratory colony at the University of Alicante were used in the study. At the beginning of the experimental period, the animals were individually housed in a modified Plexiglas cage (52.5 × 27.5 × 15 cm), which allowed the recording of FA and WR. Cages were placed in a light-tight and temperature-controlled chamber with continuous ventilation (200-300 lx during lights-on, a temperature of 23 ± 1°C, and 60 ± 20% relative humidity). Dim red light (intensity <0.5 lx) was present all the time for nocturnal manipulations. A pelleted rat diet (Panlab) and water cups were available ad libitum. The cages were cleaned, and the food and water were refilled every 10 days at random times of the day to prevent synchronization.

Apparatus. All cages were provided with a contact eatometer and a wheel for feeding and locomotor registration. The eatometer has been described in detail elsewhere (11). Briefly, it consisted of a stainless steel grid with a swinging grid mounted inside that had to be activated by the animal to eat, thus allowing the recording of FA. The axis of the wheel (9-cm wide and 25-cm diameter) was provided with an eccentric cylinder that activated a microswich each time it made a complete turn. The WR and FA of each animal were recorded on a 286 microprocessor with an I/O (CIO-DIO-48, Computer Board Inc) card at 10-min intervals. FA was measured as the duration of food contacts, whereas WR was measured in revolutions.

Procedure. As Fig. 1 shows, FA and WR activities were recorded for each animal throughout 11 lighting schedules. Initially the animals were kept in continuous darkness, after which they were subjected to a skeleton photoperiod consisting of two pulses (30 min light), one at 2000 and the other at 0800. Thirty days later, the light pulse at 2000 was removed, so that the animals were exposed for another 30 days to only 30 min light per day (at 0800 local time). The light was switched off (DD conditions) for 20 days, after which the animals were kept under a 12:12-h LD cycle (LD 12:12; lights on at 0800 local time; stage LD 13 mo old) for 30 days. After another 80 days under DD, the animals were submitted again to an LD 12:12 cycle, lights-on at 0800 (stage LD 17 mo old) for 50 days. After another 40 days under DD conditions, the animals were again submitted to LD 12:12, which was expanded after 15 days by 3 h (LD 15:9) by delaying lights-off. Finally, a second expansion of the photoperiod was achieved 15 days later by advancing lights-on by 3 h (LD 18:6). The animals were kept in this condition for 15 days.

Data treatment. In each lighting condition, feeding and locomotor circadian periods were determined by ² periodogram (17). To compare feeding and locomotor rhythms the spectral power of the first 15 harmonics was calculated by Fourier analysis (Cronobio-pc software, Panlab)

Summary data were calculated from the means of averaged waveforms for individual animals in different experimental phases. To avoid the effects of transitional periods between experimental phases, the data referring to the first week of each experimental phase were discounted.

The onset and the duration of the active phase () were determined in well-defined averaged daily waveforms according to two different criteria related to the WR pattern recorded. 1) When unimodal (as occurred under DD or 1 light pulse), -onset was estimated as the time at which the first large peak of activity with an irreversible change in slope occurred, and  was estimated as the period during which the activity profile was above the mean of daily activity counts recorded within a 24-h period. Similar procedures to estimate the onset and  have been employed by other authors (1). 2) When the pattern was bimodal, -onset was defined as the first one of three or more consecutive 10-min intervals with an activity higher than the mean, whereas -offset was defined as the last of three or more consecutive periods with an activity higher than the mean.  was calculated as the total period between the onset and offset of  measured under the light phase. All calculations concerning -onset and duration were performed in hours.

In accordance with the findings of Labyak et al. (10), the animals were classified in different chronotypes. For this, the amplitude of the morning (M) or evening (E) peak was calculated as the mean of three values: the maximum and the 10-min periods before and after this maximum. All values were expressed as the percentage of the sum of the M and E peaks (100%). Animals with >66% of their activity concentrated in the M peak were considered as chronotype A, those with more than 66% of their activity related to the E peak as chronotype C, and the rest as chronotype B.

A one-way ANOVA was used to determine differences in -onset, -duration, and the amplitude of the WR peaks between treatments. Individual comparisons were made by Scheffé's test.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Figure 1 shows a diagram indicating the sequence of lighting conditions and its duration throughout the experimental period.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Schematic representation of sequence of changes in lighting conditions during whole experimental period. LD 12:12, 12:12-h light-dark cycle; DD, constant darkness.

After exposure of the animals to a skeleton photoperiod consisting of two light pulses of 30 min separated by 12 h of darkness, a bimodal pattern in WR activity emerged (Fig. 2). However, the WR activity associated with each pulse was different: 59% of the total daily activity occurred around the 2000 pulse and only 10% related to the 0800 pulse. As Table 1 shows, the activity related to the 2000 pulse started 3.54 h before the pulse and lasted for 6.22 h. During this light pulse there was a strong decrease in WR activity (masking effect) followed by a sharp increase, with a maximum appearing 30 min after lights-off. Although no anticipatory activity was detected in relation to the 0800 pulse, the activity started to increase after the pulse, reaching the maximum level 30 min afterward. Periodogram analysis revealed a significant circadian rhythm for WR in all the animals, whereas only 30% exhibited significant circadian rhythmicity for FA. Moreover, FA showed a very irregular and interindividual variable pattern.

View larger version (76K): [in this window] [in a new window]   Fig. 2.   Three representative actograms and mean waveforms (±SE; n = 10) of wheel running (WR) during exposure to 2 pulses and 1 pulse per day. Vertical dotted line in actograms and waveforms represents bright light pulses (0800 and 2000). Mean waveforms were calculated from individual waveform averaged over whole experimental period after deleting first week of each experimental phase. Data are presented as percentage of mean of overall 24 h, so that this mean equals 100%. This procedure normalizes ranges of diel patterns across subjects.

When the 2000 pulse was eliminated a progressive shift of the WR activity associated with this pulse was observed, with most WR activity occurring around the 0800 light pulse after 10-12 days (Fig. 2). During this experimental phase, the animals showing significant daily rhythms for WR presented a unimodal pattern with  starting 6.39 h before light and lasting for 8.16 h (Table 1). As occurred in the previous phase with the 2000 pulse, light seemed to exert a negative masking effect, the WR activity increasing after lights-off to reach a maximum 30 min later. Also, during this experimental phase, the FA was arrhythmic for most animals (60%), whereas a significant circadian rhythm appeared for WR in 80% of animals.

In DD conditions (Fig. 3), WR activity started to free run from two peaks in LD, but after few days, WR exhibited a unimodal pattern similar to that observed with the 0800 light pulse, but without the reactive response to light.  for WR was ~8 h, and the total daily distance run was 1,973 ± 321 m. The free-running period was shorter than 24 h, with an average  of 23.33 ± 0.23 h. In contrast, the circadian rhythmicity was not significantly detected in FA in any animals.

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Representative actograms (left) and mean waveforms (±SE; n = 10) of WR activity (top) and feeding activity (FA; bottom) during exposure to constant darkness. x-Axis: circadian time. Mean waveforms were calculated from individual waveform averaged over whole experimental period after deleting first week of each experimental phase. Data are presented as percentage of mean of overall circadian period, so that this mean equals 100%. This procedure normalizes ranges of circadian rhythms across subjects.

Under LD 12:12 a clear bimodal pattern of WR with two main peaks associated with "dawn" (M) and "dusk" (E) was recorded. As Fig. 4, top, shows, both averaged peaks were similar in amplitude and shape, irrespective of the animal's age. In 13-mo-old animals, the M peak started 1.89 h before lights-on and lasted for 1.92 h. The E peak began 0.84 h before lights-off and lasted for 2.85 h, peaking 0.5-0.66 h after lights-off (Table 2). Although these crepuscular peaks cover only 274 min of the LD cycle (19%), the total activity occurring in both peaks reached 73% of daily WR activity.

View larger version (43K): [in this window] [in a new window]   Fig. 4.   Mean waveforms of WR (top) and FA (bottom; ±SE; n = 10) under LD 12:12 at 2 different ages, 13 (left) and 17 mo (right).

After exposure to a long period of continuous darkness and returning to LD 12:12 (LD 17 mo old) ~4 mo later, the same bimodal pattern reappeared. Both the magnitude and timing of the peaks were similar to those during the previous exposure to LD (Fig. 4, Table 2). No differences were detected in the daily amount of WR (1,551 ± 312 m/day during the LD 13 mo and 1,312 ± 222 m/day during LD 17 mo).

Under LD 12:12, FA showed an irregular and highly variable daily pattern, characterized by diurnal preferences (67% of total FA occurring during the light phase). As illustrated in Fig. 4, bottom, low levels coincided with the M and E peaks of WR. In regard to the presence of diel rhythmicity, all the records for WR showed a significant diel rhythm, whereas significant diel rhythmicity for FA could only be detected in 30% of the animals in LD 13 mo and 70% in LD 17 mo. Fourier analysis confirmed the existence of differences between WR and FA. The main differences are related to the harmonics 2 and 4, corresponding to 12 and 16 h, respectively.

Despite the high stability of the times when the M and E peaks occurred, their relative amplitude differed between animals. As Fig. 5 shows, 20-30% of the animals showed a higher M peak than E peak (chronotype A), 50-60% exhibited the same magnitude in both peaks (chronotype B), whereas 10-20% showed a higher E than M (chronotype C). An individual animal's membership of a particular chronotype did not necessarily remain the same under LD 13 mo and LD 17 mo exposure (Fig. 5, bottom). Inasmuch as before the exposure to LD 12:12 animals were exposed to DD conditions, we tested the possibility that different chronotypes were related to how the degu encounters the LD cycle from the free run in DD. However, only in four of ten animals was it possible to detect a close relationship between the phase in free running and the preferential entrainment to light-to-dark transitions.

View larger version (58K): [in this window] [in a new window]   Fig. 5.   Mean waveforms (± SE; n = 10) and superimposed actogram (all records belonging to same chronotype are represented in same actogram) of WR of 13-mo-old (left) and 17-mo-old (right) degus exposed to LD 12:12. Animals were classified in chronotypes according to criterion shown above. M, morning peak; E, evening peak. Each animal is represent by a number inside the graphs. A, chronotype A; B, chronotype B; C, chronotype C. Animals with > 66% of activity concentrated in M peak were considered chronotype A, those with 66% of activity related to E peak were considered chronotype C, and the rest were considered chronotype B.

The influence of expanded photoperiods on WR is shown in Fig. 6. Under the three photoperiods a bimodal pattern appeared, the second peak of activity being greater than the first. Under LD 12:12,  lasted 13.61 h, starting 0.35 h before lights-on and ending 1.26 h after lights-off (Table 3). When the photoperiod was increased to 15 h of light by delaying the lights-off by 3 h, the animals delayed the -onset by 0.93 h and -offset by 1.47 h. The mean waveform showed a similar pattern to that obtained under LD 12:12, with a burst of activity appearing ~2 h 30 min before lights-off (about the same local time of the previous experimental phase) and a reactive peak to darkness onset (Table 3). Under LD 18:6, no immediate response was observed in most cases, although some animals advanced their -onset without any modification of the end of  (6 of 8 animals). In these lighting conditions the waveform was similar to that seen in LD 15:9. No statistically significant differences were detected in the onset, duration, and amplitude of peaks between the three photoperiods, except for the duration of M peak under LD 18:6 (Table 3).

View larger version (51K): [in this window] [in a new window]   Fig. 6.   Representative actogram (left) and mean waveform (right; ±SE, n = 10) of WR activity during exposure to a progressively lengthening photoperiod. Shaded areas in actogram indicate dark period. Dark bars in graph represent dark period of LD cycle.

During this experimental phase the daily run was 1,074 ± 230 m/day under LD 12:12, 640 ± 162 m/day in LD 15:9, and 559 ± 111 m/day in LD 18:6.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

In regard to FA and WR activity, Octodon degus should primarily be considered as a diurnal rodent species with two crepuscular peaks in its WR activity. The profiles of general motor activity, previously reported (3, 14), exhibit some differences from the results obtained in the present experiment. Obviously, a record of general motor activity includes both true locomotor activity and also the activity related to feeding behavior, so that the record is less precise and more difficult to interpret. As previously reported by Labyak and Lee (9) in female degus, WR in LD conditions showed a strong bimodal pattern characterized by the presence of two peaks anticipating the light-to-dark and dark-to-light transitions, suggesting an endogenous control for this activity. However, the precise timing and the high amplitude of both peaks suggest the existence of an additional masking effect induced by the lighting conditions.

Similar to that reported by Lee and Labyak (12), wheel running activity of male degus shows stable free-running circadian rhythms in DD characterized by a tau <24 h.

It is noteworthy that under DD and skeleton photoperiods, FA patterns are predominantly arrhythmic in most animals on a circadian basis. Furthermore, under LD, some animals did not show any statistically significant circadian rhythm in FA. This uncoupling between WR and FA was also evident under different illumination conditions, especially under DD, when all the animals displayed circadian rhythms for WR and none for FA. In those animals in which FA exhibited a significant circadian rhythm, its periodicity coincided with that observed for WR, suggesting that both variables are under the control of a common pacemaker or different coupled pacemakers. However, the loss of FA rhythmicity under DD supports the hypothesis that circadian rhythm in FA may be the result of internal masking elicited by WR. Obviously, when an animal is running in a wheel it cannot eat simultaneously, which justifies the coincidence between the peak in WR and the fall in FA under LD conditions. However, under DD, the pattern of WR was unimodal and less clearly defined and so the internal masking would be weaker than under LD, allowing that FA was arrhythmic in all animals. Examples of such a dramatic uncoupling in other species are rare (16).

Throughout the experimental period, no important age-related differences were detected in WR, probably because this species has a long life span and the duration of our experiments was not sufficient to show such an evolution.

The crepuscular character of the degu has also been observed in natural environmental conditions. However, although it has been suggested that this behavior is the consequence of its low tolerance of high daytime temperatures (3, 15), the coincidence of the records obtained under natural conditions and those obtained under laboratory conditions involving constant temperature suggests that the pattern of motor activity is strongly defined in this species and is not exclusively controlled by environmental temperatures. It should be pointed out that the dramatic bimodal pattern in locomotor activity observed under our laboratory conditions could be enhanced by the presence of wheels together with the high ambient temperature.

According to Jaeger (7), WR in the laboratory may be analogous to foraging behavior in the wild. In the degu, field studies show that this activity is predominant at dawn and dusk and may enable the animal to accumulate food to be consumed later when foraging subsides. Similar daily foraging strategies are employed by many species, including wild rats (18)

In regard to the nature of the M and E peaks of WR, it seems clear that both peaks are endogenous components entrained to LD cycle. The following remarks support this contention. 1) Under LD 12:12, both peaks showed an anticipatory character to light-to-dark and dark-to-light transitions. 2) When the animals were transferred to DD, the free-running rhythms started to run from the two peaks. 3) The angle of phase between the M and E peaks and the lights-on and -off changed with different photoperiods.

However, the shape of the peaks was modified by the masking effect of light, bright light exerting a negative masking effect on WR, whereas the onset of darkness seems to stimulate WR activity. This seems to be the case not only under LD conditions but also under skeleton photoperiods.

When we tried to analyze more deeply the origin of the two activity peaks of WR activity a few questions arose. Are the M and E peaks the output of a single clock-controlled peak and a second peak primarily the result of masking? The anticipatory activity in both peaks to light-to-dark transitions under LD 12:12 and the fact that under DD, WR started to free run from two peaks in LD suggest an endogenous origin for both peaks.

Are the M and E peaks the output of two different oscillators or, on the contrary, are they produced by one oscillator with two separate outputs? The appearance of a unimodal pattern of WR after a few days under DD and a light pulse suggests that at least two different oscillators or groups of oscillators are involved.

Are there two single oscillators or two groups of oscillators? Although it is difficult to answer this question, the presence of components associated with one peak of activity that free run until they become entrained to the other peak supports the hypothesis that it is a multioscillatory system in which the oscillators are associated in two groups. An oscillator seems to be composed of a functional network of suprachiasmatic nuclei (SCN) neurons (2). Experimental evidences of such a multioscillatory nature have been recently shown in the mammalian SCN (13).

A noteworthy characteristic of degu's WR rhythms is the wide variation in the relative amplitudes of the M and E peaks between animals, which permits their classification into three main groups: morning type (A), with >66% of their activity related to M peak, evening type (C), with > 33% of their activity related to M peak, and intermediate type (B), in which the M peak represents 33-66% of activity. Although the number of animals studied was not sufficient to allow a statistical analysis, the percentage of animals belonging to each chronotype is similar to that described by Labyak et al. (10) using body temperature rhythms and general motor activity as variables.

It is possible to hypothesize that the circadian system of the degu that is responsible for controlling WR is composed of multiple oscillators, each with its own particular period, phase angle of entrainability and phase response curve (2). Thus different oscillators would entrain preferentially to dawn or dusk, although some of them should be able to entrain indiscriminately to both transitions. When animals are exposed to DD or one bright light pulse, all the oscillators became coupled and generated a unimodal pattern. The existence of different chronotypes could be the consequence of different proportions of M and E oscillators in each individual.

In conclusion, our results show that the degu's temporal feeding strategy seems mainly arrhythmic, whereas its WR patterns appear to be driven by a strongly circadian bimodal rhythm. The bimodal nature of the WR activity of the degu can be explained by the existence of two groups of oscillators being responsible for the M and E peaks. These would be entrained to a complete LD cycle, although some of them may be able to interchange their phase angle of entrainment.

Perspectives

In addition, these results suggest new and interesting fields of study, such as 1) the different pattern of rhythmicity between WR and FA (a strong circadian rhythm in WR while, simultaneously, FA is more arrhythmic than circadian); 2) a study of chronotype stability and the response of the "morning" and "evening" peaks to light pulses could provide important evidence to support the multioscillatory nature of circadian system; and 3) finally, a diurnal rodent characterized by the existence of different chronotypes could provide an experimental model better than nocturnal rodents for applied chronobiological studies in pharmacology, psychology, and ethology.