INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IN HOMEOTHERMIC ANIMALS, body core temperature (T_core) oscillates daily: it is higher in the active phase and lower in the inactive phase in diurnal and nocturnal animals. It has been reported that the amplitude of the T_core rhythm is largely influenced by feeding condition and/or nutritional state in rats and pigeons (7, 18, 24, 32). In rats, T_core gradually decreased during 4 days of fasting selectively in the light (inactive) phase, whereas T_core in the dark (active) phase was well maintained at the control level (32). The decrease in T_core may be an advantage for survival in decreasing heat transfer from the body to the environment, i.e., decreased energy loss. In contrast, maintenance of T_core in the dark phase may be necessary to preserve physical activity, e.g., seeking and acquiring food. However, the mechanism for the change in the T_core rhythm during fasting is not fully understood.

Homeothermic animals regulate T_core by heat production and heat loss mechanisms, although it remains unknown whether the regulation also applies to the fasting state. It is well known that, in rodents and pigeons, fasting and/or food restriction decreases metabolic heat production (3, 7, 12, 15, 18, 20, 24, 30, 31). In addition, thermal conductance from the body core to the environment, i.e., heat loss, has been reported to decrease during fasting (11, 17, 27), which indicates suppression of the heat loss mechanism. Thus suppression of the heat loss mechanism may compensate for the lowered heat production during fasting, maintaining T_core in the dark phase at the normal level. In contrast, such a response may not work in the light phase, resulting in a greater reduction of T_core. However, there is no direct evidence to support this speculation.

Animals have several mechanisms to change the efficiency of heat loss in a short period: 1) skin circulation, 2) fur and feather positioning, affecting heat insulation, and 3) postural adjustment, altering effective body surface area. In rats, the tail is a crucial site for regulation of heat loss, with its physiological and anatomic characteristics, i.e., high density of arteriovenous anastomosis (6), no fur, and a greater surface-to-volume ratio than elsewhere in the body. Young and Dawson (33) reported that rats could dissipate ~25% of basal metabolic heat production through the tail by changing tail blood flow. In addition, tail blood flow is controlled by sympathetic nerve activity (14, 19). It is known that fasting modulates sympathetic nerve activity in some organs such as the liver, adrenal gland, and adipose tissue (2, 23, 34-36). Thus we surmised that fasting also has an influence on tail blood flow.

The purpose of the present study is to examine how the daily change of T_core during fasting is generated in rats. Thus we assessed the effect of 3 days of fasting on metabolic heat production in free-moving rats. Regional body surface temperature as an index of heat loss was also evaluated by infrared thermography. We hypothesized that 1) 3 days of fasting changed the daily oscillation of T_core by affecting heat production and loss mechanisms and 2) tail blood flow was closely associated with suppression of the heat loss mechanism during fasting.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Adult male crj-Wistar rats (n = 17, Charles River Japan, Osaka, Japan) were individually housed in a cage (45 × 25 × 20 cm) at an ambient temperature (T_a) of 23°C in a 12:12-h light-dark cycle (lights on at 0700, 200 lx in the light and 0 lx in the dark). All experimental protocols were approved by the Institutional Animal Care and Use Committee of the School of Allied Health Sciences, Osaka University Faculty of Medicine.

Surgery. For the measurements of T_core and locomotor activity, a radio transmitter (15 × 30 × 8 mm; Physiotel TA10TA-F40, DataScience, St. Paul, MN) was placed in the abdominal cavity of each rat under general anesthesia with pentobarbital sodium (5 mg/100 g body wt ip). The rats were allowed to recover for >3 wk before the experiments.

Experiment 1: measurement of metabolic rate. In six rats, oxygen consumption rate (O₂, measured by indirect open-circuit calorimetry) was measured for 24 h on the last days of the control, fasting, and recovery periods. At 1500, each rat was transferred to a Plexiglas chamber (35 × 20 × 20 cm) attached to an airflow system with a constant flow rate of 2.0 l/min. The rats were kept in the chamber for two 24-h periods before the experiment to minimize stress responses, such as increases in T_core and metabolic rate. The difference in oxygen tension between the room air and the air that passed through the chamber was determined every 10 s with an electrochemical oxygen analyzer (model LCJ-700, TORAY, Tokyo, Japan). O₂ was calculated as the product of the difference in oxygen tension and airflow rate. The values were divided by the 0.75 power of body weight (Brody-Kleiber formula) and corrected to STPD condition. T_a in the chamber was continuously monitored with a thermistor and was kept at 23.0 ± 0.2°C.

Experiment 2: measurement of body surface temperature. In another group of five rats, the body surface temperature was determined by thermography (model LAIRD-S270A, Nikon, Tokyo, Japan). Each rat was placed in a plastic box (40 cm long × 30 cm wide × 70 cm high) at 1500 on the last day of the control, fasting, and recovery periods. The box was well ventilated and had an open window on the top (25 × 20 cm). T_a in the box was monitored with a thermistor and was kept at 23.0 ± 0.2°C during the experimental period. An infrared charge-coupled device camera was placed 20 cm above the box, and a digital image was taken every 5 min through the window. The surface temperatures of five areas [head, middle parts of the upper and lower back, and proximal and distal parts of the tail (one-third of the length of the tail from the root and tip, respectively)] were determined using an image analyzer program (FAIRIS, Nikon). The temperatures of the head and back were averaged as trunk temperature (T_trunk), and those of the tail as tail temperature (T_tail). T_tail was measured only when the previous and subsequent tail images could been seen: the data were excluded when the tail was under the body within the 15-min period. Those values were calibrated on the basis of our preliminary data, because the radiation rate of infrared rays differs among body sites; the actual values of skin temperature at the skin sites were measured with thermocouples in a warm or cool environment, and the values were compared with those obtained by thermography. The accuracy of the measurement was ±0.2°C.

Blood sampling. A blood sample was taken in another group of six rats at 1700 on the last day of each feeding condition; after local anesthetic (2% lidocaine gel) was applied to the tail surface, 150 µl blood were collected in microcapillary tubes through a tiny cut of the tail. Hematocrit (microcentrifuge), blood glucose (colorimetry; Arkray, Tokyo, Japan), and plasma concentrations of total protein (refractometry; Atago, Tokyo, Japan) and triglyceride (colorimetry; Wako, Osaka, Japan) were determined from the samples.

Statistics. Differences in mean values among the three feeding conditions were evaluated by ANOVA for repeated measures. A post hoc test at a specific time point was conducted by the Newman-Keuls procedure. The null hypothesis was rejected at P < 0.05. Regression analysis was conducted by the standard least squares method. Difference in slopes and intercepts for the regression lines was assessed by Student's t-test. The circadian rhythm of each physiological parameter was analyzed by fitting a cosine curve [cosinor rhythmometry (13)], and its mesor (mean level), amplitude, and acrophase (the time when the rhythm peaks) were estimated. All variables were averaged over 30 min and are shown as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figure 1 illustrates T_core, counts of locomotor activity, and O₂ on the last days in the control, fasting, and recovery periods in experiment 1. There were no significant differences in the three parameters between the control and recovery conditions. In addition, the locomotor activity counts in the fasting condition remained unchanged from the control. The arithmetic means and medians of the three parameters were always higher (P < 0.05) in the dark phase than in the light phase. For example, in the control condition, 1) the arithmetic means in the dark and light phases were 37.8 ± 0.1 and 37.1 ± 0.1°C in T_core, 24.37 ± 1.06 and 18.91 ± 0.64 ml · min¹ · kg body wt^0.75 in O₂, and 7.3 ± 1.4 and 2.4 ± 0.5 arbitrary units in locomotor activity, respectively, and 2) the medians in the dark and light phases were 37.8 ± 0.1 and 37.1 ± 0.1°C in T_core, 24.51 ± 1.07 and 18.10 ± 0.56 ml · min¹ · kg body wt^0.75 in O₂, and 6.9 ± 1.3 and 1.3 ± 0.4 units in locomotor activity, respectively.

View larger version (43K): [in this window] [in a new window]   Fig. 1.   Body core temperature (Tcore; A), counts of locomotor activity (B), and oxygen consumption rate (O2, C) during fed control, day 3 of fasting, and day 4 of recovery. Stippled area, dark phase (1900-0700). Each point is average for 30 min. Values are means ± SE for 6 rats. * Significantly different from fed control, P < 0.05.

In the fasting condition, T_core in the dark phase remained unchanged from the control. However, T_core was lower than in the control condition at 0700-1300 (P < 0.05; Fig. 1A). The reduction of T_core was significant soon after light onset, reaching its nadir at 0800 (36.0 ± 0.2°C, in contrast to control of 37.1 ± 0.1°C). The difference in T_core between the control and fasting conditions was 0.2 ± 0.1, 0.7 ± 0.1, and 0.4 ± 0.1°C at 0400-0700, 0700-1000, and 1000-1300, respectively, with the greatest difference at 0700-1000 (P < 0.05).

O₂ was lower than the control condition throughout the fasting day (P < 0.05; Fig. 1C). The arithmetic mean and median were 16.24 ± 0.69 and 16.05 ± 0.58 ml · min¹ · kg body wt^0.75 in the dark phase and 14.00 ± 0.41 and 13.32 ± 0.36 ml · min¹ · kg body wt^0.75 in the light phase, respectively. The reduction of O₂ after light onset was attenuated: the difference between the control and fasting conditions was lower at 0700-1000 than at 0400-0700 (P < 0.05, 5.24 ± 0.63 and 9.67 ± 0.93 ml · min¹ · kg body wt^0.75, respectively).

Table 1 summarizes the analysis by cosinor rhythmometry for the circadian changes in T_core, locomotor activity, and O₂. We assumed that the circadian period was constantly 24 h (the value in a normal 12:12-h light-dark cycle), because those parameters were not measured for successive days in the same conditions. The regression coefficients of the cosine curves (data not shown) were significant (P < 0.05) in any feeding condition, and the acrophases were in the middle of the dark phase. In the fasting condition, the amplitude increased in T_core and decreased in O₂ with reductions of both mesors, and both acrophases advanced by 1.8 h from the controls. However, there were no changes in the three parameters in locomotor activity.

View this table: [in this window] [in a new window]   Table 1.   Analysis of cosinor rhythmometry for daily changes in core temperature, O2, and locomotor activity

Figure 2 shows thermograms for one rat in the control condition and on day 3 of fasting in experiment 2. The thermograms are illustrated as 256-grade pseudocolor images. The tail temperature decreased to environmental temperature in the dark phase on the fasting day. The tail images could always be distinguished from the background by the difference in radiation rate: the signals from the background were always weaker than those from the tail.

View larger version (104K): [in this window] [in a new window]   Fig. 2.   Thermograms in fed and fasting conditions. Thermograms are shown as 256-grade pseudocolor images every 4 h in fed control condition and on day 3 of fasting. Lights were off between 1900 and 0700.

Figure 3 illustrates T_tail and T_trunk in the three feeding conditions in experiment 2. T_core was not different from that in experiment 1, so those data were not presented. Rats occasionally hid their tails under their bodies, and these T_tail data were not recorded or were excluded (e.g., on the control day, 4 ± 2 and 11 ± 4 of 144 data in the dark and light phases, respectively, and on the fasting day, 11 ± 2 and 22 ± 1 in the dark and light phases, respectively). There were no differences in T_tail and T_trunk between the control and recovery conditions. In the control condition, the arithmetic means and medians of T_tail and T_trunk were higher (P < 0.05) in the dark phase (30.7 ± 0.5 and 30.7 ± 0.1°C in T_tail and 28.8 ± 0.1 and 28.8 ± 0.1°C in T_trunk, respectively) than in the light phase (29.4 ± 0.1 and 29.4 ± 0.1°C in T_tail and 28.2 ± 0.1 and 28.2 ± 0.1°C in T_trunk, respectively).

View larger version (41K): [in this window] [in a new window]   Fig. 3.   Surface temperatures of tail (A) and trunk (B) determined by thermography in fed control condition, on day 3 of fasting, and day 4 of recovery. Ttail is average of surface temperatures at one-third of the length of the tail from the root and the tip. Ttrunk is average of surface temperatures of the head and middle parts of the upper and lower back. Each point is average for 30 min. Values are means ± SE for 5 rats. * Significantly different from fed control, P < 0.05.

T_tail was lower than the control throughout the fasting day (P < 0.05; Fig. 3A); however, T_trunk was lower (P < 0.05) than in the control condition only at 0600-1000 (Fig. 3B). In contrast to the control condition, T_tail in the fasting condition increased after light onset and remained at a higher level (P < 0.05) than at 0400 for 4 h. T_tail at 2300, 0400, 0700, and 1300 was 30.8 ± 0.9, 30.1 ± 0.9, 30.1 ± 0.1, and 29.3 ± 0.6°C in the control condition and 23.1 ± 0.1, 24.5 ± 0.4, 27.5 ± 0.1, and 25.0 ± 0.1°C in the fasting condition, respectively. The arithmetic mean and median were lower (P < 0.05) in the dark phase (23.9 ± 0.2 and 24.0 ± 0.1°C, respectively) than in the light phase (25.2 ± 0.2 and 25.0 ± 0.2°C, respectively).

The regression coefficients of the fitted cosine curves for daily changes of T_tail and T_trunk (data not shown) were significant in any feeding condition (P < 0.05, r = 0.41-0.72 in T_tail and 0.32-0.82 in T_trunk). In the fasting condition, the mesors of both temperatures decreased from the control (P < 0.05, from 29.1 ± 0.2 to 23.8 ± 0.2°C in T_tail and from 28.8 ± 0.1 to 27.1 ± 0.1°C in T_trunk). Moreover, the amplitude of T_tail increased from the control (P < 0.05, 0.9 ± 0.1°C) by 0.6°C; however, the amplitude of T_trunk remained unchanged (0.4 ± 0.1°C). The acrophase of T_tail was delayed by ~12 h from the control [P < 0.05, from zeitgeber time (ZT, ZT 0 = 0700) 17.8 ± 0.3 to 4.4 ± 0.3 h]; however, the acrophase of T_trunk advanced (P < 0.05) from ZT 18.8 ± 0.1 to 15.8 ± 0.1 h.

Figure 4 shows the relation between O₂ and T_core in the control and fasting conditions in experiment 1. The relation was linear in each feeding condition (P < 0.05, r = 0.96 and 0.75 in the control and fasting conditions, respectively). However, the regression slope became greater (P < 0.05) in the fasting condition: T_core at the given O₂ shifted above the control condition in the dark phase; however, the shift was less in the light phase. As time passed, the circled points denoting the period during which T_core decreased moved leftward along the regression line of the fasting condition, and the shift above the control line was totally cancelled at the end.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Relation between O2 and Tcore in fed control and fasting conditions in dark and light phases. Regression line and its equation for each condition are also presented. Each point is average over 30 min for 6 rats. Data points from 0530 to 0800, the period during which Tcore decreased, are individually circled and connected by open arrows in the order of time. Each point is average over 30 min for 5 rats.

Figure 5 illustrates the relation between T_core and T_tail and T_trunk in the control and fasting conditions in experiment 2. T_core and T_tail had a positive and linear correlation in the control condition (P < 0.05, r = 0.72; Fig. 5A); however, the correlation became negative on day 3 of fasting (P < 0.05, r = 0.64). The relation between T_core and T_trunk (Fig. 5B) was positive and linear in the control and fasting conditions (P < 0.05, r = 0.73 and 0.53, respectively), although the regression line shifted downward in the fasting condition. At the beginning of the period denoted by the circled points during which T_core decreased, T_core-T_tail shifted upward and then moved leftward along the control regression line.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Relation between Tcore and averaged Ttail (A) and Ttrunk (B) in fed control and fasting conditions in dark and light phases. Data points from 0530 to 0800, the period during which Tcore decreased, are individually circled and connected by open arrows in the order of time. Each point is average over 30 min for 5 rats.

Figure 6 shows the relation between the locomotor activity counts and O₂ in the control and fasting conditions in experiment 1. There was a linear (P < 0.05) relation between the two parameters in each condition. There were no differences among the regression slopes, except in the light phase on the fasting day. However, the activity in the light phase on the fasting day did not increase as much as in the other condition, mostly <2.0 units. The intercept (estimated O₂ at zero activity) in the control condition was higher in the dark phase than in the light phase (P < 0.05, 21.20 and 17.89 ml · min¹ · kg body wt^0.75, respectively). Moreover, those values were higher than in the fasting condition (P < 0.05, 14.64 and 13.87 ml · min¹ · kg body wt^0.75 in the dark and light phase, respectively).

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Relation between averaged locomotor activity and O2 in fed control and fasting conditions in dark and light phases. Each point is average over 30 min for 6 rats.

Table 2 shows body weight, hematocrit, and concentrations of blood glucose and plasma total protein and triglyceride in six rats. Hematocrit in the fasting condition was greater (P < 0.05) than in the control condition. The other variables decreased (P < 0.05) on day 3 of fasting; however, they returned to the control level on day 4 of the recovery, except for the augmented value of triglyceride.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

To clarify the mechanism involved in the reduction of T_core in fasted rats, which occurs selectively in the inactive phase, we assessed the daily changes in T_core and O₂ in fed and fasting conditions. In addition, regional body surface temperature as an index of heat loss was assessed by thermography. In fed conditions, T_core, O₂, T_trunk, and T_tail showed a clear daily oscillation: higher in the dark phase and lower in the light phase. However, in the fasting condition, O₂ decreased from the control condition in the dark and light phases, and its circadian rhythm amplitude was reduced. Moreover, T_tail decreased from the control greatly in the light phase of the fasting period.

Homeothermic animals regulate T_core by heat production and loss mechanisms. In the fed condition, T_core is highly correlated with O₂, i.e., metabolic heat production (Fig. 4). On the assumption that heat balance was equilibrated and evaporative heat loss was small enough, the thermal conductance from the body core to the environment was calculated as O₂/(T_core  T_a) (9). The value was higher in the dark phase than in the light phase (P < 0.05, 1.65 and 1.28 ml · min¹ · kg body wt^0.75 · °C¹, respectively). The result indicates that heat loss mechanisms were activated in the dark phase. Thus, in the control condition, metabolic heat production was the factor causing T_core to be higher in the dark phase than in the light phase.

Several studies have shown that fasting decreases the metabolic rate (3, 7, 12, 15, 18, 20, 24, 30, 31). In the present study, O₂ was lower in fasted rats during the active and inactive periods (Fig. 1C, Table 1). Despite the reduction of O₂, i.e., heat production, T_core in the dark phase on the fasting day was well maintained. The linear relation between O₂ and T_core (Fig. 4) in the fasting condition indicates that O₂ was a factor determining T_core as in the control condition. However, the increase in T_core at the given O₂ in the dark phase may suggest that T_core was maintained mainly by the suppression of heat loss mechanisms. In support of this speculation, the assumed thermal conductance in the dark phase of the fasting period decreased from the control value (P < 0.05, 1.10 and 1.65 ml·min¹ · kg body wt ^0.75·°C¹, respectively).

T_core in the light phase on the fasting day decreased strongly from the control for 3 h after light onset. However, in the fasted animals, the reduction in O₂ during the period was less than in control condition, and the circadian rhythm amplitude decreased (Fig. 1C, Table 1). Thus we suppose that heat production was not a dominant factor generating the T_core rhythm in the fasting condition, in contrast to the control condition. The regression analysis of T_core and O₂ (Fig. 4) may show that the suppression of heat loss mechanisms was attenuated in the light phase on the fasting day compared with that in the dark phase. Moreover, the suppression was completely abolished at the beginning of the light phase, which may be a mechanism for the significant reduction of T_core at that time. Thus it is supposed that heat loss mechanisms primarily determine the T_core rhythm in the fasting condition.

Body surface temperature reflects the magnitude of heat loss. At T_a below the thermoneutral zone, i.e., 28-32°C in Wistar rats, nonevaporative processes dissipate body heat (18, 35). The change in tail blood flow is an important process regulating heat loss in rats (33). T_tail in the dark phase on the fasting day decreased to the level of T_a (23°C) and was lower than that in the light phase (Fig. 3A). The regression analysis for T_core and T_tail (Fig. 5A) may indicate that tail blood flow increased with the rise in T_core in the control condition; the increase in T_tail in response to the increase in T_core was >1 (°C/°C). In the fasting condition, T_tail at the given T_core decreased; this decrease was greater in the dark phase than in the light phase. Although the results strongly suggest an attenuation of tail blood flow, it is notable that the attenuation was blunted at the beginning of the light phase (circled points in Fig. 5A), at which T_core greatly decreased. In contrast, the attenuation of the skin blood flow of the trunk may have been similar between the dark and light phases (Fig. 5B). Thus it is believed that tail blood flow was a key process determining heat loss from the body in the fasting condition, which was the factor regulating the T_core rhythm.

Despite the fact that locomotor activity was unchanged from the control condition (Fig. 1B, Table 1), O₂ greatly decreased in the fasting condition. In addition, the decrease was greater in the dark phase than in the light phase (P < 0.05, 8.13 ± 0.13 and 4.78 ± 0.71 ml · min¹ · kg body wt^0.75, respectively). The regression analysis for activity and O₂ (Fig. 6) also indicated that the estimated O₂ at zero activity decreased in the fasting condition. Furthermore, the reduction in the estimated O₂ at zero activity could explain ~80% of the decrease in measured O₂ in the dark and light phases. These results may show that the reduction of resting metabolic rate contributed to the decrease in O₂. One possible factor for the reduction of resting metabolic rate is a lack of food intake. Food intake itself induces an increase in metabolism, known as postprandial-derived thermogenesis and/or diet-induced thermogenesis (21, 22, 29, 30). Another possible factor is a decrease in energy stores and/or circulating nutrients. Recent studies have clarified that elevations in leptin and/or insulin in response to increases in body fat mass and/or blood glucose facilitate energy expenditure (3, 5, 16, 26). The 3-day fast induced a 15% reduction of body weight. Furthermore, the plasma concentrations of circulating nutrients decreased on day 3 of fasting, despite dehydration, as estimated by the increase in hematocrit (Table 1). Thus it is speculated that the reductions in energy stores and/or circulating nutrients were factors decreasing O₂ in the fasting condition. We did not assess the daily changes in plasma nutrients, because blood sampling greatly disturbs the circadian rhythm of T_core. However, the daily rhythm of plasma nutrients might be associated with the metabolic rhythm.

The significant difference in the regulation of tail blood flow between the dark and light phases may indicate that the circadian system such as the suprachiasmatic nucleus, which is known as a central oscillator in various physiological functions, was involved in the mechanism (4, 8, 10, 25, 28). Liu et al. (10) reported that T_core remained unchanged during 4 days of fasting in suprachiasmatic nucleus-lesioned rats. In addition, Alföldi et al. (1) reported that T_core decreased and T_tail increased after sleep onset and during non-rapid eye movement sleep. The result might suggest that changes in pattern and/or amount of sleep are associated with changes in T_core and T_tail in the daytime of the fasting period. However, the locomotor activity rhythm was not influenced by fasting (Fig. 1, Table 1). Although we did not assess sleep itself, it is unlikely that the sleep-wake cycle was involved in the circadian rhythm of T_tail to any great extent.

In summary, fasting alters thermoregulatory mechanisms involved in heat production and loss in rats. Metabolic heat production decreased in the dark and light phases of the fasting period. However, T_core in the dark phase was well maintained by a suppression of heat loss, which was closely related to the decrease in tail blood flow. In contrast, the decrease in tail blood flow was abolished or attenuated in the light phase of the fasting period, which could be a mechanism involved in the augmented reduction of T_core.

Perspectives