It is reported that in an environment of thermo-neutral range [1], where autonomic thermoregulatory responses are minimum, the estrus cycle affects circadian body temperature (T_b) rhythm in women [2-9] and in female rats [10-12]. In women, the change of the rhythm is characterized as the increased mean T_b and decreased circadian amplitude in the luteal phase [7,12]. The estrus cycle in female rats generally lasts 4-5 days, consisting of four phases; two days of diestrus, followed by the proestrus and estrus phases. In the proestrus phase, the mean T_b in the light phase decreases [12]. Moreover, in the proestrus phase, the tail surface temperature (Ttail) decreases in the dark phase [13], and spontaneous activity increases [14,15], compared with that in the other phases. The peak in the activity appears earlier in the proestrus rather than on the first day of diestrus [11]. An increase in the mean, peak, and nadir of T_b and a decrease in the amplitude of T_b were observed in the luteal phase, which shows a higher progesterone level compared with that in the follicular phase in women [5]. However, the mechanism involved in the changes of the rhythms remains unclear.

It has been speculated that progesterone, the level of which changes with the estrus cycle, affects the circadian T_b rhythm [2,5,16] in a thermoneutral environment. Progesterone administration suppressed activity in female rats [17]. In contrast to the influence of progesterone on the circadian T_b rhythm, the role of estradiol (E₂) still remains controversial. E₂ did not alter the circadian T_b rhythm in rats [13] and in women [2,3,5]; however, some reports showed that E₂ decreased the Ttail in the dark phase in rats [13]. E₂ influences the circadian rhythm of activity in mice [18-21] and hamsters [22,23] but not in rats [14,24]. Thus, we hypothesized that E₂ might modulate the circadian rhythm of T_b and activity.

In the present study, we compared T_b and activity between sham operated rats, ovariectomized rats, and rats administered E₂ externally, to determine the effect of E₂ on the circadian rhythm of T_b and activity.

Animals

Female Wistar rats (n=24; 224 ± 2 g; age, 8 weeks; Takasugi, Saitama, Japan) were used in the present study. They were individually housed in cages (45 cm × 25 cm × 20 cm) at an ambient temperature (T_a) of 27 ± 0.5°C with a lighting schedule of 12-h light and 12-h complete darkness (lights on at 0700 h, 300 lux at their eye level). All experimental protocols were approved by the Institutional Animal Care and Use Committee of Waseda University.

Surgery

Under inhalation anesthesia with 2% sevoflurane, a radio transmitter (26 mm × 8 mm, 2.2 g; PDT-4000 HR E-Mitter^®, Starr Life Sciences Corp., Oakmont, PA, USA) for measuring T_b and spontaneous activity was placed in the peritoneal cavity. Bilateral ovariectomy or sham (SH group; n=8) surgery was conducted through a dorsal skin incision. Silastic plates (5 mm × 25 mm × 15 mm; Silpot134, 0.2 ml; Silpot catalyst 184, 20 μl; Dow Corning Toray Co., Ltd, Tokyo, Japan) with and without 17β-estradiol (Sigma, St. Louis, MO, USA) were prepared. One plate was placed under the right dorsal skin. Eight rats were implanted with an E₂ plate (E₂(+) group) and eight other animals had a control plate without E₂ (E₂(-) group). The placement of the estradiol-containing plate results in a constant E₂ level in the plasma at least for 7 days [25]. We set a pharmacological level of E₂ to determine the definitive effect of E₂. Plasma E₂ concentration was kept at a high level in the E₂(+) group (1208 ± 85 pg/ml) for 26 days. After the surgery, the rats were injected sc with penicillin G (1,000 U, Meiji Pharmaceutical, Tokyo, Japan) to prevent post-surgical infection, and were placed in the cage at 27°C.

Experimental protocols

Before the surgery, body weight, food intake, and water intake were estimated at 0830-0900 for 5 days. Then, the entry to animal room was restricted for 2 weeks to avoid a time cue effect by the touches and sounds of experimenters during the T_b and activity measurements. After a non-contact period, the same measurements were again performed for 5 days. The signals from the radio transmitter were obtained through a receiver board ER-4000 Energizer/Receiver, Starr Life Sciences Corp.) Every 5 min, and were stored in a personal computer with a data-logging program (VitalView; Starr Life Sciences Corp.). The accuracy of the value of T_b was ± 0.1°C.

Statistics

The T_b and activity on each day were averaged every 30 min, and also for the light and dark phases and the entire day. The mean, nadir, and peak of the circadian T_b and activity rhythms and the difference between the maximum and minimum (amplitude) were estimated for each day of the 2-week measurement period, and the values were averaged. Differences in these values were assessed by two-way ANOVA with R language (R version 3.1.2, The R Foundation for Statistical Computing). Tukey-Kramer's test was used to identify the significant differences at specific time points of T_b and activity. The null hypothesis was rejected at the level of P<0.05.

The circadian T_b and activity rhythms in a day are shown in Figure 1A and 1B. T_b in the E₂(-) group was lower than that in the SH group at 0-2, 9-10, and 15-23 hours. T_b in the E₂(+) group was lower than that in the SH group at 2-4 and 9-23 hours (Figure 1A). T_b in E₂(+) was lower than that in E₂(-) at 2-16 and 21-24 hours. Activity was lower in the E₂(-) group than that in the SH group at 0-3 and 10-23 hours. Activity was lower in the E₂(+) group than that in the SH group at 7-8, 11-12, and 21-23 hours. Activity was higher in the E₂(+) group than that in the E₂(-) group at 1-3, 10-11, and 12-23 hours (Figure 1B).

Figure 1: Body temperature (T_b) and activity in a day (A and B) in the E₂(-), E₂(+), and SH groups. Values are mean ± SE (n=8 per group). Significant difference between E₂(-) and E₂(+) (*), E₂(-) and SH (#), and E₂(+) and SH (†), P<0.05.

Figure 2A and 2B show the relationship between T_b and activity during the light and dark phases. Table 1 shows the slope of the regression line between T_b and activity in the light and dark phases. The slope in the light phase was not different among groups; however in the dark phase, the slope in the E₂(-) group was the greatest among the three groups. There was no difference in the slope in the E₂(-) group between the light and dark phases; however, in the E₂(+) and SH groups, the slope was greater in the light phase than that in the dark phase.

Table 1: Slope of the regression line between body temperature (T_b) and activity in the light and dark phases in the E₂(-), E₂(+), and SH groups. Values are mean ± SE (n=8 per group). Significant difference between E₂(-) and E₂(+) (*), E₂(-) and SH (#), and light and dark phases (§), P<0.05.

Figure 2: Relationship between T_b and activity in the light (A) and dark (B) phases in the E₂(-), E₂(+), and SH groups. Values are mean ± SE (n=8 per group).

Table 2 shows the mean in the day, light, and dark phase, the amplitude, the nadir, and the peak of T_b (°C) and activity (counts). The mean in a day of T_b in SH was the greatest among all groups. The value was not different between the E₂(+) and E₂(-) groups. The mean in the light phase of T_b was not different among the groups. On the other hand, the mean in the dark phase and the amplitude of T_b in the SH group were greater than those in the E₂(-) group, and did not differ from those in the E₂(+) group. The values in the E₂(+) group were lower than those in the E₂(-) group. The nadir of T_b was not different among groups; however, the peak of T_b in the SH group was greatest among the groups; the value in the E₂(+) group was greater than that in the E₂(-) group. The mean of activity in the SH group was greater than that in the E₂(-) group, and was not different from that in the E₂(+) group. The mean in the light phase of activity in the SH group was the greatest among groups; however, the value was not different between the E₂(-) and E₂(+) groups. The mean in the dark phase of activity in the SH group was greater than that in the E₂(-) group, and was not different from that in the E₂(+) group. The value in the E₂(+) group was greater than that in the E₂(-) group. The peak and amplitude of activity in the SH group was not different from that in the E₂(-) group, and was greater than that in the E₂(+) group. The value in the E₂(+) group was greater than that in the E₂(-) group. The nadir of activity was not different among groups.

Table 2:　Mean in the day, light, and dark phase, amplitude, nadir, and peak of T_b (°C) and activity (counts) in the E₂(-), E₂(+), and SH groups. Values are mean ± SE (n=8 per group). Significant difference between E₂(-) and E₂(+) (*), E₂(-) and SH (#), E₂(+) and SH (†), and light and dark phases (§), P<0.05.

Table 3 shows the appearance time of the nadir and the peak of T_b and activity. The appearance time of the amplitude in T_b for the SH group was earlier than that in the E₂(-) group, and was not different from that in the E₂(+) group. The value in the E₂(+) group was earlier than that in the E₂(-) group. The appearance time of the nadir in T_b in the SH group was later than that in the E₂(-) group, and was not different from that in the E₂(+) group. The value in the E₂(+) group was later than that in the E₂(-) group. The appearance time of the peak in T_b in the SH group was earlier than that in the E₂(-) group, and was not different from that in the E₂(+) group. The value in the E₂(+) group was earlier than that in the E₂(-) group. The appearance time of the amplitude in activity in the SH group was not different from that in the E₂(-) and E₂(+) groups. The value in the E₂(+) group was earlier than that in the E₂(-) group. The appearance time of the nadir in activity was not different among groups. The appearance time of the peak in activity in the SH group was earlier than that in the E₂(-) group, and was not different from that in the E₂(+) group. The appearance time of the peak in activity in the E₂(+) group was earlier than that in the E₂(-) group.

Table 3: The amplitude, nadir, and peak of the appearance time of T_b and activity in the E₂(-), E₂(+), and SH groups. Values are mean ± SE (n=8 per group). Significant difference between E₂(-) and E₂(+) (*), E₂(-) and SH (#), E₂(+) and SH (†), P<0.05.

The body weight, food intake, and water intake are shown in Figure 3A-3C, respectively. Pre-surgery body weight was not different among the groups; however, post-surgery body weight in the SH group was greater than that in the E₂(-) group. The value in the E₂(+) group was lower than that in the E₂(-) group (Figure 3A). Pre- and post-surgery food intake and water intake were not different among groups (Figure 3B and 3C). Post-surgery food intake decreased from the pre-surgery food intake in all groups (Figure 3B). Water intake showed no difference between the pre- and post-surgery (Figure 3C). Plasma estradiol concentration in the E₂(+) group was higher than that in the E₂(-) and SH groups (62 ± 6 and 201 ± 60 pg/ml, respectively).

Figure 3: Body weight, food intake, and water intake (A, B, and C). Values are mean ± SE (n=8 per group). Significant difference between E₂(-) and E₂(+) (*), E₂(-) and SH (#), E₂(+) and SH (†), and pre-surgery vs. post-surgery (§), P<0.05.

The present study showed that E₂ decreased the mean of T_b, and increased the mean of activity in the dark phase. E₂ affected the timedependent relationship between T_b and activity. E₂ decreased the peak of T_b and activity, and delayed its appearance; however, E₂ did not influence the nadir of T_b and activity, but delayed their appearance. Thus, E₂ may modulate circadian rhythm of T_b and activity in female rats.

E₂ did not affect T_b in the light phase. This result coincides with those of a previous study showing that E₂ did not affect T_b in the light and dark phases [13]. Activity in the dark phase was decreased in ovariectomized mice [26] and rats [27]. The effect of E₂ on activity in the dark phase is controversial; E₂ increased it in the dark phase in mice [18,19], but did not influence it in the light and dark phases in rats [14,28]. The increased activity in the dark phase by E₂ coincided partly with the results of the previous studies.

In the E₂(+) group, the dependence of T_b on activity in the dark phase was lower than that in the light phase; however, it was not observed in the E₂(-) group (Table 1). Thus, E₂ may decrease the dependence of T_b on activity specifically in the dark phase. In literature, T_b is mainly determined by skin vasomotion and activity in rats at thermoneutral environments. E₂ decreased the Ttail in the dark phase at the thermoneutral range [14,28]. Vasoconstrictors like plasma adrenaline and arginine vasopressin have a circadian rhythm; higher in the light [29] and dark [30] phases, respectively. E₂ did not affect these [31,32]. Plasma renin activity and angiotensin related to the synthesis of vasoconstrictor angiotensin II were higher in the light phase in a day [33]. The influence of E₂ on angiotensin II is controversial as E₂ increased [34] or did not affect [35] it. Thus, it is difficult to assume that the vasoconstrictors induced the decreased Ttail by E₂ in the dark phase.

The circadian rhythm of peripheral vasodilators like endotheliumderived hyperpolarizing factor (EDHF) and endogenous hydrogen sulfide (H₂S) is unknown yet. In the light phase, E₂ affected EDHF in the mesenteric artery in female rats [36,37] and H₂S production in the mesenteric artery in ewes [38]. Thus, E₂ may contribute to peripheral vasodilation rather than vasoconstriction through EDHF and H₂S in the light phase; however, the effect in the dark phase is unclear. E₂ is considered to affect sympathetic nerves because plasma norepinephrine fluctuated along with the estrus cycle in females [39]. E₂ may modulate the skin vasomotor circadian rhythm through vasodilators and sympathetic nerves. In summary, it was speculated that the dependence of Tb on activity was decreased due to strong skin vasomotion in the dark phase in the E₂(+) group.

The effect of E₂ on circadian T_b and activity rhythm is controversial; E₂ did not influence the circadian rhythm of T_b in female rats [28] and women [40], but increased the activity and amplitude, delayed their peaks, and advanced their onset [18]. E₂ decreased the peak of T_b and activity, advanced its appearance, and delayed the appearance of the nadir in T_b. The circadian rhythm of activity in mice administered with E₂ was similar to that in mice administered with E₂ α and β receptors agonists [18]. It was speculated that the E₂ α and β receptors are related to the alteration in T_b by E₂ in the present study, though a mechanism yet unknown.

The decreased body weight by E₂ coincided with that observed in previous studies [41-43]. E₂ administration did not affect the food intake per body weight. E₂ administration in ovariectomized rats decreased the food intake in a day [44] and in the light phase [41,42,45]. Food intake in previous studies was not calculated per body weight. This may influence a difference in the results between the present and previous studies.

The estrus cycle affects water intake and drinking behavior. Drinking behavior in the estrus phase was lesser than that in other phases [46]. Isoprenaline-induced water intake in the proestrus and estrus phases was lower than that in the other phases [47]. Ovariectomy in rats resulted in increased water intake [48]. E₂ administration decreased the water intake in female rats after water deprivation [49]. E₂ administration prolonged the onset of drinking behavior in female rats administrated with NaCl solution [50]. Our result that E₂ decreased water intake coincided that of with previous studies. The result that E₂ did not influence water intake per body weight or food intake could not be compared with previous studies, because they were not calculated previously. The apparent water intake seemed to decrease due to the decreased body weight by E₂.

The present study showed that E₂ affected the time-dependent relationship between T_b and activity, and modulated the circadian rhythm of T_b and activity in female rats.

We are grateful to Prof. Kazuyoshi Tsutsui (Waseda University) and Prof. Keiko Morimoto (Nara Women’s University) for their advice for this research. The present research was partially supported by the Ministry of Education, Science, Sports, and Culture; Grant-in-Aids for Scientific Research (B), No. 20390066; Grant-in-Aid for Research Activity, No. 24800047; MEXT. KIBANKEISEI (2010); the Strategic Research Platforms for Private University; and Hayashi Memorial Foundation for Female Natural Scientists.

KN supervised the entire project. YU and KN designed the study and wrote the manuscript. YU, SM and KT performed experiments.