INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

THE SIBERIAN HAMSTER (Phodopus sungorus), a small (~35 g) rodent indigenous to the Russian and Mongolian steppes, faces severe winter conditions; during foraging it is exposed to air temperatures as low as 40°C (31). Physiological and behavioral mechanisms that appear to favor overwinter survival in this species include reductions in food intake and body mass, increases in fur length and density, and arrest of reproduction (7-9, 29). Under laboratory conditions of short-day length and low ambient temperature (T_a), Siberian hamsters also undergo bouts of torpor, a regulated reduction in body temperature (T_b) from 37 to as low as 12-15°C for 4-8 h. Torpor, which first is manifested ~6-10 wk after the onset of short-day lengths, occurs 2-4 days/wk, and persists for ~20 wk (6, 7). Torpor bouts for the most part are restricted to the rest-sleep phase of the daily cycle (i.e., subjective day for this nocturnal species) (11, 19, 23) and are associated with a large decrease in metabolic rate and thus a significant reduction in energy utilization (20). Reduced energy utilization is accompanied by a marked decrease in voluntary food intake (20, 21) that presumably decreases the time animals forage for food and consequently reduces energy spent on thermoregulation. Torpor also can be induced at any time of year, regardless of day length, by restricting food availability (10, 22). Torpor appears linked to decreased energy availability and provides energetic savings during times of energetic challenge.

The metabolic signals that provoke torpor and the mechanisms of metabolic fuel utilization during torpor are not well studied (14). Torpor is readily induced in Siberian hamsters treated with 2-deoxy-D-glucose (2-DG) (3, 4), a glucose analog that disrupts glycolysis (32). Because drugs that interfere with cellular oxidation of fatty acids, the other primary metabolic fuel, fail to induce torpor in this species (4), we suggested that onset of torpor may be triggered by reduced availability of glucose for cellular metabolism (3, 4). In deer mice (Peromyscus maniculatus) plasma glucose concentrations decrease before and during spontaneous torpor bouts, and fatty acids are elevated during daily torpor; fat is the predominant metabolic substrate used during and in the course of arousal from torpor (13, 14). Carbohydrate may be unavailable for metabolism during the initiation and expression of torpor (13). Plasma glucose also is reduced in dormice before entrance into hibernation (1). On the other hand, torpor in Syrian hamsters may be induced by blocking fatty acid but not glucose oxidation (24). Earlier studies reviewed by Musacchia and Deavers (12) indicated that hibernating ground squirrels do not maintain blood glucose concentrations within the normal euthermic range. Subsequent reports indicate that glucose utilization is reduced and blood concentrations elevated during hibernation (reviewed in Ref. 2). In contrast, Syrian and Turkish hamsters apparently regulate blood glucose concentrations during hibernation (12). Collectively, these data suggest that metabolic fuel availability may be crucial in the induction of torpor and the existence of significant interspecific variability in reliance on particular fuels.

It remains to be established that spontaneous torpor and torpor induced by food restriction in Siberian hamsters are due to reduced availability of glucose for glycolysis. Evidence in support of this conjecture is correlational and derived from pharmacological manipulations of unknown physiological relevance. If glucoprivation is causally connected to onset of torpor, then endogenous decreases in glucose concentrations should precede or coincide with the onset of spontaneous torpor. Because onset of spontaneous torpor in Siberian hamsters occurs in a circadian fashion (19, 23) at approximately the same time each day, blood glucose concentrations can be measured at the same clock (and circadian) times on days during which torpor is and is not being expressed. The association of reduced blood glucose concentrations with torpor would provide converging evidence that under physiological conditions torpor is triggered by glucoprivation (reduced glycolysis). Accordingly, blood glucose concentrations were monitored during spontaneous torpor. Additionally, it was determined whether the hypoglycemia promoted by insulin treatment would induce torpor; this also tests whether reduced blood glucose concentrations are a sufficient signal for triggering torpor.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Animals. Male and female Siberian hamsters were maintained from birth at 23 ± 2°C in a short photoperiod (10 h light/day, lights on 0800-1800) in experiment 1 or in a long photoperiod (16 h light/day, lights on 0200-1800) in experiments 2 and 3. The 10-h light photoperiod is associated with short-day, "winterlike" traits [e.g., decreased body mass and food intake and arrested reproduction (29)], and the 16-h light photoperiod is associated with long-day, "summerlike" traits. At the outset of experimentation, the animals were housed individually in polycarbonate tub cages with wood shavings for bedding and cotton for nesting (animals in cold only); food (Purina Rodent Chow no. 5015) and water were available ad libitum.

Telemetric recording of T_b. Radiotransmitters for the telemetric recording of T_b were surgically implanted in the peritoneal cavity under deep pentobarbital sodium anesthesia (80 mg/kg body mass). A single midline incision was made in the abdomen and the transmitter (model VM-FM or VM-FH-LT, Mini-Mitter, Sunriver, OR) was inserted. The wound was closed with sterile sutures and painted with 0.1% nitrofurazone ointment (Furacin). Postsurgical analgesia was provided by adding acetaminophen (Tylenol) and codeine phosphate to the drinking water (1% solution) for 2-3 days. At this time the animals were moved into an environmental chamber kept at 15°C, with an identical photoperiod. Receiver boards under the animals' cages captured radiofrequency signals, sampled every 10 min by an automated computer program (Dataquest). Torpor was defined as a decrease in T_b below 31°C for a minimum of 30 min (cf. Refs. 3 and 10).

Blood sampling and glucose analysis. Blood was collected from animals that were anesthetized with methoxyflurane (Metofane) vapors; a single drop of blood was obtained from the retro-orbital sinus using a capillary tube (4). Blood glucose concentrations were determined immediately on withdrawal using a one-touch blood glucose analyzing kit (Lifescan). All procedures were approved by the University of California Berkeley Animal Care and Use Committee.

Statistical analyses. Statistical comparisons between groups were made using ANOVA with Student-Newman-Keuls post hoc tests, where appropriate (SigmaStat, Jandel, San Rafael, CA). Statistical significance was assumed if P < 0.05 with two-tailed tests (note: SigmaStat reports actual P values up to < 0.001 for ANOVA, but only P < or > 0.05 for post hoc tests).

Experiment 1. Male and female hamsters (~equal numbers of each sex and ~10-12 wk of age) were moved to the cold chamber 2-3 days after surgery. T_b records were examined daily for occurrence of spontaneous torpor. Once spontaneous torpor had been manifested, blood samples were obtained from two groups of animals at several time points during entry into and rewarming from torpor: 1) first T_b <34°C (n = 10, group 2), 2) first T_b <31°C (n = 12, group 1), 3) first T_b >31°C (n = 7, group 2), and 4) first T_b >36°C after recovery from torpor (n = 5, group 2). A sample also was collected at the same time of day as the first T_b <31°C sample but while the animals were euthermic (i.e., on a day without torpor; n = 11, group 1). Only one blood sample was obtained during any given bout of torpor, and no animal was sampled more than twice. A single blood sample also was obtained from hamsters that never entered torpor in the cold chamber (n = 9) and others that remained in the short photoperiod outside the cold chamber at T_a of 23°C (n = 15). The latter samples were collected at approximately the same time of day as those from hamsters sampled at first T_b <31°C.

Experiment 2. To verify the effectiveness of insulin in producing hypoglycemia in this species, groups of adult female hamsters were administered 1 ml solution/kg body mass of short-acting porcine insulin (1, 5, or 10 U/kg ip, n = 10 for each group, Sigma) or saline (n = 10). Injections were given between 1200 and 1300. Glucose concentrations were determined from blood samples collected 30 and 90 min after injection.

Experiment 3. Experiment 2 was repeated but with long-acting insulin (Humilin L, recombinant DNA human insulin in zinc suspension, Lilly). First, the effects of insulin on blood glucose concentrations were verified in adult hamsters maintained at 23°C in a long photoperiod. Based on the outcome of experiment 2, a higher maximum dose of insulin was used. Groups of hamsters (n = 10 each) were treated with either saline or 5, 10, 20, 30, or 50 U/kg body mass insulin (1 ml solution/kg body mass between 0945 and 1130). Blood glucose was measured 90 and 180 min after treatment.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Experiment 1. Blood glucose concentrations decreased significantly during entry into torpor (F = 15.4, P < 0.001); by the time T_b had decreased to <34°C, glucose values had decreased by 19% compared with concentrations during euthermia (P < 0.05; Fig. 1). T_b remained below 31°C for 202 ± 37 min (range 50-360 min) based on nine torpor bouts selected at random from the records of seven animals. Glucose concentrations were 28% below euthermic values when T_b first declined to <31°C (P < 0.05; Fig. 1). Conversely, blood glucose concentrations increased as T_b increased during rewarming to >31°C and exceeded values characteristic of torpor when T_b first fell below 31 and 34°C during entry into torpor (P < 0.05 for each; Fig. 1) but were still lower than euthermic values (P < 0.05). Blood glucose concentrations were significantly below euthermic values for 3.5-4 h. With the recovery of euthermic T_b (>36°C), glucose concentrations did not differ from those recorded on a day when hamsters did not undergo torpor (P > 0.05). Blood glucose concentrations did not differ among euthermic hamsters in the several photoperiods or temperatures (P > 0.05; Fig. 2).

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Blood glucose concentrations of Siberian hamsters (n = 10/group) on a day without torpor (i.e., during euthermia; Euth), during entry into torpor as body temperature (Tb) first decreased below 34 and 31°C, and during rewarming from torpor as Tb first was >31 and >36°C. Values sharing a common letter did not differ significantly, and those with dissimilar letters differed significantly.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Blood glucose concentrations of hamsters housed at 23°C (Warm), torporous Siberian hamsters on a day without torpor (Euth), torpid hamsters after regaining euthermia (Rewarm), and hamsters kept at 15°C that never entered torpor (NTor).

Experiment 2. Exogenous insulin reduced blood glucose concentrations in a dose-dependent manner (F = 12.8, P < 0.001; Fig. 3A). The 10 U/kg insulin challenge resulted in glucose concentrations lower than those of all other groups (P < 0.05 for all comparisons). Hamsters treated with 5 U/kg insulin had significantly lower glucose values than those given saline or 1 U/kg insulin (P < 0.05 for both comparisons). This effect was evident 30 but not 90 min after injection (F = 1.3, P > 0.05; Fig. 3, A and B). Thirty minutes after treatment glucose concentrations of hamsters injected with 10 U/kg insulin were comparable to those seen during spontaneous entry into torpor (experiment 1), when T_b first decreased below 31°C (60.7 vs. 65.1 mg/dl, respectively).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Blood glucose concentrations 30 min (A) and 90 min (B) after treatment with saline (Sal), 1 (1.0 U), 5 (5.0 U), or 10 (10.0 U) U/kg of short-acting insulin. C: blood glucose concentrations were determined 180 min after treatment with saline, 5, 10, 20, 30, or 50 U/kg of long-acting insulin. Groups with dissimilar letters differed significantly.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Tb on day of and day after treatment of hamsters housed at 15°C; both animals were treated with the highest insulin dose (50.0 U/kg). * Time of injection.

Experiment 3. Treatment with long-acting insulin significantly reduced blood glucose concentrations (F = 15.3, P < 0.001; Fig. 3C), which were equivalently depressed 180 min after treatment with 20, 30, and 50 U/kg of insulin and represent significant decreases of 24, 33, and 32%, relative to hamsters treated with saline. Blood glucose at 180 min after 5 or 10 U/kg insulin treatment did not differ significantly from control values (P > 0.05); the three higher doses all induced significant decreases from saline values (P < 0.05 for each; Fig. 3C) but did not differ among each other. There also was a significant effect of time since insulin injection on blood glucose concentrations (F = 39.8, P < 0.001). The pattern of insulin effects on blood glucose at 90 min was similar to that observed after 180 min, but the effect after 20, 30, and 50 U/kg was somewhat reduced in magnitude (decreases in blood glucose of 25, 18, and 17%, respectively; not illustrated).

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

Blood glucose concentrations decrease substantially in Siberian hamsters during entrance into spontaneous torpor; glucose values were reduced by 19% by the time T_b first decreased below 34°C and 28% when T_b decreased below 31°C, the criterion for manifestation of torpor (10). Glucose concentrations increased progressively as hamsters rewarmed from torpor, such that blood glucose values were comparable to those of euthermic hamsters by the time T_b first exceeded 36°C. Similarly, blood glucose concentrations decrease in deer mice during torpor (14), and respiratory quotients indicate a shift to fat metabolism during torpor (13). Although blood glucose concentrations are decreased during expression of spontaneous torpor in Siberian hamsters, experimental induction of comparable hypoglycemia with insulin treatment failed to induce torpor. We suggest that hypoglycemia may be a consequence of torpor rather than a trigger for the hypometabolic state. It is unlikely that insulin causes hypoglycemia without attendant glucoprivation because insulin increases food intake, presumably as a consequence of glucoprivation.

2-DG reliably induces torpor in male and female Siberian hamsters kept in a low-temperature environment (3, 4). Because 2-DG disrupts cellular glycolysis (32) and is presumed to affect food intake (15, 16) and estrus (30) via its general glucoprivic action, we previously suggested that 2-DG also induces torpor by a similar mechanism. Carbohydrate availability previously also was implicated in control of hibernation in squirrels and hamsters (2, 12). Insulin-induced glucoprivation, however, failed to induce torpor in Siberian hamsters, thereby questioning whether 2-DG induces torpor by reducing the availability of glucose for cellular oxidation. Spontaneous torpor ordinarily is restricted to the light phase of the illumination cycle in Siberian hamsters (21, 23). Accordingly, we administered insulin during the light phase and nevertheless failed to observe torpor; it remains possible that insulin administration at other phases of the circadian cycle might be more effective, but there is no evidence to favor such an outcome. Other data also argue against a role of glucoprivation for induction of torpor. First, ablation of the area postrema of the brain stem, which is presumed to be critical for monitoring availability of metabolic fuels (16), eliminates the effects of 2-DG on food intake in rats (18) and estrus in Syrian hamsters (25) but does not block the torpor-inducing effects of 2-DG in Siberian hamsters, despite eliminating effects of 2-DG on food intake (H. H. Bae, J. L. Stamper, E. Heydorn, I. Zucker, and J. Dark, unpublished data). Additionally, 2-DG induces torpor in lactating hamsters, whereas food restriction fails to produce torpor in lactating dams but instead provokes pup cannibalization (28). These results suggest that although both food restriction and 2-DG result in glucoprivation they may exert their effects on torpor via different mechanisms.

It remains possible that 2-DG affects torpor via a direct, extreme glucoprivic action on one or several specific neural targets and that insulin treatments do not replicate this action of 2-DG. There are numerous reports of a mild hypothermic effect of systemic 2-DG treatment in laboratory rats, mice, and humans (e.g., Refs. 5, 15, and 27), and direct application of 2-DG to several hypothalamic areas can induce mild hypothermia (26). Injection of 2-DG into the lateral hypothalamus, ventromedial hypothalamic nucleus, anterior hypothalamic nucleus, posterior hypothalamic area, dorsomedial nucleus, and the dorsal premammillary nucleus of rats produces mild, transient (<4 h) reductions in T_b of ~1.5-3.0°C. The neural pathways implicated in intrahypothalamic hypothermia in rats may mediate the much more pronounced hypothermia observed in heterothermic species.

This experiment also leaves open an additional question: if the trigger for the onset of spontaneous torpor is not low glucose concentrations, what is it? It does not appear that metabolic fuel concentrations underlie the onset of spontaneous torpor. One clue comes from the observation that Siberian hamsters, in common with some other heterothermic mammals, enter daily torpor during the summer months if they are food restricted (22). Food restriction in some individuals first induces substantial weight loss before torpor is triggered (19), leaving open the possibility that torpor induction is not a product of metabolic fuel unavailability but rather dependent on direct feedback from body fat stores or perhaps an interaction of these factors, as for example between leptin and glucose oxidation. Central neural mechanisms apparently receive signals that convey adipose tissue availability. One or several direct adipocyte feedback mechanisms may be the critical signal(s) for torpor onset.