Behavioral Inhibition of Circadian
Responses Martin R.
to
Ralph*
Light and N.
,† Mrosovsky*
Departments of *Psychology, †Zoology, and Physiology, University of Toronto, Toronto, Ontario M5S 1A1, Canada Abstract Circadian locomotor rhythms in rodents may be synchronized by either photic or nonphotic events that produce phase shifts of the rhythm. Little is known, however, about how these two types of stimuli interact to produce entrainment. The well-characterized circadian photic response of the golden hamster was examined in situations where a short light pulse and locomotor activity, a nonphotic event, occurred simultaneously. Light-induced phase advances were attenuated when animals were active during light exposure. The results show that circadian responses to light depend upon the environmental situation in which the light is given, and call into question the implicit assumption in circadian rhythm research that phase shifting and entrainment to light-dark cycles depend simply on photic activation of well-known retinofugal pathways. Moreover, since light therapy is becoming an important component in the treatment of circadian-based disorders in humans, the results emphasize the need for evaluation of the behavioral aspects of light therapy protocols.
Key
words
phase shift, PRC, activity, behavior, arousal, motivation, nonphotic,
hamster
One of the defining characteristics of circadian systems is the ability to synchronize with cyclic environmental phenomena. This ability to be reset enables the endogenous oscillator to function as a clock that measures local time (Pittendrigh, 1960). For most organisms, the primary synchronizer (zeitgeber) is the daily light-dark (LD) cycle, and until recently, synchronization was thought to occur almost exclusively through daily light-induced adjustments (phase shifts) of the endogenous circadian rhythm (Pittendrigh, 1960, 1965; DeCoursey, 1964). It has been shown, however, that nonphotic events, particularly those that elicit behavioral arousal, can be as effective as light in producing phase shifts (Reebs and Mrosovsky, 1989a). Nonphotic stimuli may serve in entrainment as zeitgebers (Menaker and Eskin, 1966; Ehlers et al., 1988; Rusak et al., 1988; Edgar and Dement, 1991), or as signals that modify entrainment to LD cycles (Reebs and Mrosovsky, 1989b; Honrado and Mrosovsky, 1991; Mistlberger, 1991). Little is known, however, about the mechanisms or the points in the neurochemical pathways to the clock where responses to zeitgebers interact during the synchronization of rhythms with environmental cycles. In mammals, the neural circuitry between specialized photoreceptors in the eye and a circadian pacemaker in the suprachiasmatic nucleus (SCN) has been described in detail (Rusak and Zucker, 1979; Pickard, 1982). In addition, neurons that produce sustained
Correspondence should be addressed to either author at Department of Psychology, University of Toronto, George St., Toronto, Ontario M5S 1 A 1, Canada.
100 St.
353 Downloaded from jbr.sagepub.com at RMIT UNIVERSITY on July 9, 2015
responses to illumination have been found in the SCN and intergeniculate leaflet, two retinorecipient nuclei of vertebrate circadian systems (Meijer and Rietveld, 1989), and it has recently been shown that proto-oncogene expression is stimulated in these areas by brief exposure to light (Rea, 1989; Aronin et al., 1990; Komhauser et al., 1990; Rusak et al., 1990). An implicit assumption deriving from this body of work is that if the appropriate retinofugal pathways and their targets are activated at the right time of day, then phase shifts and entrainment will occur. Circadian systems, however, are responsive to both photic and nonphotic stimulation. Responses, measured as phase shifts of the free-running rhythm in a constant environment, can be elicited at different phases of the cycle. Responses to light occur during the subjective night (when the animal behaves as though it is night), whereas large, nonphotic responses occur mainly during the subjective day. Nonphotic responses in rodents, elicited by novel situations (Reebs and Mrosovsky, 1989a), by dark pulses (Boulos and Rusak, 1982), or by administration of the benzodiazepine triazolam (Turek and Losee-Olsen, 1986), require behavioral activation (Reebs et al., 1989; Van Reeth and Turek, 1989; Mrosovsky and Salmon, 1990) or a correlated central process. The capability of nonphotic cues to contribute to entrainment is evident from results showing that the rate of re-entrainment to shifted LD cycles is greatly increased by an appropriately timed pulse of locomotor activity (Mrosovsky and Salmon, 1987). The present experiments were prompted by two recent reports showing that (1) in hamsters held on short-day photoperiods, testes regression, which is normally blocked by the administration of light pulses during the night, will still occur if nighttime light pulses are self-administered in a feedback schedule requiring locomotor activity (Ferraro et al., 1990); (2) in scorpions, the circadian phase response curve (PRC) to light is typical of photic PRCs, provided that the animals remain relatively quiescent during the light pulse. Individuals whose acute response to the pulse is behavioral activation exhibit phase shifts more typical of nonphotic responses (Hohmann et al., 1990; cf. Fig. 6 in Mrosovsky et al., 1989; data from W. Hohmann and G. Fleissner). Taken together, these findings raised the possibility that a change in activity during a light pulse may influence how the circadian system responds to light. We report here that phase advance responses to light in hamsters are attenuated if simultaneous locomotor activity is elicited by placing the animal in a novel running wheel during the exposure to light, suggesting that lightinduced phase shifts are influenced by the environmental and behavioral situation in which the light pulse is given.
MATERIALS AND METHODS Golden hamsters (Mesocricetus auratus) were housed individually in running wheel cages kept in constant darkness (DD). Animals were tested at a circadian time (CT) when short light pulses usually cause large phase advances of the rhythm (CT 17-23, 5 to 11hr after the onset of activity, which is defined as CT 12 for nocturnal species). For the first experiment, animals were males obtained from Charles River, Quebec (CRL, Lak:LVG). To test for generalizability of the phenomenon, the experiment was then repeated in another laboratory, using animals obtained from Harlan Sprague-Dawley, IN (HsD: SYR).
354 Downloaded from jbr.sagepub.com at RMIT UNIVERSITY on July 9, 2015
EXPERIMENT I
Animals were 12 weeks old at the time of the light pulse and were housed individually in translucent polypropylene running-wheel cages for activity recording (Dataquest III, MiniMitter Co., Inc., Sunriver, OR). All were maintained in a 24-hour LD cycle comprising 14 hr of light and 10 hr of dark (LD 14: 10) before being placed into DD. After free-running for 10 days in DD, animals were given an activity pulse, a light pulse, or a simultaneous pulse of both. Experimental animals were placed into novel running wheels mounted in clear Plexiglas boxes (Reebs and Mrosovsky, 1989b), and allowed to habituate for 30 min prior to being given a 15-min pulse of light (8-40 lux); they were removed immediately thereafter. All such manipulations were made in the dark with the aid of an infrared viewing scope (FJW Industries, NJ). A light-only control group was given the light pulse while in the home cage without any additional disturbance. A second group of control animals was placed in the novel wheels for 45 min but not given the light pulse. While in the novel situation, animals could not exit from the wheel. The light pulse was delivered by turning on the room
lights. Eye-fitted lines through successive activity onsets for 7 days prior to the pulse and for days 5-12 following the pulse were used to project activity onsets on the day after the pulse. Phase shifts were determined as the difference in projected onset times on this day. The data for the first 4 days after the pulse were omitted to avoid complications due to transient changes in behavior. Animals were categorized as nonrunners if they failed to average 10 revolutions per minute during the light pulse (150 total). EXPERIMENT 2 Hamsters in the second experiment were obtained from Harlan Sprague-Dawley and were 15 weeks of age at the time of the light pulse. They were housed individually in runningwheel cages for activity recording (Dataquest III). After 6 days in LD 14: 10, animals were placed into DD and allowed to free-run for 8 days before being given a 20-min pulse of room light (ca. 10-20 lux). Ten animals remained in their home cages, and five were placed in novel wheels 30 min before the pulse and were removed immediately after. Nonrunners were those that failed to average 10 wheel revolutions per minute during the light pulse (200 total). Phase shifts were based on the same pre- and postpulse days used in Experiment 1, but were calculated using regression lines through activity onsets (Reebs and Mrosovsky, 1989a,b). Onsets were objectively defined as the first 10-min interval with 55 or more wheel revolutions, that was followed by another such interval within 30 min. The data for one animal were discarded because of erratic activity onsets. Ambient temperature for both
experiments
was
21°
±
2°C.
RESULTS In control animals that were not placed in novel wheels, light pulses produced the expected phase advances between CT 17 and CT 23. Animals that were induced to run in the novel wheels, however, showed smaller phase shifts (Figs. 1, 2). Attenuation of the photic response was obtained with either CRL (Experiment 1) or HsD (Experiment 2) hamsters, though the HsD animals showed greater responses to light, possibly due to the slightly longer light pulse given to these animals. Examples of individual actograms are shown in Figure 3.
355 Downloaded from jbr.sagepub.com at RMIT UNIVERSITY on July 9, 2015
FIGURE 1. Effect of stimulated activity on light-induced phase advances, Experiment 1. Animals were given light pulses in their home cages (triangles) or while confined to novel wheels (circles). Filled symbols represent individuals that were active during the pulse (> 150 wheel revolutions). Open symbols represent individuals that did not run (< 150 revolutions). Circadian time refers to the start of the light pulse. Phase shifts are shown in circadian hours (Ch), which were calculated for individual animals as the period (before the pulse) divided by 24. Histograms show the means and SEM for runners and nonrunners. The difference was significant ( p < 0.001, Student’s two-tailed t test).
Some individuals that were placed in the novel wheels in Experiment 1 exhibited very little running while the light was on, although most ran in the dark before the pulse. Statistical comparison of runners versus nonrunners in the novel wheels revealed a significant reduction in the light-induced phase shift (runners, d4>avg = 0.34 ± 0.13 [SEM] hr, n 8; nonrunners, 1.01± 0.21 [SEM] hr, n 5 ; p < 0.01, two-tailedt test). Nonrunners in novel d4>avg wheels responded with phase shifts within the range of the light-only (home cage) controls. In the home cage, most animals remained relatively inactive during the light pulse. However, out of all the hamsters in the two experiments, there were three animals that did run in their home cages at this time. Although this number is too small for statistical evaluation, it should be noted that these individuals responded with relatively small phase shifts (filled triangles in Figs. 1 and 2). Taking the significant difference between runners and nonrunners in the novel wheels together with the similar tendency in the home cage, we conclude that phase shifting to light was attenuated in the active animals. It is unlikely that wheel running simply delayed the clock to a phase in which light was less effective, because the second group of control =
=
=
FIGURE 2. Effect of stimulated activity on light-induced phase advances, Experiment 2. See Figure 1 for explanation of symbols. 200 revolutions (10/min) was used as the criterion for running, since this experiment used 20-min light pulses. The difference between the means (histogram) was significant (p < 0.005, Student’s two-tailedt test). 356 Downloaded from jbr.sagepub.com at RMIT UNIVERSITY on July 9, 2015
FIGURE 3. Examples of actograms from Experiment 1. A, B: Effects of control light pulses (no wheel running) administered at CT 18.1(shift + 1.0 hr) and CT 18.4 (shift + 1.1hr), respectively. C, D: Effects of simultaneous wheel running during light pulses at CT 18.1 (shift = 0.0 hr) and CT 18.3 (shift +0.3 hr), respectively. In each example, the light pulse was delivered on the day indicated by a horizontal arrow, at the time indicated by the star. =
=
=
in novel wheels for 45 min without receiving a phase shifts (A
Responses Martin R.
to
Ralph*
Light and N.
,† Mrosovsky*
Departments of *Psychology, †Zoology, and Physiology, University of Toronto, Toronto, Ontario M5S 1A1, Canada Abstract Circadian locomotor rhythms in rodents may be synchronized by either photic or nonphotic events that produce phase shifts of the rhythm. Little is known, however, about how these two types of stimuli interact to produce entrainment. The well-characterized circadian photic response of the golden hamster was examined in situations where a short light pulse and locomotor activity, a nonphotic event, occurred simultaneously. Light-induced phase advances were attenuated when animals were active during light exposure. The results show that circadian responses to light depend upon the environmental situation in which the light is given, and call into question the implicit assumption in circadian rhythm research that phase shifting and entrainment to light-dark cycles depend simply on photic activation of well-known retinofugal pathways. Moreover, since light therapy is becoming an important component in the treatment of circadian-based disorders in humans, the results emphasize the need for evaluation of the behavioral aspects of light therapy protocols.
Key
words
phase shift, PRC, activity, behavior, arousal, motivation, nonphotic,
hamster
One of the defining characteristics of circadian systems is the ability to synchronize with cyclic environmental phenomena. This ability to be reset enables the endogenous oscillator to function as a clock that measures local time (Pittendrigh, 1960). For most organisms, the primary synchronizer (zeitgeber) is the daily light-dark (LD) cycle, and until recently, synchronization was thought to occur almost exclusively through daily light-induced adjustments (phase shifts) of the endogenous circadian rhythm (Pittendrigh, 1960, 1965; DeCoursey, 1964). It has been shown, however, that nonphotic events, particularly those that elicit behavioral arousal, can be as effective as light in producing phase shifts (Reebs and Mrosovsky, 1989a). Nonphotic stimuli may serve in entrainment as zeitgebers (Menaker and Eskin, 1966; Ehlers et al., 1988; Rusak et al., 1988; Edgar and Dement, 1991), or as signals that modify entrainment to LD cycles (Reebs and Mrosovsky, 1989b; Honrado and Mrosovsky, 1991; Mistlberger, 1991). Little is known, however, about the mechanisms or the points in the neurochemical pathways to the clock where responses to zeitgebers interact during the synchronization of rhythms with environmental cycles. In mammals, the neural circuitry between specialized photoreceptors in the eye and a circadian pacemaker in the suprachiasmatic nucleus (SCN) has been described in detail (Rusak and Zucker, 1979; Pickard, 1982). In addition, neurons that produce sustained
Correspondence should be addressed to either author at Department of Psychology, University of Toronto, George St., Toronto, Ontario M5S 1 A 1, Canada.
100 St.
353 Downloaded from jbr.sagepub.com at RMIT UNIVERSITY on July 9, 2015
responses to illumination have been found in the SCN and intergeniculate leaflet, two retinorecipient nuclei of vertebrate circadian systems (Meijer and Rietveld, 1989), and it has recently been shown that proto-oncogene expression is stimulated in these areas by brief exposure to light (Rea, 1989; Aronin et al., 1990; Komhauser et al., 1990; Rusak et al., 1990). An implicit assumption deriving from this body of work is that if the appropriate retinofugal pathways and their targets are activated at the right time of day, then phase shifts and entrainment will occur. Circadian systems, however, are responsive to both photic and nonphotic stimulation. Responses, measured as phase shifts of the free-running rhythm in a constant environment, can be elicited at different phases of the cycle. Responses to light occur during the subjective night (when the animal behaves as though it is night), whereas large, nonphotic responses occur mainly during the subjective day. Nonphotic responses in rodents, elicited by novel situations (Reebs and Mrosovsky, 1989a), by dark pulses (Boulos and Rusak, 1982), or by administration of the benzodiazepine triazolam (Turek and Losee-Olsen, 1986), require behavioral activation (Reebs et al., 1989; Van Reeth and Turek, 1989; Mrosovsky and Salmon, 1990) or a correlated central process. The capability of nonphotic cues to contribute to entrainment is evident from results showing that the rate of re-entrainment to shifted LD cycles is greatly increased by an appropriately timed pulse of locomotor activity (Mrosovsky and Salmon, 1987). The present experiments were prompted by two recent reports showing that (1) in hamsters held on short-day photoperiods, testes regression, which is normally blocked by the administration of light pulses during the night, will still occur if nighttime light pulses are self-administered in a feedback schedule requiring locomotor activity (Ferraro et al., 1990); (2) in scorpions, the circadian phase response curve (PRC) to light is typical of photic PRCs, provided that the animals remain relatively quiescent during the light pulse. Individuals whose acute response to the pulse is behavioral activation exhibit phase shifts more typical of nonphotic responses (Hohmann et al., 1990; cf. Fig. 6 in Mrosovsky et al., 1989; data from W. Hohmann and G. Fleissner). Taken together, these findings raised the possibility that a change in activity during a light pulse may influence how the circadian system responds to light. We report here that phase advance responses to light in hamsters are attenuated if simultaneous locomotor activity is elicited by placing the animal in a novel running wheel during the exposure to light, suggesting that lightinduced phase shifts are influenced by the environmental and behavioral situation in which the light pulse is given.
MATERIALS AND METHODS Golden hamsters (Mesocricetus auratus) were housed individually in running wheel cages kept in constant darkness (DD). Animals were tested at a circadian time (CT) when short light pulses usually cause large phase advances of the rhythm (CT 17-23, 5 to 11hr after the onset of activity, which is defined as CT 12 for nocturnal species). For the first experiment, animals were males obtained from Charles River, Quebec (CRL, Lak:LVG). To test for generalizability of the phenomenon, the experiment was then repeated in another laboratory, using animals obtained from Harlan Sprague-Dawley, IN (HsD: SYR).
354 Downloaded from jbr.sagepub.com at RMIT UNIVERSITY on July 9, 2015
EXPERIMENT I
Animals were 12 weeks old at the time of the light pulse and were housed individually in translucent polypropylene running-wheel cages for activity recording (Dataquest III, MiniMitter Co., Inc., Sunriver, OR). All were maintained in a 24-hour LD cycle comprising 14 hr of light and 10 hr of dark (LD 14: 10) before being placed into DD. After free-running for 10 days in DD, animals were given an activity pulse, a light pulse, or a simultaneous pulse of both. Experimental animals were placed into novel running wheels mounted in clear Plexiglas boxes (Reebs and Mrosovsky, 1989b), and allowed to habituate for 30 min prior to being given a 15-min pulse of light (8-40 lux); they were removed immediately thereafter. All such manipulations were made in the dark with the aid of an infrared viewing scope (FJW Industries, NJ). A light-only control group was given the light pulse while in the home cage without any additional disturbance. A second group of control animals was placed in the novel wheels for 45 min but not given the light pulse. While in the novel situation, animals could not exit from the wheel. The light pulse was delivered by turning on the room
lights. Eye-fitted lines through successive activity onsets for 7 days prior to the pulse and for days 5-12 following the pulse were used to project activity onsets on the day after the pulse. Phase shifts were determined as the difference in projected onset times on this day. The data for the first 4 days after the pulse were omitted to avoid complications due to transient changes in behavior. Animals were categorized as nonrunners if they failed to average 10 revolutions per minute during the light pulse (150 total). EXPERIMENT 2 Hamsters in the second experiment were obtained from Harlan Sprague-Dawley and were 15 weeks of age at the time of the light pulse. They were housed individually in runningwheel cages for activity recording (Dataquest III). After 6 days in LD 14: 10, animals were placed into DD and allowed to free-run for 8 days before being given a 20-min pulse of room light (ca. 10-20 lux). Ten animals remained in their home cages, and five were placed in novel wheels 30 min before the pulse and were removed immediately after. Nonrunners were those that failed to average 10 wheel revolutions per minute during the light pulse (200 total). Phase shifts were based on the same pre- and postpulse days used in Experiment 1, but were calculated using regression lines through activity onsets (Reebs and Mrosovsky, 1989a,b). Onsets were objectively defined as the first 10-min interval with 55 or more wheel revolutions, that was followed by another such interval within 30 min. The data for one animal were discarded because of erratic activity onsets. Ambient temperature for both
experiments
was
21°
±
2°C.
RESULTS In control animals that were not placed in novel wheels, light pulses produced the expected phase advances between CT 17 and CT 23. Animals that were induced to run in the novel wheels, however, showed smaller phase shifts (Figs. 1, 2). Attenuation of the photic response was obtained with either CRL (Experiment 1) or HsD (Experiment 2) hamsters, though the HsD animals showed greater responses to light, possibly due to the slightly longer light pulse given to these animals. Examples of individual actograms are shown in Figure 3.
355 Downloaded from jbr.sagepub.com at RMIT UNIVERSITY on July 9, 2015
FIGURE 1. Effect of stimulated activity on light-induced phase advances, Experiment 1. Animals were given light pulses in their home cages (triangles) or while confined to novel wheels (circles). Filled symbols represent individuals that were active during the pulse (> 150 wheel revolutions). Open symbols represent individuals that did not run (< 150 revolutions). Circadian time refers to the start of the light pulse. Phase shifts are shown in circadian hours (Ch), which were calculated for individual animals as the period (before the pulse) divided by 24. Histograms show the means and SEM for runners and nonrunners. The difference was significant ( p < 0.001, Student’s two-tailed t test).
Some individuals that were placed in the novel wheels in Experiment 1 exhibited very little running while the light was on, although most ran in the dark before the pulse. Statistical comparison of runners versus nonrunners in the novel wheels revealed a significant reduction in the light-induced phase shift (runners, d4>avg = 0.34 ± 0.13 [SEM] hr, n 8; nonrunners, 1.01± 0.21 [SEM] hr, n 5 ; p < 0.01, two-tailedt test). Nonrunners in novel d4>avg wheels responded with phase shifts within the range of the light-only (home cage) controls. In the home cage, most animals remained relatively inactive during the light pulse. However, out of all the hamsters in the two experiments, there were three animals that did run in their home cages at this time. Although this number is too small for statistical evaluation, it should be noted that these individuals responded with relatively small phase shifts (filled triangles in Figs. 1 and 2). Taking the significant difference between runners and nonrunners in the novel wheels together with the similar tendency in the home cage, we conclude that phase shifting to light was attenuated in the active animals. It is unlikely that wheel running simply delayed the clock to a phase in which light was less effective, because the second group of control =
=
=
FIGURE 2. Effect of stimulated activity on light-induced phase advances, Experiment 2. See Figure 1 for explanation of symbols. 200 revolutions (10/min) was used as the criterion for running, since this experiment used 20-min light pulses. The difference between the means (histogram) was significant (p < 0.005, Student’s two-tailedt test). 356 Downloaded from jbr.sagepub.com at RMIT UNIVERSITY on July 9, 2015
FIGURE 3. Examples of actograms from Experiment 1. A, B: Effects of control light pulses (no wheel running) administered at CT 18.1(shift + 1.0 hr) and CT 18.4 (shift + 1.1hr), respectively. C, D: Effects of simultaneous wheel running during light pulses at CT 18.1 (shift = 0.0 hr) and CT 18.3 (shift +0.3 hr), respectively. In each example, the light pulse was delivered on the day indicated by a horizontal arrow, at the time indicated by the star. =
=
=
in novel wheels for 45 min without receiving a phase shifts (A
