- Email: [email protected]
Calories Affect Zeitgeber Properties of the Feeding Entrained Circadian Oscillator
Physiology & Behavior, Vol. 62, No. 5, pp. 995–1002, 1997 © 1997 Elsevier Science Inc. All rights reserved. Printed in the U.S.A. 0031-9384/97 $17.00 1 .00
S0031-9384(97)00204-7
Calories Affect Zeitgeber Properties of the Feeding Entrained Circadian Oscillator FRIEDRICH K. STEPHAN1 Neuroscience Program, Department of Psychology, Florida State University, Tallahassee, FL 32306 –1051, USA Received 5 November 1996; Accepted 7 March 1997 STEPHAN, F. K. Calories affect zeitgeber properties of the feeding entrained circadian oscillator. PHYSIOL BEHAV 62(5) 995–1002, 1997.—Rats with suprachiasmatic (SCN) lesions readily entrain to daily meals by increasing their activity prior to food access. Although previous experiments suggest that entrainment requires a nutritive meal, it is not yet clear what the parameters of the entraining stimulus are. The first experiment investigated the role of caloric content in resetting the feeding entrainable oscillator (FEO). Rats were entrained to 20 g chow/day until anticipatory activity was stable. The food access time was then delayed by 8 h and rats received either 0, 2, 6, or 16 g of chow for two days. Sixteen g of chow produced large delays on the next two cycles, while 0 and 2 g produced no delays. Two of 8 rats receiving 6 g showed delays, indicating that 22 kcal is near the threshold. In a second experiment, the effects of bulk were investigated. After an 8 h phase delay, rats received 0, 6, 10 or 16 g of chow mixed with cellulose for a total of 21 g for all groups. The 10 and 16 g chow groups delayed while the 0 g chow group did not. In the 6 g groups some animals phase shifted while others did not. Thus, the addition of non-nutritive bulk to the phase shifted meal had little or no effect on resetting the FEO and it appears that caloric content rather than gastric distention provides an effective zeitgeber for the FEO. © 1997 Elsevier Science Inc. Circadian
Feeding
Entrainment
Phase shift
Calories
supported by a study which provided a diet missing 2 of the 3 major macronutrients ad lib and one of these was made available for 1 or 2 h/day. None of the rats anticipated these meals. However, when the missing macronutrients were presented in larger amounts at two different times and were the only source of food, rats anticipated the meals regardless of macronutrient type (6). We have observed anticipatory licking in some rats to a glucose or sucrose solution when these were the only source of calories. In the presence of ad lib chow, the solutions failed to induce anticipation (2). It is also of interest that the timing of anticipation appears to be relatively insensitive to the duration of food access and presumably to the amount of food consumed (18), suggesting perhaps a threshold phenomenon of some kind. Attempts to entrain AA using water restriction indicate that water may not be an effective zeitgeber for the FEO. If chow is provided ad lib, some rats display AA but they also eat a substantial amount of chow during that time (7). When chow is restricted to the 12 h dark period and water to 1 h in the light, rats failed to show anticipatory wheel running but some displayed approach behavior to the water bin (3). Water restricted hamsters also anticipate water access but consume a substantial portion of their daily food intake during water access. Interestingly, hamsters with SCN lesions are more likely to show this AA (4). Another approach was to investigate whether induced need states would support entrainment of AA. Adrenalectomized or DOCA treated rats were placed on restricted saline access but neither treatment elicited anticipatory wheel running to NaCl (11).
BILATERAL lesion of the suprachiasmatic nuclei (SCN) abolish or severely attenuate circadian rhythms when animals are maintained in ad libitum feeding conditions (13). However, it is clear that the SCN are not the only structures that generate circadian rhythms. When meals are restricted to a particular time of day, intact rats, as well as rats with SCN lesions, begin to ‘‘anticipate’’ meal time with increased wheel running, approaches to the feeder, or unreinforced bar pressing (5). In addition, body temperature and serum corticosterone levels are among the physiologic functions that increase prior to meal time. These functions have limits of entrainment in the circadian range (i.e., food must be presented at intervals between 22 and 31 h), they free run when animals are food deprived and display transients in response to phase shifts of food access, indicating mediation by a circadian oscillator (5). Consequently, it has been proposed that the SCN represent a light-entrainable oscillator which is anatomically and functionally distinct from a feeding entrainable oscillator (FEO) which has not yet been localized (14,15). A number of studies have attempted to establish which properties of food are necessary to entrain the FEO. Nutritive substances are clearly sufficient, but perhaps not always necessary, to entrain anticipatory behavior in a wide variety of conditions. Most intact rats on ad lib chow and given a palatable and nutritive meal for 2 h/day anticipate these meals. However, when the amount of the meal was reduced to 4 g, only 2 of 13 rats anticipated and none of the rats given a palatable but non-nutritive meal showed anticipation, suggesting that both meal size and calories affect the expression of AA (8). This interpretation was 1
E-mail: [email protected]
995
996
STEPHAN
In another study, some adrenalectomized rats showed anticipatory bar pressing for a saline solution, albeit only in the presence of a light– dark cycle and when access time was unsignaled (12). Nevertheless, despite some evidence that animals can anticipate water access or other biological need states, it seems clear that nutritive meals are much more effective as zeitgebers and that, in contrast to properties of light that affect the LEO, the zeitgeber properties of meals that entrain the FEO are still in need of study. Because animals cannot be maintained on low calorie diets for prolonged periods of time, earlier studies (e.g., 8) provided ad lib ‘‘background’’ diets in order to maintain rats long enough to observe possible AA to small meals or to non-nutritive stimuli. However, it is possible that ad lib background diets reduce the probability of observing AA by reducing the saliency of the meal or other factors. By analogy, bright light pulses presented to animals in dim light are not very effective in entraining or phase shifting circadian rhythms. To circumvent this problem, a phase shift procedure was used in the present study. Rats were first entrained to a high calorie daily meal. The meal time was then phase delayed by 8 h and the caloric content of the phase shifted meal was varied. If rats consume a sufficient amount of food, this procedure reliably induces large delaying transients as the FEO seeks to re-entrain to the new food access time (16,17). Transients provide compelling evidence that the stimulus affected the clock, i.e., that it had zeitgeber properties. This was assessed for two consecutive cycles with a range of calories from 0 to about 60 kcal per meal. In a second experiment, chow was mixed with cellulose to test for potential volume effects on the FEO, i.e., those arising from gastric distention.
ambiguous cases, contacts with the water pedal were examined to confirm the onset of AA. Most rats alternated between the food and water pedals during the anticipation period. In addition, the data were examined quantitatively. The average time per 30 min spent on the food pedal over a 2 h segment, beginning 6 h before the pre-phase shift food access time, was computed as a measure of baseline activity. The onset of anticipation was based on an at least 3-fold increase (often more than a 100-fold) sustained for at least one h. While this procedure was adequate in about 70% of the cases, if visual inspection indicated a different onset, subjective judgment was used. If no clear determination could be made or in a few cases of equipment failure, the data were omitted from analysis. Statistical analysis was done using analysis of variance and post-hoc comparisons of means used a Bonferroni correction for multiple comparisons (SPSS, TM). Histology After the conclusion of experiment 1, animals were anesthetized with sodium pentobarbital and perfused with saline, followed by 10% formalin. After Experiment 2, rats were anesthetized with methoxyfluorane, blood was collected via cardiac puncture for a pilot experiment, the head was removed and stored in 10% formalin. Brains were removed and stored in sugar formalin. Frozen sections were cut through the area of the lesion at 40 m, mounted on slides and stained with cresyl violet for light microscopic examination of the damage. EXPERIMENT 1
Procedure METHOD
For both experiments, thirty adult male Sprague–Dawley rats were housed in hanging cages equipped with contact drinkometers. Bilateral electrolytic lesions were aimed at the SCN (anterior 1.3, lateral 1 0.2 and ventral 9.5 mm) by passing 1.5 mA anodal current for 15 s through a stainless steel electrode with an uninsulated tip of about 0.5 mm. At least 10 days of drinking data were examined with periodogram analysis and 16 rats without significant peaks in the circadian range were selected for each experiment. Rats were maintained in constant light (;120 lux) throughout the study to prevent light– dark transitions from serving as external time cues.
Sixteen rats were given approximately 30 g of powdered rat chow (Purina 5002 TM, mixed with 15 mL water to reduce spillage) per day and access time was limited from 1000 to 1300 for 14 days. Food trays were loaded daily after 1300. On Day 15, food access time was delayed by 8 h to 1800. Four rats each were provided with 0 g chow, 2 g chow 1 1 mL water, 6 g chow 1 3 mL water, or 16 g chow 1 8 mL water for 2 days at the new access time. Food trays were loaded just prior to 1800. The food trays for the 0 g group moved into the available condition when the other groups received food. For the next 5 days all rats were given 16 g
Apparatus Rats were transferred to plastic boxes which were covered with a gabled superstructure. Two compartments attaches to one side provided access to food or water in a stainless steel tray or a glass jar, respectively. The food and water containers were mounted on a sliding carrier, operated by air pressure under computer control. In order to eat or drink, rats had to place their front paws on a hinged pedal, about 12 cm above the cage floor. Pressure on the pedal opened a microswitch and activated a timer. These pedals were accessible whether the food containers were in the food available or unavailable position. The number of seconds of pedal contact per 10 min was stored on disk. Water was available ad lib throughout the experiment and food access time was under computer control. Data Analysis The results were printed as event records in 10 min time bins. In most cases, the onset of anticipatory activity could determined directly from the event records to the nearest 30 min, based on a sustained increase in time spent on the food pedal. In a few
FIG. 1. Average phase delays on the first and second day after an 8 h delay of meal time. n 5 7 for the 0 g group and n 5 8 for the remaining groups. SEM are also shown.
CALORIC MEALS ENTRAIN THE FEO
997
FIG. 2. Event records of two representative rats in the 0 g (left) and 2 g (right) chow groups. Arrows indicate the pre- and post-phase shift meal times. Vertical deflections indicate time in contact with the pedal in front of the food tray. Numbers to the left are the day of the year.
chow 1 8 mL water to induce delaying transients and re-entrainment in all rats. Access time was then returned to 1000 for 12 days. The rats were then again exposed to the 8 h phase delay, counterbalancing the amounts of chow provided. The results of this second phase shift were difficult to interpret (see below) and the rats were placed into hanging cages with ad lib food for 30 days. Rats were then returned to the boxes with food access from 1000 to 1300 for 14 days. The phase shift procedure was repeated, again counterbalancing food amounts. An attempt to phase shift rats a fourth time failed because of a prolonged power failure. RESULTS
Histology Light microscopic examination revealed that 4 of 16 rats did not have complete SCN lesions. In two rats the rostral tip of the
SCN was spared and in the other two, a unilateral SCN remnant was visible. The phase shift responses of these rats was compared to that of rats with complete lesions and did not appear to be different. Furthermore, examination of the event records during restricted feeding did not show a free running component. Consequently, the data for these rats were included in the analysis. Phase Shifts The main conclusions of this experiment are based on the first and third phase shift. A few rats in the 16 g chow group ate only about 14 g on the day of the phase shift but all the remaining rats consumed all the chow provided so that the phase delays were not influenced by within group differences in kcal consumed. Figure 1 shows that 0 and 2 g of chow resulted in small delays of less than 1 h on the first and second day after the phase shift. These delays
998
STEPHAN
FIG. 3. Event records of two representative rats in the 6 g (left) and 16 g (right) chow groups.
were most likely due to initiation of a free running rhythm. In the 6 g group, 6 rats were similar to the 0 and 2 g groups but two rats displayed delaying transients, suggesting that this amount might be near the threshold of having an effect on the FEO. All rats in the 16 g chow group showed large delays of 2 to 3 h per day after the phase shift. The delays in this group were significantly larger compared to the other groups (p , 0.001). The 2 and 6 g groups were not significantly different from the 0 g group. Representative event records of two rats from each group are shown in Figs. 2 and 3. Most rats in the 0, 2, or 6 g group did display delaying transients after the amount of food was increased to 16 g although some rats (e.g., T 53-7) required two additional days before they re-entrained completely to the new access time. Individual differences in the onset of anticipatory activity within
and between groups indicate that rats did not appear to be influenced by the activity of rats in other boxes. The results of the second phase shift were difficult to interpret because all rats, including those in the 16 g group, continued to show elevated activity at the pre-shift phase position. Figure 4 shows event records for two of the four rats in the 16 g chow group. Because the increase in activity typically occurred between 0700 and 1000, external noise from people, elevators, etc. could have been a factor in triggering this activity. Food trays were loaded at 0900 for four days after the phase shift and this may have resulted in some of the activity after that time. However, for most rats, increased activity began about 1 to 2 h before food was delivered. Alternatively, rats could have developed a ‘‘phase memory’’ after prolonged entrainment to a fixed access time (see DISCUSSION).
CALORIC MEALS ENTRAIN THE FEO
999 phase shift to accustom rats to its presence in the diet. On the next day, the food access time was delayed by 8 h to 1400 and four rats each received either 0 g chow 1 21 g cellulose, 6 g chow 1 15 g cellulose, 10 g chow 1 11 g cellulose, or 16 g chow 1 5 g cellulose. Thus all rats had access to a meal of 21 g moistened with 10 mL water. These meals were given for two days after which all rats received 16 g chow. The food access time remained at 1400 for 18 days. Food trays were loaded immediately prior to food access, i.e., between 1350 and 1400. A second 8 h delay was then scheduled, moving access time to 2200. Food trays were loaded between 0800 and 1000. Four rats each received the same meals as above in counterbalanced order. A third 8 h delay was introduced 20 days later moving access time to 0600. Four days prior to each phase shift, all rats received the 16 g chow 1 5 g cellulose mixture as above. Food trays were loaded after food access had ended (0900 to 0915). RESULTS
Histology Based on light microscopic examination, 11 rats appeared to have complete SCN lesions. The rostral tip of the SCN was spared in 3 rats and a lateral remnant remained in 1 rat. The lesion was too caudal in one rat sparing much of the SCN. Examination of the phase shift magnitude as well as the food intakes indicated no differences between the rats with complete and partial lesions and the event records did not show a free running rhythm in any of the rats. Consequently, the data from all animals were included in the analysis. Phase Shifts In this experiment, the results of the second phase shift (1400 to 2200) were difficult to evaluate. Because of the late access time, food was put into the food trays between 0800 and 1000. Most of the rats attempted to eat and spent a considerable time on the approach pedal. This activity overlapped with the phase shifted approach behavior so that the onset of the latter could not be assessed reliably. Consequently, the data from the first and third phase shift were analyzed. Figure 5 shows the total intake and proportion of intake esti-
FIG. 4. Event records of two rats in the 16 g chow group for the second phase shift. Note the persistence of anticipatory activity at the pre-shift phase position.
EXPERIMENT 2
Procedure Sixteen rats were placed into boxes with ad lib food for 7 days, followed by restricted access time from 0600 to 0900. Food trays were stocked immediately after food access had ended. For 12 days, rats received about 20 g of powdered chow/day. For an additional 4 days, rats received a mixture of 16 g chow 1 5 g cellulose. The cellulose was mixed with a 0.2% sodium saccharin solution in 2:1 ratio to increase palatability and 10 mL water was added to reduce spillage. The cellulose was introduced prior to the
FIG. 5. Average phase delays (bottom), total volume consumed and estimated proportion of chow (top) on the first and second day after an 8 h delay of meal time. n 5 8 for each group.
1000
STEPHAN
mated to be chow (top half) and the phase delays on the first and second day after the phase shift (bottom half). The consumption of 10 or more grams of the cellulose/saccharin mixture without chow did not produce any significant phase delays. In the 6 g group, one rat delayed by 6 h on Day 1 and two others by 2.5 h on the first day after the shift. The same rats delayed again on the second day, indicating a delaying response of the FEO. However, the group response was not significantly different from the 0 g group (p 5 0.18). The average total intake was around 15 g and the proportion of the total estimated to be chow was around 4 g. All rats in the 10 and 16 g groups showed robust delays on both days after the phase shift. Both groups were significantly different from the 0 g group (p , 0.001) but were not significantly different from each other (p 5 0.40). The proportion of the total consisting of chow consumed in the 10 g group was estimated to be about 7 g on day 1 and 8 g on day 2. Rats in the 16 g group consumed nearly all of their 21 g allocation, 16 g of which was chow. Representative event records for 1 rat in each group are shown in Fig. 6. In order to test whether the amount of chow consumed affected the magnitude of the phase delay correlation coefficients were computed. In the 6 g group the correlation was 0.57 but failed to reach significance (p 5 0.14). When the 6 and 10 g groups were combined, the correlation increased to 0.72 (p 5 0.002), suggesting that in this range of chow consumption (4 to 8 g), caloric intake from chow affects the magnitude of the delay. A comparison of the mean total intake (chow 1 cellulose) on Day 1 showed that the 16 g group consumed more than the other 3 groups (p , 0.05), which did not differ reliably from each other. On Day 2, all chow groups ingested more than the 0 g group and the 16 g group consumed more than the 6 g group (p , 0.05), but the 6 g group’s total intake was not reliably different from that of the 10 g group. Furthermore, in the 0 g chow group, the intake of non-nutritive bulk ranged from 7 to 17 g but none of the rats displayed meaningful phase delays. Together, these results indicate that the amount consumed did not affect the phase shift of anticipatory activity to a significant degree. It should also be noted that the volume of the cellulose/saccharine/water mixture was nearly twice that of powdered chow when equated for weight. DISCUSSION
The results of both experiments support previous findings that entrainment of the FEO requires a nutritive (i.e., caloric) signal and that palatable, but non-nutritive meals do not entrain the FEO (6,8). By using a phase shift procedure that made prolonged exposure to low calorie meals unnecessary, the experiments enabled us to estimate the number of calories required to affect the FEO, i.e., if a phase shifted stimulus induces transients it must necessarily be capable of entraining the underlying oscillator. In the first experiment, the amounts of food selected for the phase shifted meals were not optimal. The 2 g amount was based on the average meal size of rats at this body weight and was not expected to cause a phase delay. The 6 g amount was expected to be sufficient to phase delay the FEO, but it appeared to be near a threshold value, i.e., 2 of 8 rats delayed (the 0 g and 16 g values served as negative and positive controls, respectively). The phase shift procedure also permitted the assessment of adding nonnutritive bulk to the meals on phase shifts of anticipatory behavior. Although meal size could not be held entirely constant since rats failed to consume all of the meals provided, the ingestion of 10 g or more of a cellulose/saccharine mixture was equivalent to food deprivation, i.e., the 0 g chow group in Experiment 1, while 16 g chow 1 5 g cellulose, which was nearly totally consumed, was equivalent to 16 g chow alone in Experiment 1. Because of the
results of Experiment 1, the intermediate chow values selected for the second experiment were 6 and 10 g. In the 6 g chow group, the results were similar to those in Experiment 1 in that some, but not all rats displayed phase delays. In the 10 g chow group, all animals responded with a magnitude comparable to the 16 g chow group, although they consumed only about 8 g of chow. There was no clear indication that meal amount had an effect on the delaying transients so that distention cues arising from the gastrointestinal tract do not seem to play a role in entraining the FEO. Considered together, the results suggest that the FEO requires about 7 g of chow with a caloric value of 25.2 kcal to decelerate in response to a phase delay of the zeitgeber. Of course, this value may depend on the age, weight, fat reserves (10), time allotted to eat and other variables. In an entrainment experiment that varied access duration, AA could be entrained to access times up to 10 or 12 h. However, even with these long access times, event records indicated that rats spend a disproportionate amount of time eating during the first hour of access (18), suggesting that the first few meals entrained the FEO. Within a narrow range of intakes (4 – 8 g), there was a significant correlation between chow consumption and the magnitude of the phase delay. However, as discussed above, further increases in chow intake did no longer affect the magnitude of delays. While it is difficult to make valid comparisons between zeitgeber properties of light and food, it is nevertheless instructive that for nocturnal rodents the threshold intensity for inducing phase shifts is considerably higher than absolute visual threshold and that the visual entrainment pathway has a long time constant of integration (9). This appears to have adaptive value in that the circadian system is relatively insensitive to low light levels as well as to changes in light intensity, permitting stable entrainment despite changes in environmental illumination. A similar argument can be made for the FEO. A single small meal may not be sufficient to compensate for the calories expended in finding it and therefore should not entrain or reset the FEO to a new phase position. In both experiments, a potential problem with masking of the phase shift response was encountered. In Experiment 1, no clear cause for the failure to phase shift in response to large meals could be established. However, since the pre-shift meal time was 1000 and anticipatory behavior started around 0700 to 0800, the persistence of activity around this time may have been due to exogenous cues arising from the arrival of personnel in the building. Since different rats became active at different times, such cues were probably indistinct and different for individual rats. Alternatively, after prolonged exposure to the 1000 meal time rats could have displayed a kind of ‘‘phase memory’’ which served to override the response to the zeitgeber. In a previous experiment, rats returned to the previous phase of AA when 1 or 2 phase shifted food pulses were followed by food deprivation (17). In the second experiment, the rats clearly responded to the loading of the food trays and the resulting behavior interfered with assessment of delaying transients. To avoid such problems, it seems advisable to reload food hoppers immediately before or after the scheduled mealtime and to phase shift meal time to a period which is less likely to provide exogenous cues. At this time, it is still difficult to explain the anticipation of water when food is provided at night and water for 1 h in the light period (3). Intact rats were used in that experiment and although the SCN is not necessary to support anticipatory behavior to periodic meals, it may nevertheless be involved in timing the access to both food and/or water. A less likely possibility is that the digestion of dry laboratory chow may be significantly delayed in the absence of water so that water may trigger food metabolism sufficiently to deliver a caloric signal. When both food and water
CALORIC MEALS ENTRAIN THE FEO
1001
FIG. 6. Event records of one rat in each group. The amount of chow presented is shown in the right margin. Chow was mixed with a cellulose/saccharin mixture for a total volume of 21 g.
were restricted to 1 h/day at different times day, all rats anticipated food but failed to anticipate water, suggesting that water at best is a very weak signal to the FEO (3). While hamsters with SCN lesions anticipate restricted water access, they also consume a substantial fraction of their daily food intake during water access, complicating the interpretation of this result (4). An additional caveat arises from the results of a pilot experiment conducted prior to sacrifice of the rats used in Experiment 2. In order to assess the role of specific macronutrients, some rats received a glucose (10 g glucose 1 12 mL water) solution mixed with cellulose while others received a saccharin solution mixed with cellulose. In the glucose group, only 2 of 7 rats showed a phase delay on Day 1 and 6 of 7 rats a delay on Day 2. Unexpectedly, in the saccharin group 1 of 8 rats delayed on Day 1 and 6 of 8 rats showed a delay on Day 2. Since the activity onsets were different for individual rats, this did not appear to be due to masking. The phase shift was repeated once more with all rats receiving non-nutritive combinations of saccharin and cellulose. Three rats showed delaying transients on Day 1 and one additional rat on Day 2. However, 5 of 6 rats that
had delayed on Day 2 after the first pilot phase shift failed to do so when delayed again. There does not seem to be a straightforward explanation for these results. Considering the results of recent work showing that pairing air flow with light leads to entrainment to air flow as well as c-fos induction in the SCN by air flow, a form of conditioning may be involved (1). Repeated phase shifts with exposure to nutritive signals may lead to associations with external cues (e.g., sliding in of the food tray, noise from the air pump) or internal cues that become sufficient to substitute for a nutritive signal. On the other hand, this fails to explain why some rats failed to delay during the repeated exposure. We are currently investigating the role of macronutient specificity and hope to shed some light on these puzzling results. ACKNOWLEDGEMENTS
The author thanks Ms. Ping Li for technical assistance. This work was supported by NIH grant R01-DK50224. A preliminary version of experiment 1 was presented at the meeting of the Society for Research on Biological Rhythms, May 1996.
1002
STEPHAN REFERENCES
1. Amir, S.; Stewart, J. Resetting of the circadian clock by a conditioned stimulus. Nature 79:542–545; 1996. 2. Bays, M. E.; Stephan, F. K. Entrainment of anticipatory licking to nutritious solutions. Soc. Res. Biol. Rhythms Abst. No. 118, 82; 1992. 3. Mistlberger, R. Anticipatory activity rhythms under daily schedules of water access in the rat. J. Biol. Rhythms 7:149 –160; 1992. 4. Mistlberger, R. E. Circadian properties of anticipatory activity to restricted water access in suprachiasmatic-ablated hamsters. Am. J. Physiol. 264:R22–R29; 1993. 5. Mistlberger, R. Circadian food-anticipatory activity: formal models and physiological mechanisms. Neurosci. Biobehav. Rev. 18:1–25; 1994. 6. Mistlberger, R. E.; Houpt, T. A.; Moore-Ede, M. C. Food-anticipatory rhythms under 24-h schedules of limited access to single macronutrients. J. Biol. Rhythms 5:35– 46; 1990. 7. Mistlberger, R. E.; Rechtschaffen, A. Periodic water availability is not a potent zeitgeber for entrainment of circadian locomotor rhythms in rats. Physiol. Behav. 34:17–22; 1985. 8. Mistlberger, R.; Rusak, B. Palatable daily meals entrain anticipatory activity rhythms in free-feeding rats: dependence on meal size and nutrient content. Physiol. Behav. 41:219 –226; 1987. 9. Nelson, D. E.; Takahashi, J. S. Sensitivity and integration in a visual pathway for circadian entrainment in the hamster (Mesocricetus auratus). J. Physiol. 439:115–145; 1991.
10. Persons, J. E.; Stephan, F. K. Diet-induced obesity attenuates anticipation of food access in rats. Physiol. Behav. 54:55– 64; 1993. 11. Rosenwasser, A. M.; Schulkin, J.; Adler, N. T. Circadian wheel running activity of rats under schedules of limited daily access to salt. Chronobiol. Int. 2:115–119; 1985. 12. Rosenwasser, A. M.; Schulkin, J.; Adler, N. T. Anticipatory appetitive behavior of adrenalectomized rats under circadian salt-access schedules. Animal Learn. Behav. 16:324 –329; 1988. 13. Rusak, B.; Zucker, I. Neural regulation of circadian rhythms. Physiol. Rev. 59:449 –526; 1979. 14. Stephan, F. K. The role of period and phase in interactions between feeding- and light-entrainable circadian rhythms. Physiol. Behav. 36: 151–158; 1986. 15. Stephan, F. K. Coupling between feeding- and light-entrainable circadian pacemakers in the rat. Physiol. Behav. 38:537–546; 1986. 16. Stephan, F. K. Resetting of a feeding-entrainable circadian clock in the rat. Physiol. Behav. 52:985–995; 1992. 17. Stephan, F. K. Resetting of a circadian clock by food pulses. Physiol. Behav. 52:997–1008; 1992. 18. Stephan, F. K.; Becker, G. Entrainment of anticipatory activity to various durations of food access. Physiol. Behav. 46:731–741; 1989.
S0031-9384(97)00204-7
Calories Affect Zeitgeber Properties of the Feeding Entrained Circadian Oscillator FRIEDRICH K. STEPHAN1 Neuroscience Program, Department of Psychology, Florida State University, Tallahassee, FL 32306 –1051, USA Received 5 November 1996; Accepted 7 March 1997 STEPHAN, F. K. Calories affect zeitgeber properties of the feeding entrained circadian oscillator. PHYSIOL BEHAV 62(5) 995–1002, 1997.—Rats with suprachiasmatic (SCN) lesions readily entrain to daily meals by increasing their activity prior to food access. Although previous experiments suggest that entrainment requires a nutritive meal, it is not yet clear what the parameters of the entraining stimulus are. The first experiment investigated the role of caloric content in resetting the feeding entrainable oscillator (FEO). Rats were entrained to 20 g chow/day until anticipatory activity was stable. The food access time was then delayed by 8 h and rats received either 0, 2, 6, or 16 g of chow for two days. Sixteen g of chow produced large delays on the next two cycles, while 0 and 2 g produced no delays. Two of 8 rats receiving 6 g showed delays, indicating that 22 kcal is near the threshold. In a second experiment, the effects of bulk were investigated. After an 8 h phase delay, rats received 0, 6, 10 or 16 g of chow mixed with cellulose for a total of 21 g for all groups. The 10 and 16 g chow groups delayed while the 0 g chow group did not. In the 6 g groups some animals phase shifted while others did not. Thus, the addition of non-nutritive bulk to the phase shifted meal had little or no effect on resetting the FEO and it appears that caloric content rather than gastric distention provides an effective zeitgeber for the FEO. © 1997 Elsevier Science Inc. Circadian
Feeding
Entrainment
Phase shift
Calories
supported by a study which provided a diet missing 2 of the 3 major macronutrients ad lib and one of these was made available for 1 or 2 h/day. None of the rats anticipated these meals. However, when the missing macronutrients were presented in larger amounts at two different times and were the only source of food, rats anticipated the meals regardless of macronutrient type (6). We have observed anticipatory licking in some rats to a glucose or sucrose solution when these were the only source of calories. In the presence of ad lib chow, the solutions failed to induce anticipation (2). It is also of interest that the timing of anticipation appears to be relatively insensitive to the duration of food access and presumably to the amount of food consumed (18), suggesting perhaps a threshold phenomenon of some kind. Attempts to entrain AA using water restriction indicate that water may not be an effective zeitgeber for the FEO. If chow is provided ad lib, some rats display AA but they also eat a substantial amount of chow during that time (7). When chow is restricted to the 12 h dark period and water to 1 h in the light, rats failed to show anticipatory wheel running but some displayed approach behavior to the water bin (3). Water restricted hamsters also anticipate water access but consume a substantial portion of their daily food intake during water access. Interestingly, hamsters with SCN lesions are more likely to show this AA (4). Another approach was to investigate whether induced need states would support entrainment of AA. Adrenalectomized or DOCA treated rats were placed on restricted saline access but neither treatment elicited anticipatory wheel running to NaCl (11).
BILATERAL lesion of the suprachiasmatic nuclei (SCN) abolish or severely attenuate circadian rhythms when animals are maintained in ad libitum feeding conditions (13). However, it is clear that the SCN are not the only structures that generate circadian rhythms. When meals are restricted to a particular time of day, intact rats, as well as rats with SCN lesions, begin to ‘‘anticipate’’ meal time with increased wheel running, approaches to the feeder, or unreinforced bar pressing (5). In addition, body temperature and serum corticosterone levels are among the physiologic functions that increase prior to meal time. These functions have limits of entrainment in the circadian range (i.e., food must be presented at intervals between 22 and 31 h), they free run when animals are food deprived and display transients in response to phase shifts of food access, indicating mediation by a circadian oscillator (5). Consequently, it has been proposed that the SCN represent a light-entrainable oscillator which is anatomically and functionally distinct from a feeding entrainable oscillator (FEO) which has not yet been localized (14,15). A number of studies have attempted to establish which properties of food are necessary to entrain the FEO. Nutritive substances are clearly sufficient, but perhaps not always necessary, to entrain anticipatory behavior in a wide variety of conditions. Most intact rats on ad lib chow and given a palatable and nutritive meal for 2 h/day anticipate these meals. However, when the amount of the meal was reduced to 4 g, only 2 of 13 rats anticipated and none of the rats given a palatable but non-nutritive meal showed anticipation, suggesting that both meal size and calories affect the expression of AA (8). This interpretation was 1
E-mail: [email protected]
995
996
STEPHAN
In another study, some adrenalectomized rats showed anticipatory bar pressing for a saline solution, albeit only in the presence of a light– dark cycle and when access time was unsignaled (12). Nevertheless, despite some evidence that animals can anticipate water access or other biological need states, it seems clear that nutritive meals are much more effective as zeitgebers and that, in contrast to properties of light that affect the LEO, the zeitgeber properties of meals that entrain the FEO are still in need of study. Because animals cannot be maintained on low calorie diets for prolonged periods of time, earlier studies (e.g., 8) provided ad lib ‘‘background’’ diets in order to maintain rats long enough to observe possible AA to small meals or to non-nutritive stimuli. However, it is possible that ad lib background diets reduce the probability of observing AA by reducing the saliency of the meal or other factors. By analogy, bright light pulses presented to animals in dim light are not very effective in entraining or phase shifting circadian rhythms. To circumvent this problem, a phase shift procedure was used in the present study. Rats were first entrained to a high calorie daily meal. The meal time was then phase delayed by 8 h and the caloric content of the phase shifted meal was varied. If rats consume a sufficient amount of food, this procedure reliably induces large delaying transients as the FEO seeks to re-entrain to the new food access time (16,17). Transients provide compelling evidence that the stimulus affected the clock, i.e., that it had zeitgeber properties. This was assessed for two consecutive cycles with a range of calories from 0 to about 60 kcal per meal. In a second experiment, chow was mixed with cellulose to test for potential volume effects on the FEO, i.e., those arising from gastric distention.
ambiguous cases, contacts with the water pedal were examined to confirm the onset of AA. Most rats alternated between the food and water pedals during the anticipation period. In addition, the data were examined quantitatively. The average time per 30 min spent on the food pedal over a 2 h segment, beginning 6 h before the pre-phase shift food access time, was computed as a measure of baseline activity. The onset of anticipation was based on an at least 3-fold increase (often more than a 100-fold) sustained for at least one h. While this procedure was adequate in about 70% of the cases, if visual inspection indicated a different onset, subjective judgment was used. If no clear determination could be made or in a few cases of equipment failure, the data were omitted from analysis. Statistical analysis was done using analysis of variance and post-hoc comparisons of means used a Bonferroni correction for multiple comparisons (SPSS, TM). Histology After the conclusion of experiment 1, animals were anesthetized with sodium pentobarbital and perfused with saline, followed by 10% formalin. After Experiment 2, rats were anesthetized with methoxyfluorane, blood was collected via cardiac puncture for a pilot experiment, the head was removed and stored in 10% formalin. Brains were removed and stored in sugar formalin. Frozen sections were cut through the area of the lesion at 40 m, mounted on slides and stained with cresyl violet for light microscopic examination of the damage. EXPERIMENT 1
Procedure METHOD
For both experiments, thirty adult male Sprague–Dawley rats were housed in hanging cages equipped with contact drinkometers. Bilateral electrolytic lesions were aimed at the SCN (anterior 1.3, lateral 1 0.2 and ventral 9.5 mm) by passing 1.5 mA anodal current for 15 s through a stainless steel electrode with an uninsulated tip of about 0.5 mm. At least 10 days of drinking data were examined with periodogram analysis and 16 rats without significant peaks in the circadian range were selected for each experiment. Rats were maintained in constant light (;120 lux) throughout the study to prevent light– dark transitions from serving as external time cues.
Sixteen rats were given approximately 30 g of powdered rat chow (Purina 5002 TM, mixed with 15 mL water to reduce spillage) per day and access time was limited from 1000 to 1300 for 14 days. Food trays were loaded daily after 1300. On Day 15, food access time was delayed by 8 h to 1800. Four rats each were provided with 0 g chow, 2 g chow 1 1 mL water, 6 g chow 1 3 mL water, or 16 g chow 1 8 mL water for 2 days at the new access time. Food trays were loaded just prior to 1800. The food trays for the 0 g group moved into the available condition when the other groups received food. For the next 5 days all rats were given 16 g
Apparatus Rats were transferred to plastic boxes which were covered with a gabled superstructure. Two compartments attaches to one side provided access to food or water in a stainless steel tray or a glass jar, respectively. The food and water containers were mounted on a sliding carrier, operated by air pressure under computer control. In order to eat or drink, rats had to place their front paws on a hinged pedal, about 12 cm above the cage floor. Pressure on the pedal opened a microswitch and activated a timer. These pedals were accessible whether the food containers were in the food available or unavailable position. The number of seconds of pedal contact per 10 min was stored on disk. Water was available ad lib throughout the experiment and food access time was under computer control. Data Analysis The results were printed as event records in 10 min time bins. In most cases, the onset of anticipatory activity could determined directly from the event records to the nearest 30 min, based on a sustained increase in time spent on the food pedal. In a few
FIG. 1. Average phase delays on the first and second day after an 8 h delay of meal time. n 5 7 for the 0 g group and n 5 8 for the remaining groups. SEM are also shown.
CALORIC MEALS ENTRAIN THE FEO
997
FIG. 2. Event records of two representative rats in the 0 g (left) and 2 g (right) chow groups. Arrows indicate the pre- and post-phase shift meal times. Vertical deflections indicate time in contact with the pedal in front of the food tray. Numbers to the left are the day of the year.
chow 1 8 mL water to induce delaying transients and re-entrainment in all rats. Access time was then returned to 1000 for 12 days. The rats were then again exposed to the 8 h phase delay, counterbalancing the amounts of chow provided. The results of this second phase shift were difficult to interpret (see below) and the rats were placed into hanging cages with ad lib food for 30 days. Rats were then returned to the boxes with food access from 1000 to 1300 for 14 days. The phase shift procedure was repeated, again counterbalancing food amounts. An attempt to phase shift rats a fourth time failed because of a prolonged power failure. RESULTS
Histology Light microscopic examination revealed that 4 of 16 rats did not have complete SCN lesions. In two rats the rostral tip of the
SCN was spared and in the other two, a unilateral SCN remnant was visible. The phase shift responses of these rats was compared to that of rats with complete lesions and did not appear to be different. Furthermore, examination of the event records during restricted feeding did not show a free running component. Consequently, the data for these rats were included in the analysis. Phase Shifts The main conclusions of this experiment are based on the first and third phase shift. A few rats in the 16 g chow group ate only about 14 g on the day of the phase shift but all the remaining rats consumed all the chow provided so that the phase delays were not influenced by within group differences in kcal consumed. Figure 1 shows that 0 and 2 g of chow resulted in small delays of less than 1 h on the first and second day after the phase shift. These delays
998
STEPHAN
FIG. 3. Event records of two representative rats in the 6 g (left) and 16 g (right) chow groups.
were most likely due to initiation of a free running rhythm. In the 6 g group, 6 rats were similar to the 0 and 2 g groups but two rats displayed delaying transients, suggesting that this amount might be near the threshold of having an effect on the FEO. All rats in the 16 g chow group showed large delays of 2 to 3 h per day after the phase shift. The delays in this group were significantly larger compared to the other groups (p , 0.001). The 2 and 6 g groups were not significantly different from the 0 g group. Representative event records of two rats from each group are shown in Figs. 2 and 3. Most rats in the 0, 2, or 6 g group did display delaying transients after the amount of food was increased to 16 g although some rats (e.g., T 53-7) required two additional days before they re-entrained completely to the new access time. Individual differences in the onset of anticipatory activity within
and between groups indicate that rats did not appear to be influenced by the activity of rats in other boxes. The results of the second phase shift were difficult to interpret because all rats, including those in the 16 g group, continued to show elevated activity at the pre-shift phase position. Figure 4 shows event records for two of the four rats in the 16 g chow group. Because the increase in activity typically occurred between 0700 and 1000, external noise from people, elevators, etc. could have been a factor in triggering this activity. Food trays were loaded at 0900 for four days after the phase shift and this may have resulted in some of the activity after that time. However, for most rats, increased activity began about 1 to 2 h before food was delivered. Alternatively, rats could have developed a ‘‘phase memory’’ after prolonged entrainment to a fixed access time (see DISCUSSION).
CALORIC MEALS ENTRAIN THE FEO
999 phase shift to accustom rats to its presence in the diet. On the next day, the food access time was delayed by 8 h to 1400 and four rats each received either 0 g chow 1 21 g cellulose, 6 g chow 1 15 g cellulose, 10 g chow 1 11 g cellulose, or 16 g chow 1 5 g cellulose. Thus all rats had access to a meal of 21 g moistened with 10 mL water. These meals were given for two days after which all rats received 16 g chow. The food access time remained at 1400 for 18 days. Food trays were loaded immediately prior to food access, i.e., between 1350 and 1400. A second 8 h delay was then scheduled, moving access time to 2200. Food trays were loaded between 0800 and 1000. Four rats each received the same meals as above in counterbalanced order. A third 8 h delay was introduced 20 days later moving access time to 0600. Four days prior to each phase shift, all rats received the 16 g chow 1 5 g cellulose mixture as above. Food trays were loaded after food access had ended (0900 to 0915). RESULTS
Histology Based on light microscopic examination, 11 rats appeared to have complete SCN lesions. The rostral tip of the SCN was spared in 3 rats and a lateral remnant remained in 1 rat. The lesion was too caudal in one rat sparing much of the SCN. Examination of the phase shift magnitude as well as the food intakes indicated no differences between the rats with complete and partial lesions and the event records did not show a free running rhythm in any of the rats. Consequently, the data from all animals were included in the analysis. Phase Shifts In this experiment, the results of the second phase shift (1400 to 2200) were difficult to evaluate. Because of the late access time, food was put into the food trays between 0800 and 1000. Most of the rats attempted to eat and spent a considerable time on the approach pedal. This activity overlapped with the phase shifted approach behavior so that the onset of the latter could not be assessed reliably. Consequently, the data from the first and third phase shift were analyzed. Figure 5 shows the total intake and proportion of intake esti-
FIG. 4. Event records of two rats in the 16 g chow group for the second phase shift. Note the persistence of anticipatory activity at the pre-shift phase position.
EXPERIMENT 2
Procedure Sixteen rats were placed into boxes with ad lib food for 7 days, followed by restricted access time from 0600 to 0900. Food trays were stocked immediately after food access had ended. For 12 days, rats received about 20 g of powdered chow/day. For an additional 4 days, rats received a mixture of 16 g chow 1 5 g cellulose. The cellulose was mixed with a 0.2% sodium saccharin solution in 2:1 ratio to increase palatability and 10 mL water was added to reduce spillage. The cellulose was introduced prior to the
FIG. 5. Average phase delays (bottom), total volume consumed and estimated proportion of chow (top) on the first and second day after an 8 h delay of meal time. n 5 8 for each group.
1000
STEPHAN
mated to be chow (top half) and the phase delays on the first and second day after the phase shift (bottom half). The consumption of 10 or more grams of the cellulose/saccharin mixture without chow did not produce any significant phase delays. In the 6 g group, one rat delayed by 6 h on Day 1 and two others by 2.5 h on the first day after the shift. The same rats delayed again on the second day, indicating a delaying response of the FEO. However, the group response was not significantly different from the 0 g group (p 5 0.18). The average total intake was around 15 g and the proportion of the total estimated to be chow was around 4 g. All rats in the 10 and 16 g groups showed robust delays on both days after the phase shift. Both groups were significantly different from the 0 g group (p , 0.001) but were not significantly different from each other (p 5 0.40). The proportion of the total consisting of chow consumed in the 10 g group was estimated to be about 7 g on day 1 and 8 g on day 2. Rats in the 16 g group consumed nearly all of their 21 g allocation, 16 g of which was chow. Representative event records for 1 rat in each group are shown in Fig. 6. In order to test whether the amount of chow consumed affected the magnitude of the phase delay correlation coefficients were computed. In the 6 g group the correlation was 0.57 but failed to reach significance (p 5 0.14). When the 6 and 10 g groups were combined, the correlation increased to 0.72 (p 5 0.002), suggesting that in this range of chow consumption (4 to 8 g), caloric intake from chow affects the magnitude of the delay. A comparison of the mean total intake (chow 1 cellulose) on Day 1 showed that the 16 g group consumed more than the other 3 groups (p , 0.05), which did not differ reliably from each other. On Day 2, all chow groups ingested more than the 0 g group and the 16 g group consumed more than the 6 g group (p , 0.05), but the 6 g group’s total intake was not reliably different from that of the 10 g group. Furthermore, in the 0 g chow group, the intake of non-nutritive bulk ranged from 7 to 17 g but none of the rats displayed meaningful phase delays. Together, these results indicate that the amount consumed did not affect the phase shift of anticipatory activity to a significant degree. It should also be noted that the volume of the cellulose/saccharine/water mixture was nearly twice that of powdered chow when equated for weight. DISCUSSION
The results of both experiments support previous findings that entrainment of the FEO requires a nutritive (i.e., caloric) signal and that palatable, but non-nutritive meals do not entrain the FEO (6,8). By using a phase shift procedure that made prolonged exposure to low calorie meals unnecessary, the experiments enabled us to estimate the number of calories required to affect the FEO, i.e., if a phase shifted stimulus induces transients it must necessarily be capable of entraining the underlying oscillator. In the first experiment, the amounts of food selected for the phase shifted meals were not optimal. The 2 g amount was based on the average meal size of rats at this body weight and was not expected to cause a phase delay. The 6 g amount was expected to be sufficient to phase delay the FEO, but it appeared to be near a threshold value, i.e., 2 of 8 rats delayed (the 0 g and 16 g values served as negative and positive controls, respectively). The phase shift procedure also permitted the assessment of adding nonnutritive bulk to the meals on phase shifts of anticipatory behavior. Although meal size could not be held entirely constant since rats failed to consume all of the meals provided, the ingestion of 10 g or more of a cellulose/saccharine mixture was equivalent to food deprivation, i.e., the 0 g chow group in Experiment 1, while 16 g chow 1 5 g cellulose, which was nearly totally consumed, was equivalent to 16 g chow alone in Experiment 1. Because of the
results of Experiment 1, the intermediate chow values selected for the second experiment were 6 and 10 g. In the 6 g chow group, the results were similar to those in Experiment 1 in that some, but not all rats displayed phase delays. In the 10 g chow group, all animals responded with a magnitude comparable to the 16 g chow group, although they consumed only about 8 g of chow. There was no clear indication that meal amount had an effect on the delaying transients so that distention cues arising from the gastrointestinal tract do not seem to play a role in entraining the FEO. Considered together, the results suggest that the FEO requires about 7 g of chow with a caloric value of 25.2 kcal to decelerate in response to a phase delay of the zeitgeber. Of course, this value may depend on the age, weight, fat reserves (10), time allotted to eat and other variables. In an entrainment experiment that varied access duration, AA could be entrained to access times up to 10 or 12 h. However, even with these long access times, event records indicated that rats spend a disproportionate amount of time eating during the first hour of access (18), suggesting that the first few meals entrained the FEO. Within a narrow range of intakes (4 – 8 g), there was a significant correlation between chow consumption and the magnitude of the phase delay. However, as discussed above, further increases in chow intake did no longer affect the magnitude of delays. While it is difficult to make valid comparisons between zeitgeber properties of light and food, it is nevertheless instructive that for nocturnal rodents the threshold intensity for inducing phase shifts is considerably higher than absolute visual threshold and that the visual entrainment pathway has a long time constant of integration (9). This appears to have adaptive value in that the circadian system is relatively insensitive to low light levels as well as to changes in light intensity, permitting stable entrainment despite changes in environmental illumination. A similar argument can be made for the FEO. A single small meal may not be sufficient to compensate for the calories expended in finding it and therefore should not entrain or reset the FEO to a new phase position. In both experiments, a potential problem with masking of the phase shift response was encountered. In Experiment 1, no clear cause for the failure to phase shift in response to large meals could be established. However, since the pre-shift meal time was 1000 and anticipatory behavior started around 0700 to 0800, the persistence of activity around this time may have been due to exogenous cues arising from the arrival of personnel in the building. Since different rats became active at different times, such cues were probably indistinct and different for individual rats. Alternatively, after prolonged exposure to the 1000 meal time rats could have displayed a kind of ‘‘phase memory’’ which served to override the response to the zeitgeber. In a previous experiment, rats returned to the previous phase of AA when 1 or 2 phase shifted food pulses were followed by food deprivation (17). In the second experiment, the rats clearly responded to the loading of the food trays and the resulting behavior interfered with assessment of delaying transients. To avoid such problems, it seems advisable to reload food hoppers immediately before or after the scheduled mealtime and to phase shift meal time to a period which is less likely to provide exogenous cues. At this time, it is still difficult to explain the anticipation of water when food is provided at night and water for 1 h in the light period (3). Intact rats were used in that experiment and although the SCN is not necessary to support anticipatory behavior to periodic meals, it may nevertheless be involved in timing the access to both food and/or water. A less likely possibility is that the digestion of dry laboratory chow may be significantly delayed in the absence of water so that water may trigger food metabolism sufficiently to deliver a caloric signal. When both food and water
CALORIC MEALS ENTRAIN THE FEO
1001
FIG. 6. Event records of one rat in each group. The amount of chow presented is shown in the right margin. Chow was mixed with a cellulose/saccharin mixture for a total volume of 21 g.
were restricted to 1 h/day at different times day, all rats anticipated food but failed to anticipate water, suggesting that water at best is a very weak signal to the FEO (3). While hamsters with SCN lesions anticipate restricted water access, they also consume a substantial fraction of their daily food intake during water access, complicating the interpretation of this result (4). An additional caveat arises from the results of a pilot experiment conducted prior to sacrifice of the rats used in Experiment 2. In order to assess the role of specific macronutrients, some rats received a glucose (10 g glucose 1 12 mL water) solution mixed with cellulose while others received a saccharin solution mixed with cellulose. In the glucose group, only 2 of 7 rats showed a phase delay on Day 1 and 6 of 7 rats a delay on Day 2. Unexpectedly, in the saccharin group 1 of 8 rats delayed on Day 1 and 6 of 8 rats showed a delay on Day 2. Since the activity onsets were different for individual rats, this did not appear to be due to masking. The phase shift was repeated once more with all rats receiving non-nutritive combinations of saccharin and cellulose. Three rats showed delaying transients on Day 1 and one additional rat on Day 2. However, 5 of 6 rats that
had delayed on Day 2 after the first pilot phase shift failed to do so when delayed again. There does not seem to be a straightforward explanation for these results. Considering the results of recent work showing that pairing air flow with light leads to entrainment to air flow as well as c-fos induction in the SCN by air flow, a form of conditioning may be involved (1). Repeated phase shifts with exposure to nutritive signals may lead to associations with external cues (e.g., sliding in of the food tray, noise from the air pump) or internal cues that become sufficient to substitute for a nutritive signal. On the other hand, this fails to explain why some rats failed to delay during the repeated exposure. We are currently investigating the role of macronutient specificity and hope to shed some light on these puzzling results. ACKNOWLEDGEMENTS
The author thanks Ms. Ping Li for technical assistance. This work was supported by NIH grant R01-DK50224. A preliminary version of experiment 1 was presented at the meeting of the Society for Research on Biological Rhythms, May 1996.
1002
STEPHAN REFERENCES
1. Amir, S.; Stewart, J. Resetting of the circadian clock by a conditioned stimulus. Nature 79:542–545; 1996. 2. Bays, M. E.; Stephan, F. K. Entrainment of anticipatory licking to nutritious solutions. Soc. Res. Biol. Rhythms Abst. No. 118, 82; 1992. 3. Mistlberger, R. Anticipatory activity rhythms under daily schedules of water access in the rat. J. Biol. Rhythms 7:149 –160; 1992. 4. Mistlberger, R. E. Circadian properties of anticipatory activity to restricted water access in suprachiasmatic-ablated hamsters. Am. J. Physiol. 264:R22–R29; 1993. 5. Mistlberger, R. Circadian food-anticipatory activity: formal models and physiological mechanisms. Neurosci. Biobehav. Rev. 18:1–25; 1994. 6. Mistlberger, R. E.; Houpt, T. A.; Moore-Ede, M. C. Food-anticipatory rhythms under 24-h schedules of limited access to single macronutrients. J. Biol. Rhythms 5:35– 46; 1990. 7. Mistlberger, R. E.; Rechtschaffen, A. Periodic water availability is not a potent zeitgeber for entrainment of circadian locomotor rhythms in rats. Physiol. Behav. 34:17–22; 1985. 8. Mistlberger, R.; Rusak, B. Palatable daily meals entrain anticipatory activity rhythms in free-feeding rats: dependence on meal size and nutrient content. Physiol. Behav. 41:219 –226; 1987. 9. Nelson, D. E.; Takahashi, J. S. Sensitivity and integration in a visual pathway for circadian entrainment in the hamster (Mesocricetus auratus). J. Physiol. 439:115–145; 1991.
10. Persons, J. E.; Stephan, F. K. Diet-induced obesity attenuates anticipation of food access in rats. Physiol. Behav. 54:55– 64; 1993. 11. Rosenwasser, A. M.; Schulkin, J.; Adler, N. T. Circadian wheel running activity of rats under schedules of limited daily access to salt. Chronobiol. Int. 2:115–119; 1985. 12. Rosenwasser, A. M.; Schulkin, J.; Adler, N. T. Anticipatory appetitive behavior of adrenalectomized rats under circadian salt-access schedules. Animal Learn. Behav. 16:324 –329; 1988. 13. Rusak, B.; Zucker, I. Neural regulation of circadian rhythms. Physiol. Rev. 59:449 –526; 1979. 14. Stephan, F. K. The role of period and phase in interactions between feeding- and light-entrainable circadian rhythms. Physiol. Behav. 36: 151–158; 1986. 15. Stephan, F. K. Coupling between feeding- and light-entrainable circadian pacemakers in the rat. Physiol. Behav. 38:537–546; 1986. 16. Stephan, F. K. Resetting of a feeding-entrainable circadian clock in the rat. Physiol. Behav. 52:985–995; 1992. 17. Stephan, F. K. Resetting of a circadian clock by food pulses. Physiol. Behav. 52:997–1008; 1992. 18. Stephan, F. K.; Becker, G. Entrainment of anticipatory activity to various durations of food access. Physiol. Behav. 46:731–741; 1989.