Circadian rhythms of locomotor activity in naked mole-rats (Heterocephalus glaber)

Physiology & Behavior 71 (2000) 1 ± 13

Circadian rhythms of locomotor activity in naked mole-rats (Heterocephalus glaber) Alexandra P. Riccio, Bruce D. Goldman* Department of Physiology and Neurobiology, University of Connecticut, Box U-154, Room 2, Building #4, 3107 Horsebarn Hill Road, Storrs, CT 06269, USA Received 29 September 1999; received in revised form 14 December 1999; accepted 27 March 2000

Abstract A wide variety of organisms exhibit various circadian rhythms in their behavior and physiology. Circadian rhythms are regulated by internal clocks that are generally entrained primarily by the environmental light:dark (L:D) cycle. There have been few studies of circadian rhythms in fossorial species that inhabit an environment where day ± night variations are minimal and where exposure to light occurs infrequently. In this study, circadian patterns of wheel-running activity were examined in naked mole-rats (Heterocephalus glaber). Naked mole-rats are fossorial and eusocial, living in colonies of 60 ± 70 animals with only one breeding female. Most individual mole-rats that ran on wheels (65%) exhibited robust circadian rhythms of locomotor activity, entrained to various L:D cycles, and free-ran in constant darkness (DD) with taus averaging 23.5 h. The remainder of the animals either free-ran or were arrhythmic under the various L:D cycles. Mole-rats generally failed to entrain to non-24-h T-cycles with period lengths ranging from T = 23 h to T = 25 h. There was considerable inter-individual variation in the circadian patterns of locomotor activity in naked mole-rats as is observed in other subterranean mammals that have been studied. In contrast to the results obtained when mole-rats were individually housed with access to running wheels, circadian rhythms of general locomotor activity were typically not observed for animals monitored while they were housed in a colony setting. However, clear nocturnal rhythms of general locomotor activity were displayed by four males while residing in their home colonies. Two of these males exhibited the physical appearance of a disperser morph Ð subordinate individuals that are believed to leave their home colonies to achieve reproductive opportunities elsewhere. All four of these males were among the largest males in their respective colonies. These results demonstrate that although naked mole-rats are not frequently exposed to light, the species has retained the capacity to exhibit locomotor patterns of circadian rhythmicity and has the ability to entrain to 24-h L:D cycles. The possible adaptive function of this circadian capacity is discussed. D 2000 Elsevier Science Inc. All rights reserved. Keywords: Naked mole-rat; Circadian rhythm; Locomotor activity; Entrainment

1. Introduction A wide variety of organisms exhibit daily rhythms of behavioral and physiologic changes. These rhythms are mostly generated by an endogenous circadian timing system that is entrained by environmental cues, or zeitgebers. The light:dark (L:D) cycle is the most widely used zeitgeber and is the most potent cue for circadian entrainment in most organisms [24]. Since many species of subterranean mammals have reduced ocular structures and inhabit a niche where day± night fluctuations in temperature and humidity

* Corresponding author. Tel.: +1-860-486-2984; fax: +1-860-4863303. E-mail address: [email protected] (B.D. Goldman).

are minimized, these species are of special interest for comparative studies of the mammalian circadian system [11]. Do these fossorial mammals exhibit circadian rhythms? If they do, is light a major zeitgeber for these rhythms as it is for species that spend a significant part of their active period above ground and receive much greater light exposure? If light is an important zeitgeber, how do subterranean mammals manage to receive adequate photic cues to maintain entrainment? There have been few detailed investigations of circadian rhythms in highly fossorial (i.e., specialized for burrowing) organisms. Most of these studies have focused on various species of mole-rats [2,11,17 ±19,25,26]. Blind mole-rats (Spalax ehrenbergi) showed considerable inter-individual variability with respect to locomotor activity and body temperature (Tb) rhythms. Many individuals exhibited ro-

0031-9384/00/$ ± see front matter D 2000 Elsevier Science Inc. All rights reserved. PII: S 0 0 3 1 - 9 3 8 4 ( 0 0 ) 0 0 2 8 1 - X

2

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

bust rhythms, but a substantial number failed to show rhythmicity. Of those subjects that did show clear circadian rhythms of locomotor activity, some were primarily diurnal and others were primarily nocturnal [2,11,22,25,26]. As with other mammals, blind mole-rats exhibited induction of c-fos in the suprachiasmatic nuclei (SCN) following exposure to a light pulse during subjective night but not during subjective day [30,31]. The SCN comprise the main circadian clock in mammals [20,21], and induction of c-fos in the SCN appears to be related to the phase resetting action of light [1,29]. The subject of the present investigation is the naked mole-rat (Heterocephalus glaber), a subterranean rodent that is native to semi-arid regions of Kenya, Ethiopia, and Somalia. Naked mole-rats are of special interest because they are eusocial, living in colonies that average about 60 ± 70 individuals. Each colony generally includes a single breeding female and one to three breeding males. The remaining animals do not engage directly in reproduction but do aid in general colony maintenance, including care of the young. A caste system of sorts exists within each colony. Some of the animals are more involved with defense against predation and others are more active in foraging and maintenance of the communal burrow [14,15]. One study of circadian rhythms in naked mole-rats revealed that there was no overall circadian rhythm of sleep/wake patterns when observing colony behavior as a whole [6]. A second preliminary study reported circadian rhythms of locomotor activity and body temperature (Tb) in naked mole-rats, but data are shown for just 5 days in one individual, and no attempt was made to determine whether the rhythms were entrainable by light [12]. Our objective in the present study was to determine whether naked mole-rats retain a potentially functional circadian system and, if so, whether these animals are capable of using light cues to entrain their rhythms. In addition to expanding our understanding of the circadian systems of subterranean mammals, it was also our goal to advance the knowledge of the biology of this very unusual species. It was not our immediate objective to determine whether circadian rhythms of locomotor activity have an important function in the lives of naked mole-rats. Nevertheless, some of the data obtained in the course of the study allow us to speculate about a possible functional role for circadian timing in this species. 2. Materials and methods Twenty naked mole-rats were purchased in 1990 from John Visser, Durbanville, South Africa. These were all descendents from populations that were captured in Kenya, and all our animals are derived from this original stock. The animals are maintained in groups of various sizes, ranging from 2 to 48 animals. Pairs of mole-rats are housed in large polypropylene tubs (43  21  16 cm) with clear Plexiglas

covers. Larger groups and colonies are housed in units consisting of combinations of the large tubs and smaller (271512 cm) tubs connected to one another by lengths of acrylic tubing inserted through holes drilled in the sides of the tubs. Each tub is fitted with a Plexiglas lid. Dried corncob bedding is provided. Animals are provided a diet consisting primarily of sweet potato supplemented with apples, carrots, squash, oatmeal, and dog biscuits (Bonz). Food was always present, and new food was introduced from day-to-day at various times between 0900 and 1700 h. During experimental procedures animals were usually supplied with sufficient food to last for 2 ±4 days, so feedings generally only occurred 2 ± 3 times weekly. When animals were housed in continuous darkness (DD), feeding was accomplished during the natural light hours. Mole-rats do not drink and obtain all required water from their food. Room temperature is maintained at 30 ±32°C and the standard photoperiod is 12L:12D, with lights-on at 0500 ±1700 h. All mole-rats used in these studies were adults, ranging from 10 to 50 months of age. These were relatively young animals, as naked mole-rats frequently survive more than 15 years in captivity [14]. None of the animals were breeders within their colonies of origin. During experiments, molerats were housed in temperature and light-controlled environmental chambers (Environair, East Longmeadow, MA). The lighting in these chambers was regulated by a Chrontrol timer (Lindburg Enterprises, San Diego, CA). Each chamber was equipped with two 40-W fluorescent bulbs; these were wrapped in brown paper to reduce the light level to 20 ± 80 lux (lx) in the initial portion of the experiment. During the remainder of the experiment, the bulbs were unwrapped and provided a light level of approximately 200 lx. The temperature in the chambers was maintained at 30°C for some of the studies and 32°C in other experiments. There was no apparent difference in circadian locomotor rhythms of animals housed at 30°C as compared to 32°C. To measure locomotor activity, mole-rats were housed individually in polypropylene tubs (41  20  20 cm) equipped with running wheels and fitted with Plexiglas lids. The wire wheels were wrapped with duct tape to prevent irritation of the animals' feet. Data were collected by a computer running Datacol 3 (Mini-Mitter, Sunriver, OR). Wheel-running data were plotted as actograms using the TAU software program. Animals that did not show significant amounts of wheel-running activity ( < 20 wheel revolutions/ day) for at least 21 days were removed from the experiment. Individual activity records were tested for circadian rhythmicity and period length by using the c2 periodogram analysis that is provided with the TAU program. Sex difference in percentages of animals running on wheels vs. nonrunners was tested by c2 analysis. 2.1. Experiment 1: Entrainment of wheel-running rhythms A total of 23 animals were used in Experiments 1± 3. However, the number of subjects varied between

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

3

Fig. 1. Free-running rhythms of wheel-running activity in two naked mole-rats housed under 12L:12D with a light intensity of 20 ± 80 lx. These actograms and those in Figs. 3 ± 6 are double-plotted (48 h across) with data for consecutive days displayed in descending order. The dark phase of the L:D cycle is represented by the black bars at the top of each record and by stippling. Both these animals free-ran through the L:D cycle with tau (indicated at the right of each record) lengthening when activity onset fell during the light phase and shortening when activity onset occurred during the dark phase. Values for tau during various phases are given at the right of each actogram.

experiments, and some animals were used in more than one experiment. Twenty-three animals (14 males, 9 females) were placed in running wheel cages in an environmental cabinet. Each individual was subjected to a 12L:12D cycle (lights-on at 0500 ±1700 h, the same as in the animal rooms) for at least 21 days. For five of the animals, the light intensity was initially 20± 80 lx; the light intensity was later increased to approximately 200 lx. Other subjects were exposed only to the higher light intensity. Animals were considered to be running if they showed a minimum of 20 wheel rotations/day for 7 consecutive days. Mole-rats were considered entrained to the L:D cycle if activity onset occurred at approximately the same time of day (e.g., with a period length of 24 h) for at least 5 consecutive days. In cases where animals were not entrained, but showed a free-running rhythm, the period length of the rhythm (tau) was determined from the times of activity onset over 10 consecutive days. Subjects that failed to meet the minimum requirement of running during exposure to 12L:12D were not used in further experiments. Except where otherwise specified, all further light cycles employed a light intensity of approximately 200 lx.

2.2. Experiment 2: Phase shift of wheel-running rhythms Four animals that showed wheel-running behavior and entrainment to a 12L:12D cycle (lights-on from 0500 to 1700 h) for a minimum of 21 days were subjected to a 6-h phase advance of the L:D cycle (lights-on 2300± 1100 h). One additional animal that exhibited a freerunning rhythm in the initial 12L:12D cycle was subjected to the same phase advance of the L:D cycle. Activity was monitored and recorded for another 21 days. Thirty-four days after all animals (n = 5) were

Table 1 Wheel-running patterns in 12L:12D (200 lx) Running/total Entrained/running Free-running/running Arrhythmic/running

Males

Females

Total

12/14 8/12 3/12 1/12

3/9 * 2/3 0/3 1/3

15/23 10/15 3/15 2/15

* Statistically different percentages of males vs. females running on wheels (c2 = 4.29; p < 0.05).

4

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

then subjected to a 23-h T-cycle (1L:22D) for 41 days. Following the T = 23-h cycle, the animals were placed

Fig. 2. Wheel-running rhythm in 12L:12D and re-entrainment after a 6-h phase advance of the L:D cycle. Light intensity was approximately 200 lx. During the first 40 days, the animal was entrained to a 12L:12D photoschedule. When the L:D cycle was phase advanced by 6 h, the mole-rat phase advanced its activity and re-entrained to the new 12L:12D cycle. After release from the phase advanced L:D cycle into DD (bottom third of record), the animal began to free-run, starting from the phase achieved during entrainment to the previous L:D schedule.

subjected to the 6-h phase advance, they were placed in DD for 40 days. 2.3. Experiment 3: T-cycles with 1-h light cycle The five animals subjected to DD in Experiments 2 and 7 additional mole-rats that were also in DD for 40 days, were subjected to a 1L:23D cycle for 26 days with lights-on from 0900 to 1000 h. Six of the animals were

Fig. 3. Wheel-running rhythms during exposure to T-cycles with light intensity of 200 lx. This animal was initially entrained to 1L:23D (T = 24 h). When subsequently exposed to 1L:22D (T = 23 h), the animal exhibited two separate bouts of activity. The shorter activity bout coincided with the light pulse (indicated by diagonal bars), whereas the major activity bout had a tau = 23.5 h and was not entrained to the photoschedule. After release from 1L:23D into DD (middle third of record), the major activity bout continued to free-run (tau = 23.5 h); the shorter bout that was previously associated with the light pulse was not apparent during DD. During exposure to 1L:24D (T = 25 h; bottom third of record), the mole-rat again exhibited two bouts of activity. One bout was associated with the light pulse and the other, longer bout, initially continued to free-run with a tau that was the same as in DD. When the 1-h light pulse fell near the onset of this longer bout of activity, its period length increased. However, the increase was not sufficient to allow for entrainment to the T = 25-h cycle.

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

into DD for 21 days. The original 12 mole-rats plus two new animals were subjected to a 25-h T-cycle (1L:24D) for 29 days. Since mole-rats generally failed to entrain to 23- and 25h T-cycles, a T-cycle with a period length closer to 24 h was tested. The same 14 animals were subjected to a 1L:23D cycle for 34 days and then placed into a 24.25-h T-cycle (1L:23.25D) for another 31 days. Following the 24.25-h T-cycle, animals were released into DD for 22 days. Finally, seven males were subjected to a T = 24.5-h cycle (1L:23.5D) with an illumination level of 1000 ±2000 lx. Lighting of this intensity was provided by two Satellite lamps (Northern Light Technologies, Montreal). The animals remained in this T-cycle for 54 days and were then released into DD for an additional 20 days. 2.4. Experiment 4: General locomotor activity of mole-rats housed with their colonies Each of the mole-rats in several colonies was implanted subcutaneously with a chip (AVID, Norco, CA) that uniquely identified that individual. Each colony was housed in

5

a series of rat tubs that were interconnected by acrylic tubes inserted through holes in the tubs. All but one of the tubs was covered with Plexiglas lids. A tub at one end of the housing system was left uncovered. Mole-rats are sensitive to air currents and are aware when their burrow systems are opened. Therefore, we presumed that the uncovered cage might be perceived as a potential site of exit from the artificial tunnel system (see Discussion). Subjects' entrances to and exits from the uncovered tub were monitored via a detector (identity-tag reader, AVID) placed at the entrance to the tub (Fig. 6). The system used for individual identification was interfaced with a computer system so that the events could be recorded and collated separately for each individual in the colony. For purposes of description, the phrase `general locomotor activity' will be used to refer to events of movement through the region of the tube where the animal was detected by the tag reader. This method does not, of course, measure total locomotor activity since the system only allows us to record from the area under the tag reader. However, our casual observations suggest that when mole-rats leave their nest chamber they usually traverse most or all of the colony housing.

Fig. 4. Wheel-running rhythm in T = 24.25 h with light intensity of 200 lx. These animals were initially entrained to 1L:23 D(T = 24 h) and were then exposed to 1L:23.25D (T = 24.25 h). During the first 10 days of the 24.25-h T-cycle, the animal represented in panel A appeared to be entrained. However, during the remainder of the L:D cycle, the mole-rat exhibited a tau with a period length slightly less than 24.25 h. During this time, the animal was active for a brief period just before the light pulse, stopped running during the pulse, and then resumed activity after the end of the light pulse. When released into DD (bottom portion of record), both animals free-ran from the time of activity onset on the prior day.

6

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

All procedures employed in this research were approved by the IACUC at the University of Connecticut. 3. Results 3.1. Experiment 1: Entrainment of wheel-running rhythms Of the 23 animals tested, 15 showed significant amounts of wheel-running activity. Of the 9 females, 3 ran on the wheels, whereas 12 of 14 males ran (Table 1); the sex difference in percentage of animals exhibiting wheel-running behavior was statistically significant (p < 0.01, c2 test). Of the 15 animals that ran on wheels during exposure to 12L:12D at 200 lx, 10 were entrained, 3 were free-running, and 2 were arrhythmic. Three of the five individuals that were initially exposed to 20 ± 80 lx illumination failed to entrain. These three individuals were all males, and they all exhibited taus that varied according to the phase of the L:D cycle in which the onset of activity occurred (Fig. 1). When activity onset occurred during the light portion of the L:D cycle, these animals exhibited taus that ranged from 23.5 to 23.9 h; when activity onset was during the dark phase their taus shortened and ranged from 23.2 to 23.4 h. All but one of the mole-rats that exhibited robust circadian rhythmicity began wheel running within 3 days after being placed in running wheel cages, and 10 of the 13 rhythmic animals showed entrainment to the 12L:12D cycle at 200 lx. One animal started to run 28 days after access to a wheel, and showed circadian rhythmicity within the subsequent 2 days. The majority of the animals that exhibited entrainment were active mostly during the dark portion of the L:D cycle. The duration of heightened wheel-running activity (alpha, a) averaged 4.9 h and the average phase angle of entrainment relative to lights-off was ÿ 1.40 h (ranging from ÿ 9.0 to + 1.3 h); most animals began their activity between 2.5 h before lights-off and the onset of darkness.

activity during the hour of the light phase and this was separate from the main bout of activity, which ranged from 5 to 10 h in duration. Three of the twelve animals failed to entrain and continued to free-run. The remaining animal became arrhythmic in this photoperiod and eventually stopped running for 35 days. The duration of a during each circadian cycle in 1L:23D averaged 6.7 h. During 12L:12D, the activity onset of all animals tested fell during the 2.5-h period immediately preceding the dark phase (Experiment 1). A similar result occurred during the 1L:23D cycle where there was a clustering of activity onset times < 1 h before lights-off

3.2. Experiment 2: Phase shifting of wheel-running rhythms Four of five mole-rats shifted their active phase by approximately 6 h in response to a 6-h phase advance of the L:D cycle (Fig. 2). It took 8± 10 days for each animal to re-entrain to the new L:D cycle. The fifth individual was free-running during the previous L:D cycle, and continued to free-run during the phase shift, failing to entrain to the new L:D cycle. This animal exhibited a tau = 23.8 h. 3.3. Experiment 3: T-cycles with 1-h light cycle Twelve animals were transferred from DD to 1L:23D, and all continued to free-run until the time when activity onset fell within the light phase. Eight of the animals entrained with the majority of their activity occurring during the dark phase. One of these animals showed a short bout ( < 1 h) of

Fig. 5. Wheel-running rhythms in four mole-rats exposed to a 1L:24.5D (T = 24.5 h) with a light intensity of 1000 ± 2000 lx. All animals had been exposed to 1L:23D (T = 24 h) prior to beginning the T = 24.5-h cycle. The animal represented by actogram ``A'' was the only individual that entrained to T = 24.5 h. Following release into DD (bottom portion of record), this animal initiated a free-run with tau < 24 h. One animal (``D'') failed to entrain even to T = 24-h and showed only a very small phase delay when the 1-h light pulse occurred at about the time of activity onset. Most typically (``B'' and ``C''), mole-rats exhibited phase delays when light fell near the time of activity onset, but the delays were apparently not of sufficient magnitude to effect entrainment to T = 24.5 h. Two animals (``A'' and ``C'') consistently showed a brief bout of activity during the light pulse, whereas in the other two individuals (``B'' and ``D'') wheel running stopped during the light pulse. In each case, these appeared to be masking effects of light, separate from the effect of light on the main circadian bout of activity.

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

Fig. 6. Diagram of housing for mole-rat colony and recording of general locomotor activity of individuals in the colony. A series of polypropylene tubs are interconnected by acrylic tubes inserted through holes in the tubs. Each animal bears a subcutaneous chip that is uniquely coded. Events are recorded when an animal passes through the region of a tube surrounded by the identity-tag reader.

(six animals), with two animals beginning their activity 3.1 and 5.9 h after lights-off. A few animals exposed to the 1L:23D cycle began their activity shortly ( < 0.1 h) before lights-on or at the time of lights-on. None of the animals entrained to the 23-h T-cycle (1L:22 D). However, one animal exhibited two separate and

7

distinct bouts of activity, and the shorter bout fell entirely during the 1-h light pulse (Fig. 3). This probably represented a direct response to illumination rather than a circadian regulated event, since the brief activity bout was no longer apparent after the animal was released into DD. The longer activity bout (a averaged over 10 consecutive days was 10 h) of this mole-rat had a tau of 23.5 h, and was clearly not entrained to the 23-h T-cycle. The other four animals freeran with taus ranging from 23.1 to 23.6 h and averaging ( ‹ SEM) 23.3 ‹ 0.10 h. When placed into DD, all animals free-ran from the time of activity onset on the previous day with period lengths ranging from 23.0 to 23.7 h. None of the animals showed clear entrainment to the 25h T-cycle. One individual exhibited a short bout of activity ( < 1 h) during the light phase that persisted for 21 days. This could have been a non-circadian response to light since the short bout of activity was confined to the 1-h light phase; therefore, we did not consider this animal to be truly entrained. This animal's circadian rhythm of activity eventually became less robust after about 16 days of exposure to the 25-h T-cycle. Another animal exhibited two bouts of activity in the 25-h T-cycle; one short bout ( < 1 h) occurred during the light phase and a longer bout (mean a = 9.3 h) with a tau of 23.5 h for the first 12 days and a tau of 23.9 h for the remainder of the T-cycle. This was the same animal that exhibited two bouts of activity under the 23-h T-cycle (Fig. 3). Again, we did not consider this animal to be entrained because the period length of its main bout of activity was clearly less than the 25-h period length of the

Fig. 7. Actograms depicting general locomotor activity of three mole-rats housed with their colony (total = 30 animals). These actograms are single-plotted with consecutive days of activity plotted in descending order. The middle actogram is from a male mole-rat that exhibited the physical characteristics described for disperser morphs [23]. Dark phases of the L:D cycle are depicted by heavy bars. The colony was exposed, in sequence, to 16L:8D, 12L:12D, continuous dim illumination (indicated by solid bar across entire record), and 6L:18D. The disperser exhibited a robust nocturnal rhythm of activity, whereas the other two individuals did not show rhythmic activity and tended to be more active during the light phase.

8

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

L:D cycle; the additional short bout of activity probably reflected a direct response to the brief light phase rather than a circadian regulated event. None of the 14 animals subjected to the 24.25-h T-cycle exhibited unequivocal evidence of entrainment. However, one animal exhibited two distinct bouts of activity: a short bout ( < 1 h) that occurred before the 1-h light pulse and a longer bout averaging 4.3 h (averaged over 10 consecutive days) that began approximately at lights-off (Fig. 4). Since

this animal's tau was slightly < 24.25 h, it might be that light was having only a masking effect on activity. However, this animal exhibited a period length that was >24 h, whereas all our other mole-rats exhibited free-running taus < 24 h in DD. This suggests that light was delaying activity, but not enough to entrain the animal to a T-cycle of 24.25 h. When released into DD this animal free-ran from the point of activity onset on the previous day with a tau < 24 h. Another animal appeared entrained, but became arrhythmic

Fig. 8. Actograms depicting general locomotor activity of three mole-rats housed with their colony (total = 12 animals). The actograms are single-plotted, and the dark phase of the light cycle (12L:12D) is indicated by the heavy bars above the record and by stippling. The actogram in the center shows the nocturnal activity of the disperser morph, whereas the actograms on the left and right show the more scattered, and primarily diurnal, activity of two other colony members. Photographs of these animals are shown below their respective actographs. Note that the disperser (center) has considerable fatty tissue, particularly evident in the neck region. Several days of data were lost (just below middle of each actogram) due to a power failure.

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

9

Fig. 9. Actograms of three male mole-rats from a colony of 12 animals. The actograms are single-plotted. The dark phase is indicated by heavy bars and stippling. The animals were initially exposed to 12L:12D. They were released into DD for 12 days (middle of record), followed by a 12L:12D cycle that was phase delayed by 4 h relative to the initial 12L:12D cycle. Finally, the animals were again exposed to the original 12L:12D cycle. The record on the far right is from the disperser depicted in Fig. 8. The arrow indicates the time when the photograph of Fig. 8 was taken. This animal stopped exhibiting its nocturnal rhythm of activity while in DD and did not regain rhythmicity when light cycles were resumed. Several days of data were lost (during the initial 12L:12D photoschedule) due to a power failure.

when released into DD so that a masking effect of light could not be ruled out. Of the remaining 12 animals, 5 became arrhythmic and 7 free-ran with periods lengths averaging 23.6 h. Even when the light intensity was increased to 1000 ± 2000 lx, only one of seven animals entrained to a T = 24.5-h cycle (Fig. 5A). One animal failed to entrain to either T = 24 h (1L:23D) or T = 24.5 h at this light intensity (Fig. 5D). The remaining five subjects did entrain to T = 24 h, but not to T = 24.5 h (Fig. 5B,C). 3.4. Experiment 4: General locomotor activity of mole-rats housed with their colonies Recordings of general activity were made for several colonies totaling more than 100 individuals. Individuals in these colonies were generally arrhythmic or had weak indications of circadian rhythmicity. In the large majority of animals more activity occurred during the light phase than during the dark. One individual presented a striking exception to this pattern. This animal showed a robust nocturnal pattern of activity with activity beginning at approximately the onset of the dark phase. The animal was a relatively large male with a large amount of subcutaneous adipose tissue, especially evident in the neck region. Thus, the animal had the physical characteristics of a disperser morph Ð nonbreeding mole-rats that are thought to leave their home colonies to possibly acquire breeding opportunities elsewhere [23]. This male was a member of a colony of 30

individuals, and all others in the colony showed the more typical Ð scattered and primarily diurnal Ð pattern of general locomotor activity (Fig. 7). Later, a second male exhibiting the physical characteristics of a disperser was found in another colony of 12 animals. This animal was the largest male in the colony. General locomotor activity was recorded for all the animals in this colony as described above. Again, the male with the physical attributes of a disperser exhibited a nocturnal rhythm of activity, whereas the other members of the colony were not clearly rhythmic and tended to be more active during the day (Fig. 8). After several weeks, the `disperser' male stopped showing its nocturnal rhythm and, shortly thereafter lost approximately 20% of its body weight (decrease from 61.8 to 49.3 g) and ceased showing the body fat accumulation that characterizes dispersers. At about this time, two other males in this colony (the next largest males after the disperser) began to exhibit robust nocturnal rhythms of general activity (Fig. 9). These males did not exhibit the large accumulation of subcutaneous fat that has been described for dispersers. 4. Discussion Circadian rhythms probably evolved as adaptations that allowed organisms to prepare for relatively predictable environmental changes associated with the day ±night cycle [24]. Subterranean organisms are subject to far less circa-

10

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

dian variation in habitat characteristics as compared to organisms that live above ground. For example, diurnal changes in ambient temperature and humidity are diminished in the underground habitat. Despite these environmental characteristics, several species of almost exclusively fossorial rodents have retained circadian systems that are capable of driving clear rhythms of behavioral and physiological change [2,11,17,18,30]. In the past two decades, considerable attention has focused on a particular group of subterranean rodents Ð the mole-rats of the family Bathyergidae. These animals have been of interest not only for their fossorial life style, but also because some of the bathyergids exhibit a eusocial reproductive pattern that was formerly unknown among mammals. The most studied of the bathyergids is the naked mole-rat, a species that lives in large colonies inhabiting extensive burrow systems [3]. In the present study, individual naked mole-rats expressed robust circadian rhythms of wheel-running activity under L:D cycles and in DD. Most subjects that exhibited circadian patterns of wheel running readily entrained to 12L:12D or 1L:23D light cycles, provided the illumination was approximately 200 lx. When the light intensity was only 20 ± 80 lx, most animals failed to entrain to 12L:12D. Rather, free-running patterns were observed with period lengths of 23.2± 23.9 h. The lower light levels were not completely without effect, however, since the period length of the free-run usually lengthened (i.e., became closer to 24 h) when the onset of activity fell within the light phase. This observation suggests that light was exerting a phase delaying action when exposure occurred during the early part of the active phase. The phase delays evoked by 20 ±80 lx illumination apparently were not of sufficient magnitude to effect entrainment to the 24-h light cycle. This is quite different from results in laboratory rats, mice, and hamsters, where entrainment to L:D cycles is readily accomplished with light intensities of a few lux [8]. The onset of wheelrunning activity Ð the phase at which light appeared to induce phase delays Ð probably corresponds to early subjective night, since when mole-rats entrained to 12L:12D at 200 lx, the onset of activity generally fell around the beginning of the dark phase and most of the activity occurred during the dark. This pattern of entrainment is typical of nocturnal rodents. There are at least two possible explanations for the relatively high threshold for entrainment to light in naked mole-rats as compared to other rodents: (a) This may be a species-specific trait. In this case, it could reflect a partial `degeneration' of the circadian system, perhaps indicating that entrainment to light cues is not important for naked mole-rats in their burrow habitats. However, this species lives in open, arid or semi-arid areas where the animals might be exposed to high intensity illumination when opening their burrow to extrude loose earth [3]. Comparative studies in other rodents have suggested that species differences in the illumination threshold for circadian responses may be related to habitat; species inhabiting woodlands

have a lower threshold than those living in open areas [7]. Therefore, it remains possible that naked mole-rats do use light cues to track local time in the field. (b) It is possible that the relative insensitivity to light is a laboratory artifact. Our mole-rats are raised in rooms where they are exposed to light for 12 h/day. In the field these animals spend the large majority of time in the total darkness of the burrows. Thus, it may be that exposure to light for even 12 h/day is sufficient to induce some degree of retinal degeneration, as has been reported in laboratory rats following chronic exposure to continuous illumination [28], and this might result in reduced photic sensitivity. Two observations in the present studies demonstrate conclusively that naked mole-rats were entrained to light cycles rather than to some other (uncontrolled) cue: (a) When subjected to a 6-h phase advance of the L:D cycle, the animals exhibited a corresponding phase shift of their wheel-running rhythms. (b) When released into DD, molerats exhibited free-running rhythms, and the phase of the initial portion of the free-run corresponded to the phase attained during previous entrainment to L:D. One mole-rat exhibited a rhythm with a period length of 24 h in DD. It may be that this animal had a tau that was so close to 24 h as to be indistinguishable from a 24-h rhythm; alternatively, the animal may have been entrained to some unknown zeitgeber. The same individual did appear to entrain to a 12L:12D cycle and also exhibited a phase shift when the L:D cycle was phase advanced by 6 h. These observations indicate that this individual was capable of entraining to light cues. One prominent feature of the present study was the variability among individual naked mole-rats with respect to running on wheels. Of the 23 animals tested, 15 exhibited significant amounts of wheel running and 13 out of 15 of the animals that did run showed robust circadian rhythms. Similar variability with respect to the number of animals running on wheels was observed in another subterranean mammal, the blind mole-rat (S. ehrenbergi). However, in comparison to blind mole-rats, the naked mole-rats that ran on wheels were generally more consistent in exhibiting clear rhythmic patterns and also with respect to the phase relation between activity and the L:D cycle. It has been suggested that inter-individual variability in the expression of circadian patterns in subterranean species may reflect different individual strategies (i.e., rhythmic vs. arrhythmic; nocturnal vs. diurnal) in adapting to an environment where circadian variations in most environmental parameters are relatively small, so that there may not be an overwhelming advantage for any one particular strategy [11]. Another feature that varied between individuals was the masking action of light on wheel running. For many animals, light did not appear to directly stimulate or inhibit wheel running and only acted via its circadian phase resetting actions. However, some individuals consistently ran during 1-h light pulses (Figs. 3 and 5A,C), whereas other subjects stopped running during 1-h light pulses (Figs.

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

4 and 5B,D). The nature of these non-circadian effects of light appeared to correlate with the duration of the main (circadian) bout of activity: Animals that ran during 1-h light pulses had long circadian bouts (long a); subjects that stopped running during the light pulse had relatively short circadian bouts of activity. It is not clear what significance, if any, these acute responses to light might have for the natural behavior of naked mole-rats. It would be interesting to establish whether the two opposite responses to light are correlated with the behavioral roles of individuals within their colonies (caste system). Male naked mole-rats may be more likely to exhibit wheel-running behavior as compared to females, since 12 of 14 males ran on wheels as compared to only 3 of 9 females ( p < 0.01). Only non-breeding animals were used in the running wheel experiments, so we have no information on the possible circadian capacities of the reproductives. In the present study, all animals were tested during chronic isolation from their colonies. Under these conditions, naked mole-rats are reported to exhibit changes in endocrine status. Non-breeding females become ovulatory, with increased levels of urinary progesterone following isolation [10], whereas non-breeding males exhibit increased urinary testosterone during isolation [9]. It is possible that changes in endocrine status following isolation might influence the propensity to run on wheels and/or the expression of circadian rhythms in mole-rats. Alternatively, isolation may act in some other way to promote expression of rhythmic behavior. However, a recent study in our laboratory indicated that naked mole-rats need not be completely isolated from their colony mates to express running wheel rhythms. All animals in each of two colonies were isolated in wheel cages for 22± 23 h each day and were returned to their respective colonies for the remaining 1 ±2 h. This short period of social contact was sufficient to maintain colony recognition, so that animals did not exhibit aggression toward their colony mates. In this study, 6 of 18 individuals exhibited clear circadian rhythms of wheel running (C. Jack and B. Goldman, unpublished data). It is interesting that a species which has specialized for an almost exclusively subterranean existence has retained not only a capacity for exceptionally precise endogenous circadian timekeeping, but is also capable of entrainment to L:D cycles. It has been noted that whereas fossorial mammals spend most of their lives in darkness, they occasionally open their burrows to extrude freshly dug soil to the surface [11,24,25]. Animals could presumably receive light cues at these times. Retention of a light entrainable circadian system could be useful for timing of particular activities, especially those that carry a high risk of predation. Our observations of circadian rhythms of locomotor activity in naked mole-rats may appear to contradict results of a study that failed to reveal sleep/wake rhythms in most animals of this species; in that study, only 2 of 47 individuals showed significant circadian cycles [6].

11

However, we believe that the differences may be explained by the different measurements that were used in the respective studies. The previous study examined resting/active states in mole-rats by visual observation of animals in a colony setting [6]. Work in our laboratory also failed to reveal clear evidence of circadian rhythms of general locomotor activity in most individuals when mole-rats were studied while residing in their colonies. In these studies, clear circadian rhythms were detected in only 4 of more than 100 individuals from several colonies. This contrasts sharply with the robust wheelrunning rhythms seen in 13 of 23 animals that were housed individually. The four mole-rats that showed circadian patterns of locomotor activity while they remained with their colonies are of particular interest, since their behavior in a colony setting may provide clues to natural functions of the circadian system in this species. One of these individuals was a non-breeding male in a colony of 30 animals. This animal had the morphological characteristics described for `disperser morphs'. Dispersers are non-breeding males that exhibit a tendency to leave the confines of a laboratory colony if provided with a single opening in the `burrow' system. It has been suggested that these animals might represent males that sometimes leave their home colonies in the field and outbreed by either invading other established colonies or by soliciting one or more foreign individuals to form a new colony. Dispersers attempt to mate with individuals from other colonies when paired with them in a neutral arena [23]. Our male also attempted to mate when paired with `foreign' mole-rats. This individual exhibited a nocturnal activity pattern, with entrances to the uncovered tub (Fig. 1) beginning at about the time of lights-off in a 16L:8D photoschedule. A delay of lights-on by 4 h to achieve a 12L:12D cycle (time of lights-off unchanged) failed to result in a change in the time of activity. However, when the room was subsequently placed in continuous dim red illumination for 3 weeks, the activity of this animal appeared to free-run with tau < 24 h (Fig. 7). Thus, it appears that this mole-rat was exhibiting a circadian pattern of activity based on an endogenous oscillation that was entrainable to light. The three remaining individuals that exhibited clear rhythms of general locomotor activity were non-breeding males in a colony of 12 animals. They were the three largest of the seven males in the colony, and one of them had the morphological characteristics described for dispersers (Fig. 8). As with the disperser from the larger colony, these three mole-rats were distinctly nocturnal with respect to locomotor activity. It is interesting that the activity of all four rhythmic males consisted, at least in part, of excursions into the single uncovered tub at one end of the colony housing system. It is tempting to speculate that the mole-rats might have been attempting to leave their respective colonies. If naked mole-rats do disperse above ground, it probably would be safer for this presumably high-risk activity to

12

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

occur at night. It is possible that naked mole-rats may use their capacity to exhibit circadian timing only with respect to a relatively narrow range of tasks, and this might include above ground forays that occur with low frequency and involve only a small number of individuals. This could help explain why our laboratory and others have had difficulty detecting evidence of rhythms in `natural' behaviors of this species. There is strong evidence for dispersal of naked mole-rats from their home colonies in the field. Stanton Braude captured and marked animals from 54 colonies in Meru National Park, Kenya, and then returned them to their burrows. Animals were recaptured annually and 16 of the marked animals were found separated from their original colonies. One of these was found in a different colony, whereas the others were alone or in pairs or small groups. In addition to the 16 marked animals, 15 unmarked adults were found in these nascent colonies along with the marked animals. Most of the marked animals were found 150 ± 300 m away from their original colonies, and 2 individual had traveled more than 2 km. Based on the composition of nascent colonies that contained marked animals, it was estimated that naked mole-rat dispersers are outbreeding in at least 88% of the cases of new colony formation [4]. Stanton Braude (personal communication) has never observed a naked mole-rat during an above ground foray, but he has reports from four individuals who claim to have sighted animals above ground; in each of these cases the sightings occurred at night. This information is intriguing in view of the distinctly nocturnal patterns of locomotor activity observed for putative dispersers in our laboratory colonies. Even though dispersal is probably an infrequent event, it is likely to have considerable importance for the evolutionary history of the species, as it may represent the main (or only) opportunity for outbreeding. A preference for outbreeding among naked mole-rats that are undergoing reproductive development was demonstrated in a laboratory study. Non-breeding animals were removed from their home colonies and housed with members of the opposite sex from both their home colonies (familiar siblings) and from other colonies (unfamiliar distant kin). In this mating choice paradigm, mating behavior was observed only between males and females that were unrelated to each other, and in eight of nine breeding pairs that developed the members of the pair were from different colonies [5]. Circadian rhythms of general locomotor activity were reported for another eusocial bathyergid, the Damaraland mole-rat (Cryptomys damarensis). In one study, two colonies (five and four individuals, respectively) of Damaraland mole-rats exhibited circadian rhythms when summed activity of all colony members was monitored via motion detectors. Most activity occurred during the light phase in 16L:8D and 12L:12D photoperiods, and free-runs were observed when the colonies were in DD [17]. A second study revealed fluctuations in locomotor activity and body

temperature in two Damaraland mole-rats, but the variations were not clearly circadian [16]. Circadian rhythms of general locomotor activity were observed in the solitary Cape mole-rat (Georychus capensis), and these animals were primarily nocturnal [19]. Another solitary subterranean species, the blind mole-rat, also has a well-documented capacity to exhibit circadian variations of locomotor activity [2,11,25,26,30]. Why has the capacity to display circadian rhythms been retained in subterranean mammals? It is possible that these species have merely retained intact circadian systems as they evolved from ancestral species that spent considerable time above ground, and that their rhythmic capabilities are put to little or no use in the field. This point of view might be supported by the observation of more inter-individual variability in certain aspects of circadian function as compared to what is typically seen in other mammals [11]. Another possibility is that circadian capacities have been retained because they are still useful, albeit perhaps for just a few specific functions. For example, it might be that circadian rhythms could be very important for helping subterranean mammals to determine when to make their extremely infrequent forays above ground, or to establish optimal times for removing loose earth from the burrows Ð activities that might expose the animals to predation. It has also been proposed that circadian systems might be valuable for providing internal temporal order [24], and this could provide selective advantage for the retention of rhythms even in organisms that inhabit relatively constant environments. Further, whereas subterranean mammals may frequently encounter minimal day ± night variation in environmental conditions, virtually all these species are subjected to important seasonal changes. In mammals, the circadian system is an essential component of the photoperiodic system that measures day length and uses changes in day length to help regulate seasonal adaptations [32]. Thus, the ability to use photoperiod information to time seasonal adjustments could provide another adaptive rationale for the retention of the circadian system in subterranean species. It has been noted that whereas there are two groups of subterranean mammals that have completely lost external eyes Ð the mole-rats of the family Spalacidae and the golden moles of the family Chrysochloridae Ð all these species have retained small retinal structures located beneath the skin [13]. Careful study of these rudimentary retinas has been carried out only for the blind mole-rat, where neural projections from the retina to the SCN are apparently capable of allowing for entrainment of circadian rhythms to the L:D cycle, as is the case for sighted mammals [11,26,27]. These types of observations might be taken as support for the notion that subterranean mammals may continue to use their circadian systems and that they might use light as a zeitgeber for entrainment to local time. Field studies will be required to determine if, and how, circadian rhythms have a role in the lives of these species.

A.P. Riccio, B.D. Goldman / Physiology & Behavior 71 (2000) 1±13

Acknowledgments The authors thank Sharry Goldman and Matthew Goldman for aid in preparation of the figures. We are grateful to Stanton Braude, Paul Sherman, and Sharry Goldman for helpful comments regarding preparation of the manuscript. We also thank Kevin Kavanagh and Charlene Taylor for their excellent services in care of the animals. References [1] Aronin N, Sagar SM, Sharp FR, Schwartz WJ. Light regulates expression of a Fos-related protein in rat suprachiasmatic nuclei. Proc Natl Acad Sci USA 1990;87(1):5959 ± 62. [2] Ben-Shlomo R, Ritte U, Nevo E. Activity pattern and rhythm in the subterranean mole-rat superspecies, Spalax ehrenbergi. Behav Gen 1995;25(1):239 ± 45. [3] Braude SH. Which naked mole-rats volcano? In: Sherman PW, Jarvis JUM, Alexander RD, editors. The biology of the naked molerat: Princeton Univ. Press, 1991. pp.185 ± 94. [4] Braude SH. Dispersal and new colony formation in wild naked molerats: evidence inbreeding as the system of mating. Behav Ecol 2000;11(1):7 ± 12. [5] Ciszek D. New colony formation in the ``highly inbred'' eusocial naked mole-rat: outbreeding is preferred. Behav Ecol 2000;11(1):1 ± 6. [6] Davis-Walton J, Sherman PW. Sleep arrhythmia in the eusocial naked mole-rat. Naturwissenschaften 1994;81(1):272 ± 5. [7] DeCoursey PJ. Photoentrainment of circadian rhythms: an ecologist's viewpoint. In: Hiroshige T, Honma K, editors. Circadian clocks and ecology. Sapporo: Hokkaido Univ. Press, 1989. pp. 187 ± 206. [8] DeCoursey PJ. Circadian photoentrainment in nocturnal mammals: ecological overtones. Biol Behav 1990;15(1):213 ± 38. [9] Faulkes CG, Abbott DH. Social control of reproduction in breeding and non-breeding male naked mole-rats (Heterocephalus glaber). J Reprod Fertil 1991;93(1):427 ± 35. [10] Faulkes CG, Abbott DH, Jarvis JUM. Social suppression of ovarian cyclicity in captive and wild colonies of naked mole-rats, Heterocephalus glaber. J Reprod Fertil 1990;88(1):559 ± 68. [11] Goldman BD, Goldman SL, Riccio AP, Terkel J. Circadian patterns of locomotor activity and body temperature in blind mole-rats, Spalax ehrenbergi. J Biol Rhythms 1997;12(1):348 ± 61. [12] Herold N, Spray S, Horn T, Henriksen SJ. Measurements of behavior in the naked mole-rat after intraperitoneal implantation of a radiotelemetry system. J Neurosci Methods 1998;81(1):151 ± 8. [13] Hickman GC. The Chrysochloridae: studies toward a broader perspective of adaptation in subterranean rodents. In: Nevo E, Reig OA, editors. Evolution of subterranean mammals at the organismal and molecular levels. New York: Alan R. Liss, 1990. pp. 23 ± 48. [14] Jarvis JUM. Reproduction of naked mole-rats. In: Sherman PW, Jarvis JUM, Alexander RD, editors. The biology of the naked mole-rat. Princeton, NJ: Princeton Univ. Press, 1991. pp. 384 ± 425.

13

[15] Lacey EA, Sherman PW. Social organization of naked mole-rat colonies: evidence for divisions of labor. In: Sherman PW, Jarvis JUM, Alexander RD, editors. The biology of the naked mole-rat. Princeton, NJ: Princeton Univ. Press, 1991. pp. 275 ± 336. [16] Lovegrove BG, Heldmaier G. The amplitude of circadian body temperature rhythms in three rodents (Aethomys namaquensis, Thallomys paedulcus, and Cryptomys damarensis) along an arboreal ± subterranean gradient. Aust J Zool 1994;42(1):65 ± 78. [17] Lovegrove BG, Heldmaier G, Ruf T. Circadian activity rhythms in colonies of `blind' molerats, Cryptomys damarensis (Bathyergidae). S Afr Tydskr Dierkd 1993;28(1):46 ± 55. [18] Lovegrove BG, Muir A. Circadian body temperature rhythms of the solitary cape mole rat, Georychus capensis (Bathyergidae). Physiol Behav 1996;60(1):991 ± 8. [19] Lovegrove BG, Papenfus ME. Cricadian activity rhythms in the solitary cape molerat (Georychus capensis: Bathyergidae) with some evidence of splitting. Physiol Behav 1995;58(1):679 ± 85. [20] Meijer JH, Rietveld WJ. Neurophysiology of the suprachiasmatic circadian pacemaker in rodents. Physiol Rev 1989;69(1):671 ± 707. [21] Morin LP. The circadian visual system. Brain Res Rev 1994;67(1): 102 ± 27. [22] Nevo E, Guttman R, Haber M, Erez E. Activity patterns of evolving mole rats. J Mammal 1982;63(1):453 ± 63. [23] O'Riain MJ, Jarvis JUM, Faulkes CG. A dispersive morph in the naked mole-rat. Nature 1996;380(1):619 ± 21. [24] Pittendrigh CS. Temporal organization: reflections of a Darwinian clock-watcher. Annu Rev Physiol 1993;55(1):17 ± 54. [25] Rado R, Terkel J. Circadian activity of the blind mole rat, Spalax ehrenbergi, monitored by radio telemetry, in seminatural and natural conditions. In: Spanier E, Steinberger Y, Luria M, editors. Environmental and ecosystem stability: environmental quality, vol. IV-B. Jerusalem, Israel: ISEEQS Pub., 1989. pp. 392 ± 400. [26] Rado R, Gev H, Goldman BD, Terkel J. Light and circadian activity in the blind mole rat. In: Riklis E, editor. Photobiology. New York: Plenum, 1991. pp. 581 ± 9. [27] Rado R, Wollberg Z, Terkel J. Sensitivity to light of the blind molerat: behavioral and neuroanatomical study. Isr J Zool 1992;38(1): 323 ± 31. [28] Reiter RJ, Klein DC. Observations on the pineal gland, Harderian glands, the retina, and the reproductive organs of adult female rats exposed to continuous light. J Endocrinol 1971;51(1):117 ± 25. [29] Rusak B, Robertson HA, Wisden W, Hunt SP. Light pulses that shift rhythms induce gene expression in the suprachiasmatic nucleus. Science 1990;248(1):1237 ± 40. [30] Tobler I, Herrmann M, Cooper HM, Negroni J, Nevo E, Achermann P. Rest ± activity of the blind mole-rat, Spalax ehrenbergi, under different lighting conditions. Behav Brain Res 1998;96(1):173 ± 83. [31] Vuillez P, Herbin M, Cooper HM, Nevo E, Pevet P. Photic induction of Fos immunoreactivity in the suprachiasmatic nuclei of the blind mole rat (Spalax ehrenbergi). Brain Res 1994;654(1):81 ± 4. [32] Underwood H, Goldman BD. Vertebrate circadian and photoperiodic systems: role of the pineal gland and melatonin. J Biol Rhythms 1987;2(1):279 ± 315.