INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE GOLDEN-MANTLED ground squirrel (Spermophilus lateralis) displays marked annual rhythms in body mass, hibernation, and reproduction (9, 22). Body mass increases during spring and summer to reach peak values just before squirrels enter hibernation in late summer. Because food consumption is greatly reduced or suppressed during the hibernation season, body mass declines steadily despite the greatly reduced energy expenditure during hibernation (22). Males emerge from hibernation in the early spring, and mating activity begins when the females emerge a few weeks later, coincident with the annual trough in male body mass (2; also see Ref. 18 for Richardson's ground squirrels). As the mating season draws to a close in mid-May (2), the cycle repeats, and males again fatten in preparation for the next hibernation season. Plasma concentrations of pituitary gonadotropins and gonadal steroids also vary seasonally, with high values during the reproductive season and basal or undetectable concentrations the remainder of the year (1, 16; reviewed in Ref. 24).

Annual rhythms that persist for two or more cycles in the absence of seasonal changes in day length, temperature, food availability, and other factors are designated circannual rhythms (CARs); they share many formal properties with circadian rhythms (7). The body mass rhythm of S. lateralis is the most extensively studied mammalian CAR. Among squirrels of both sexes, the period () of the free-running CAR of body mass deviates from 12 mo (26). The  for the body mass and reproduction rhythms of golden-mantled ground squirrels maintained at 23°C and on a 14:10-h light-dark cycle, is ~10.5 mo (11) and can be entrained to a period of ~12 mo by a simulated natural photocycle (13).

Circadian rhythms are influenced by ovarian hormone secretions. In female hamsters and rats (reviewed in Ref. 19), increased estradiol secretion is associated with a shorter circadian  and phase advances in activity onset (20). Similarly, chronic estradiol treatment markedly changes the waveform of the CAR of body mass such that the mass-gain phase is lengthened and the mass-loss phase is shortened (25). When estradiol treatments are restricted to the mass-gain and mass-loss phases of the annual cycle, they produce substantial phase delays and phase advances, respectively, in the body mass CAR of ovariectomized S. lateralis (14).

It is not known whether estradiol has similar phase-shifting actions on CARs of male ground squirrels. Some steroid hormones affect traits equally in both sexes. For example, testosterone terminates hibernation in both female and male hamsters (8); for other traits, e.g., the induction of lordosis by estradiol, female responsiveness greatly exceeds that of males (6). The main purpose of the present experiment was to assess sex differences in the phase-shifting effects of estradiol on CARs of squirrels.

The circadian system of some rodents is sexually differentiated; e.g., male Syrian hamsters gonadectomized in adulthood, unlike their ovariectomized female counterparts, show no changes in circadian  during treatment with estradiol (27). If the circadian-circannual analogy holds in this instance (7), then CARs of male squirrels should be less responsive than those of females to phase-shifting actions of estradiol. To evaluate this hypothesis, we determined whether CARs in body mass of male squirrels show responses to estradiol similar to or different from those previously described for females (14). Adult male gonadectomized squirrels were treated with estradiol either during the mass-gain or mass-loss phase of their initial circannual cycle; the effects on subsequent body mass CARs were then determined.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Male golden-mantled ground squirrels, S. lateralis, born in the laboratory to wild-caught pregnant females trapped in Lassen County, CA, were individually housed at 23 ± 2°C on a 14:10-h light-dark cycle, with lights on at 0700, Pacific Standard Time. Each cage was provided with pine shavings and a continuous supply of water and food (Simonson rat pellets, maintenance diet). Animals were weighed weekly throughout their lives, beginning at 35 days of age.

Experiment 1: Estradiol Treatment During the Mass-Gain Phase of the Circannual Cycle

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Experimental treatments for experiments 1 and 2. Each complete oscillation represents one annual cycle of body mass, the actual duration of which varied from 10 to 13 mo, depending on treatment (see Table 1). Hatched areas indicate the treatment period during which squirrels were implanted with capsules containing estradiol benzoate (EB squirrels) or with empty capsules (blank squirrels). In both experiments, peak (PK) and trough (TR) numbers begin with the first peak and trough for each animal's lifetime (i.e., numbering of peaks and troughs is the same for both experiments). Animals were housed on a 14:10-h light-dark cycle throughout both experiments.

When a squirrel had reached the subsequent peak in its body mass, defined by the failure to increase body mass for two consecutive weeks, its capsule was removed, and weekly weighings continued until the end of the experiment.

Experiment 2: Estradiol Treatment During the Mass-Loss Phase

When squirrels reached the subsequent trough in body mass, defined by a failure to lose mass on two consecutive weeks, capsules were removed; weekly weighings continued until the end of the experiment.

Capsules

Surgery

Blood Sampling

Hormone Assays

Data Analysis

In experiment 1, although animals were initially assigned to EB and blank groups to balance the treatment groups for body mass and calendar date of body mass phase reference points occurring before capsule implantation, the removal of animals from analysis for the reasons cited above resulted in treatment groups that were not balanced. One animal was therefore removed from the blank group, so that the groups were balanced for median dates of peak 1 and trough 1. Of the 36 squirrels that started each experiment, data from 12 and 13 individual squirrels were used for the final analysis in experiments 1 and 2, respectively, resulting in sample sizes similar to those in the corresponding study on female squirrels (14), which was used as a basis for comparison.

The weekly body masses were analyzed with a time-dependent clustering algorithm (3) that yielded dates of mid-peaks and mid-troughs in the annual cycle. Dates are shown as day number within the experiment, where day 0 is the same calendar date for each individual. The  of the CAR, measured from a given phase reference point to its recurrrence during the next annual cycle, was computed as the difference between mid-peak or mid-trough values obtained from cluster analysis; for example, peak-to-peak  for year 1 was obtained by subtracting the date of mid-peak 1 from the date of mid-peak 2. In experiment 1, the duration of the mass-gain phase of the cycle during which EB treatment occurred was computed by subtracting the date of mid-trough 1 from the date of mid-peak 2 (Fig. 1). In experiment 2, the duration of the mass-loss phase during which EB treatment occurred was calculated by subtracting the date of mid-peak 1 from the date of mid-trough 1 (Fig. 1).

Mid-peak and mid-trough dates, peak-to-peak and trough-to-trough  values, duration of rising or falling body mass phase during hormone treatment, and plasma estrogen concentrations were compared between EB- and blank-treated groups using the nonparametric Mann-Whitney U-test; these data were not normally distributed. Peak masses, trough masses, and peak-to-trough amplitudes were compared across three cycles with one-way analysis of variance; Fisher's protected least significant difference (PLSD) test was used for post hoc comparisons of between-cycle values.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Experiment 1

Absolute body mass peaks and troughs. Peak, trough (Table 2), and amplitude of each body mass cycle were statistically indistinguishable when the two treatment groups were compared. Both groups showed a tendency toward lower peak masses during the first than during the second or third cycles, but only the difference between the first and third peaks was statistically significant (P < 0.02 for values of both treatment groups combined; Table 2). Trough body masses did not differ significantly from each other during each of the 3 yr of testing (Table 2).

Effects of EB during the mass gain phase. Estrogen treatment extended the median duration of the mass gain phase by 19 days in EB trough compared with blank trough squirrels (P < 0.05; Table 3). The  of the CAR, measured between troughs 1 and 2, was 54 days longer in EB trough than in blank trough squirrels (P < 0.004, Table 3). The  for peaks 1-2 and 2-3 did not differ between the treatment groups.

Experiment 2

Absolute body mass peaks and troughs. Body mass peak, trough (Table 2), and amplitude during each cycle were statistically indistinguishable when the two treatment groups were compared. Both groups attained significantly lower peak masses during the first cycle than in either the second or third cycles (P < 0.02, Fisher's PLSD in both cases for values of both treatment groups combined; Table 2). Mean trough masses also increased during the course of the experiment, but only the difference between the first and third years was statistically significant (P < 0. 03, Fisher's PLSD; Table 2). Peak-to-trough amplitude in body mass was also lower in the first than third cycles (P < 0.01, Fisher's PLSD).

Effects of EB during the mass-loss phase. Median duration of the mass-loss phase was shorter by 12 days in EB- than blank-treated groups, but this difference fell short of significance (P = 0.10; Table 3).

The accelerated body mass loss between peak 1 and trough 1 resulted in earlier median trough and peak dates on all subsequent cycles after capsules had been removed, e.g., trough 2 was advanced by 66 days in EB squirrels (P < 0.01) and peak 3 by 64 days (P = 0.08; Table 3).

Although the median  values for peak 1-2 and trough 1-2 were shorter for EB than blank squirrels, by 61 and 39 days, respectively, the differences were not significant (P = 0.12 in both cases).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Estradiol treatment affected the waveform of the circannual cycle of male ground squirrels; gain in body mass was significantly decelerated when estradiol treatment occurred during the mass-gain portion of the annual cycle. Mass loss was accelerated by estradiol treatment, but dose reduction to prevent catastrophic mass loss probably contributed to a smaller, statistically insignificant effect during the mass-loss portion of the annual cycle. Estradiol treatment also affected the  of the annual cycle in males; treatment during the mass-gain and mass-loss phases, respectively, lengthened and shortened  of the annual cycle during which treatment occurred, although, in the latter case, the effect only approached statistical significance (Tables 2 and 3).

We propose that the main effects of estrogen treatment on the body mass cycle are to decelerate mass gain and accelerate mass loss, both of which appear to be due to shifts in energy balance that increase expenditures and/or decrease uptake and storage. Estrogens increase locomotor activity in female mice and rats but apparently not in golden-mantled ground squirrels (see Ref. 15 for review) and affect fat metabolism in female ground squirrels (5); the efficiency of energy utilization also may be under the control of estrogens (14, 25). The present experiment does not directly address the question of whether estradiol's effects on the circannual mechanism are direct or indirect (i.e., secondary to direct effects on body mass). Results of other experiments, however, suggest that body mass itself does not influence the body mass cycle. Mrosovsky (21) and Heller and Poulson (9) subjected ground squirrels to food deprivation that induced weight decreases greater than those resulting from EB treatment in the present experiment but found that these body mass changes had little or no effect on the  of the ensuing annual cycle. Body mass reductions achieved by lipectomy (4) also failed to provoke a change in the waveform of the CAR. Finally, changes in the length of the mass-gain or mass-loss phase were associated with long-term changes in the CAR, including changes in  of the entire annual cycle and/or corresponding phase delays or advances of subsequent phase markers; this, together with the results of other experiments, suggests that EB treatment influences the underlying circannual oscillator as well as the physiological functions ("hands of the clock") controlled by this oscillator.

Similarly, estrogen treatment of female ground squirrels significantly lengthened the mass-gain phase and shortened the mass-loss phase, with corresponding lengthening and shortening, respectively, of  of the annual body mass cycle that included hormone treatment (14). However, female squirrels treated with EB during the mass-gain phase showed only insignificant phase delays in the peak a full annual cycle after the termination of treatment (Table 3); data on the corresponding phase marker for females treated during the mass-loss phase were not subjected to statistical analysis. Because later cycles were not monitored for females, the two sexes cannot be compared on this measure. Although the CAR of females appears to be more responsive than that of males during EB treatment (Table 4), the effects of estradiol on  of the annual cycle that included the treatment phase are greater in males than in females (Table 4). A comparison of the two studies permits general conclusions about effects of estradiol on CARs, but, because of unavoidable procedural differences, we cannot discount subtle sex differences. For example, the treatments were not directly comparable for EB-peak males studied in experiment 2 and females investigated by Lee and Zucker (14); because several animals in the earlier experiment died as a consequence of EB treatment, the health of the males was safeguarded by reducing the dose of EB to one-quarter of that given to females. Also, despite an initially large cohort of squirrels in each group, final sample sizes were six to seven per group, and, for this reason, some of the conclusions must be tempered. Finally, direct comparisons also are limited by the absence of data on hormone concentrations produced by the implants in the females treated with EB capsules (14).

Sexually differentiated responses to estradiol in adult rats and hamsters are well-documented for circadian rhythms (reviewed in Ref. 19) and have been demonstrated also for several seasonal rhythmic phenomena in sheep (e.g., Refs. 10 and 17) and squirrels (23). The present study suggests that, if sex differences exist among adult squirrels in the responsiveness of their CARs to estradiol, they are less robust than those described for circadian rhythms; e.g., the  of the circadian rhythm of female hamsters is shortened by estradiol treatment (20), but males are completely unresponsive to this treatment (27). Clearly, the CARs of male squirrels are responsive to estradiol; whether they are as responsive as those of females remains uncertain.

Additional studies will be necessary to establish whether there are subtle sex differences in responsiveness to estradiol in the circannual substrate of ground squirrels. Even in the absence of such differences, differences in patterns of estrogen secretion in the two sexes render it likely that endogenous estradiol at physiological concentrations plays a greater role in phasing the CAR of female than of male squirrels.