Torpor, a state characterized by a well-orchestrated reduction of metabolic rate and body temperature (Tb), is employed for energetic savings by organisms throughout the animal kingdom. The nucleotide AMP has recently been purported to be a primary regulator of torpor in mice, as circulating AMP is elevated in the fasted state, and administration of AMP causes severe hypothermia. However, we have found that the characteristics and parameters of the hypothermia induced by AMP were dissimilar to those of fasting-induced torpor bouts in mice. Although administration of AMP induced hypothermia (minimum Tb = 25.2 ± 0.6°C) similar to the depth of fasting-induced torpor (24.9 ± 1.5°C), ADP and ATP were equally effective in lowering Tb (minimum Tb: 24.8 ± 0.9°C and 24.0 ± 0.5°C, respectively). The maximum rate of Tb fall into hypothermia was significantly faster with injection of adenine nucleotides (AMP: −0.24 ± 0.03; ADP: −0.24 ± 0.02; ATP: −0.25 ± 0.03°C/min) than during fasting-induced torpor (−0.13 ± 0.02°C/min). Heart rate decreased from 755 ± 15 to 268 ± 17 beats per minute (bpm) within 1 min of AMP administration, unlike that observed during torpor (from 646 ± 21 to 294 ± 19 bpm over 35 min). Finally, the hypothermic effect of AMP was blunted with preadministration of an adenosine receptor blocker, suggesting that AMP action on Tb is mediated via the adenosine receptor. These data suggest that injection of adenine nucleotides into mice induces a reversible hypothermic state that is unrelated to fasting-induced torpor.
- core body temperature
- heart rate
in the face of cool ambient temperature (Ta) and diminished caloric availability, small mammals like mice can depress metabolic rate for enormous energetic savings. During the state of torpor, core Tb in mice can be as deep as 20°C and can last from as little as a few minutes up to ∼14 h (15, 17). Consistent with the fact that torpor in mice is dependent upon the lack of caloric availability, circulating hormones generated in the periphery that relay messages of both feeding status and fat availability play a critical role in determining 1) whether a mouse enters torpor and 2) the depth and duration of that torpor bout. For example, administration of ghrelin, a stomach- derived hormone that is normally released during fasting periods (2, 40), deepens the torpor bout via neuropeptide Y neurons within the arcuate nucleus of the hypothalamus (18). Leptin, a fat-derived hormone, cues a response opposite of ghrelin. Administration of leptin prevents deep bouts of fasting-induced torpor (13, 14, 16). Mice missing leptin (i.e., ob/ob mice) undergo unusually long bouts of torpor (21, 36) and can enter torpor even in the fed state (41). We have recently shown that sympathetically mediated depression of circulating leptin during fasting is a requisite for initiation of the torpor bout (37).
Within the hypothalamus, the arcuate nucleus plays a major role in regulating both food intake and metabolic rate (34), and as such, may be involved in mediating fasting-induced torpor. Ghrelin and leptin induce robust changes in both activity of neurons within the arcuate nucleus, as well as altered expression of neurotransmitters within this region (reviewed in Refs. 22 and 46). Mice that undergo perinatal injections of monosodium glutamate (MSG) develop major lesions within the arcuate nucleus, become obese (5, 18, 39), and fail to enter torpor upon fasting (18).
Zhang et al. (44) have recently suggested that the nucleotide AMP may play a role in torpor. They showed that 1) AMP levels are elevated in the plasma of fasted mice and 2) peripheral administration of AMP to mice induces a hypothermic state, interpreted as torpor. These authors suggested (44) that the elevated levels of AMP during a fast would activate AMP-activated protein kinase (AMPK), an important intracellular sensor of energy availability (20, 23), or other enzymes involved with flux through glycolysis and gluconeogenesis, ultimately resulting in a fall in core Tb (44). If delivery of exogenous AMP truly does evoke a torpor response, the ramifications to the field of hibernation and torpor would be enormous. We wished to directly test the hypothesis that AMP induces a torpor response in mice. Using adenine nucleotides, adenosine, and aminophylline, an adenosine receptor blocker, in telemetrically monitored mice, we now provide convincing evidence that, unlike the conclusions from previous research (44), administration of AMP does not appear to induce a state of true torpor.
MATERIALS AND METHODS
Female C57BL mice weighing 22–24 g (3 mo old) and pregnant C57BL mice were obtained from Jackson Laboratories (Bar Harbor, ME). Upon receipt, the mice were housed at 28–30°C and maintained on a 12:12-h light-dark cycle. Mice were maintained at this warm ambient temperature throughout the studies unless otherwise noted so as to minimize cold stress on the mice (38). The studies performed were approved by the local Institutional Animal Care and Use Committee.
ATP, ADP, AMP, inosine monophosphate (IMP), adenosine, and aminophylline were obtained from Sigma (St. Louis, MO). Sterile saline served as the solvent for all solutions.
Implantation of EKG and blood pressure telemeters.
Temperature telemeters or EKG telemeters (TAF20, ETAF20; Data Sciences International, St. Paul, MN) were implanted into the abdominal cavity as previously described (18, 33, 36). Both telemeter types weigh 3.8–3.9 g with a volume of 1.8 cubic centimeters. Temperature readings from the telemeters were calibrated at three temperatures, 20°C, 30°C, and 38°C. Mice were maintained on a heating pad for 24 h following the surgery and then housed individually at 28–30°C for 10 days to allow time for recovery.
Cardiovascular, temperature, and activity data collection.
Data from the telemeters were recorded at 500 Hz. Tb and activity levels were collected once per minute for 1 s from the temperature telemeters. HR, Tb, and activity levels were collected once per minute for 5 s from the EKG telemeters. Activity levels were measured using software from Data Sciences International, which calculates activity by quantifying the change in signal strength transmitted from the telemeter as the animal moves about the cage.
Experimental protocols—effect of nucleotides.
Mice with temperature telemeters (n = 6) were moved to an ambient temperature of 20°C. After 48 h at 20°C, the mice were fasted at the onset of the dark cycle to obtain torpor bouts for each mouse. Mice were refed 23 h after the onset of the fast. After a week of recovery from the fast at 28–30°C, these same mice were rehoused at 20°C, and randomly assigned to one of 4 dosages for each nucleotide (AMP, ADP, ATP): 40, 100, 400, and 800 mg/kg. Mice received an intraperitoneal injection ∼7 h after the onset of the dark cycle, the approximate time when fasted mice typically enter torpor. On the following day, mice were again randomized to a new dosage. This procedure was repeated for three days, such that each mouse received every dose. After the fourth day, the mice were returned to 28–30°C housing for 1 wk. The animals underwent the same procedure using another randomized nucleotide. The process was repeated until each mouse received each dose of the three nucleotides. Mice also received an intraperitoneal injection of IMP at 800 mg/kg and vehicle (saline) using the same procedure.
Experimental protocols: MSG treatment.
The pups from a C57BL litter were treated with MSG perinatally, as described previously (18). At 2 mo of age, female MSG-treated mice (n = 3) were implanted with temperature telemeters. After 10 days of surgical recovery, these mice were fasted for 23 h while housed at 20°C. These MSG-treated mice were allowed to recover from the fast at 28–30°C for 1 wk. At that time, these mice received an intraperitoneal injection of 800 mg/kg AMP while housed at 20°C.
Experimental protocols: effect of AMP on heart rate.
Another set of C57BL mice (n = 6) were implanted with ECG telemeters. After 10 days of recovery from the surgery, the mice were fasted for 23 h at 20°C to induce a torpor bout. These mice were housed at 28–30°C for 1 wk after the fast. The mice were again housed at 20°C and injected with 800 mg/kg of AMP.
Experimental protocols: effect of adenosine and adenosine receptor blockers.
Female mice (n = 8) were implanted with temperature telemeters. After 10 days of recovery from the surgery, mice were housed at 20°C and received an intraperitoneal injection of either saline or aminophylline (20 mg/kg). Aminophylline was used as an antagonist to the adenosine receptor (3, 4). The first injection was followed 10 min later with a second intraperitoneal injection of either adenosine (100 mg/kg) or AMP (100 mg/kg). The mice were randomized into one of four conditions: 1) saline-adenosine, 2) saline-AMP, 3) aminophylline-adenosine, and 4) aminophylline-AMP. Over four consecutive days, every mouse received each of the four treatments such that each mouse was her own control. The dose of aminophylline was based on results from previous experiments, and was large enough to block adenosine receptors but not to induce seizures (4). The dose of adenosine was based on the solubility of adenosine in saline. AMP dosage was used to match that of adenosine.
All results are reported as means ± SE. The effect of nucleotides, saline, or fasting-induced torpor bouts on rate of Tb decline was statistically assessed using a repeated measures ANOVA, followed by a post hoc Bonferroni test for statistical significance. An unpaired t-test was used to compare minimum Tb in fasted wild-type mice vs. fasted MSG-treated mice, whereas a paired t-test was used to compare minimum Tb in fasted MSG-treated mice vs. AMP-injected MSG-treated mice. A repeated-measures ANOVA, with a 4 × 2 design, was used to assess differences between saline and aminophylline on the minimum Tb during the hypothermic bout induced by adenosine or AMP, followed by a Bonferroni post hoc test. Significance levels of P < 0.05 were accepted.
Peripheral AMP, ADP, and ATP dose dependently lower Tb in wild-type mice.
When C57BL mice were fasted at an ambient temperature of 20°C, all of them entered a torpor bout, reaching a minimum Tb of 24.9 ± 1.5°C. Peripheral administration of AMP at 20°C led to a dose-dependent decrease in Tb (Fig. 1A). Minimum Tbs attained at the higher doses of AMP (27.3 ± 1.3, 26.3 ± 1.1°C for 400 and 800 mg/kg, respectively) reached values indicative of torpor. Peripheral administration of ADP also led to a dose-dependent decrease in core Tb (Fig. 1B). Surprisingly, peripheral administration of ATP also resulted in a dose-dependent fall in minimum Tb (Fig. 1C). Two other controls were performed: 1) injection of saline and 2) injection of IMP, a breakdown product of AMP, at 800 mg/kg. Intraperitoneal administration of saline did not lead to a significant decrease in Tb (data not shown), whereas administration of IMP caused a decrease in Tb of 3°C (to 33.0 ± 0.5°C).
The maximum rate of Tb decline was calculated over a 30-min window during saline injection, during a fasting- induced torpor bout, and after administration of 800 mg/kg of AMP, ADP, or ATP. As Figure 2 shows, the maximum rate of Tb decline was significantly greater with the nucleotides than during fasting-induced torpor.
AMP lowers core Tb in mice unable to enter torpor.
We have previously shown that ablation of the arcuate nucleus within the hypothalamus using perinatal administration of MSG results in adult mice that do not enter torpor upon fasting (18). To test the hypothesis that the arcuate nucleus is required for the hypothermic effect of AMP, mouse pups were treated perinatally with MSG. As adults (body weight = 24.6 ± 0.6 g), these mice were fasted. As Fig. 3B shows, MSG-treated mice did not enter fasting-induced torpor, as we have seen previously (18). However, when injected peripherally with AMP at 800 mg/kg, all showed a hypothermic bout (Fig. 3C), with a minimum Tb of 23.4 ± 0.6°C (Fig. 3D).
Peripheral AMP drastically depresses heart rate.
To further examine the physiological effects of AMP in mice, C57BL mice were implanted with EKG/Tb telemeters. These mice all entered torpor bouts when fasted. Fig. 4A shows the Tb tracing of the mice at the initiation of the torpor bout, which was ∼7 h after initiation of the fast. Heart rate fell in these fasted mice at the onset of torpor over these first 35 min from 646 ± 21 beats/min to 294 ± 19 beats/min (Fig. 4C). One week later, these same mice received an intraperitoneal injection of AMP at 800 mg/kg. As Fig. 4A shows, the rate of core Tb decline was greater with injection of AMP than during fasting, as discovered earlier (Fig. 2). Amazingly, the heart rate of mice injected with AMP fell from 755 ± 15 to 268 ± 17 bpm within 1 min (Fig. 4C). A representative ECG tracing from a mouse 1 min before AMP injection and 1 min after AMP injection is shown in Fig. 4B. When plotted as a function of Tb (Fig. 4C), heart rate of the mice was significantly higher during fasting-induced torpor than after administration of AMP.
Aminophylline, an adenosine receptor blocker, blocks the hypothermic effect of adenosine and AMP.
Because AMP induced such a drastic depression in heart rate and that adenosine is known to produce a profound bradycardia (9–11, 26, 28, 35), it was suspected that AMP induced a hypothermic state via the adenosine receptor. To test this hypothesis, mice were pretreated with either saline or aminophylline (20 mg/kg), an adenosine receptor antagonist (3, 4). Pretreatment with saline or aminophylline occurred 10 min before an injection of either adenosine (100 mg/kg) or AMP (100 mg/kg). Mice that received saline followed by adenosine underwent a bout of hypothermia, with a minimum Tb of 28.9 ± 1.2°C (Fig. 5). The adenosine-induced bout of hypothermia was nearly identical to that bout of hypothermia induced by AMP, with a minimum Tb of 29.0 ± 1.3°C. When mice were pretreated with aminophylline, the hypothermic bout induced by adenosine was significantly blunted, as expected (Fig. 5). Importantly, aminophylline also blocked the hypothermic effect of AMP to a similar extent as the block on adenosine (Fig. 5).
There is a great amount of interest in identifying specific factors that induce torpor in animals, as well as the mechanisms causing the reduction of metabolic rate. The therapeutic benefits of a regulated, well-controlled hypothermia are numerous (19). Recent investigations have used H2S, 2-deoxy-d-glucose (2-DG), and AMP to induce a hypothermic state (6, 8, 43, 44). Mice breathing H2S gas undergo a rapid decrease in metabolic rate, with a subsequent decrease in core Tb, to as low as 15°C, in a dose-dependent manner (6). Glucose deprivation via delivery of 2-DG results in hypometabolism and accompanying hypothermia (8, 43). However, given the large surface area-to-volume ratio of the mouse relative to the human, extrapolation of these studies in terms of a therapeutic hypothermia for humans should be taken carefully, in that core Tb falls much more quickly in a mouse than in humans.
The mechanisms of hypothermia induced by H2S, 2-DG, and AMP are not well understood, although it has been shown that the hypothermia induced by 2-DG has thermoregulatory patterns that are different from those observed in fasting-induced torpor (43). We sought to determine whether AMP truly mimics torpor and to elucidate the mechanism by which AMP induces hypothermia. We were able to repeat the observation (44) that peripherally administered AMP induces a dose- dependent hypothermic state in mice (Fig. 1A). We had hypothesized that AMP-induced hypothermia was mediated via activation of hypothalamic AMPK, an intracellular sensor of energy status (20, 23), particularly because this enzyme is so sensitive to ghrelin and leptin (1, 7, 23, 24), known modulators of the torpor response. However, particularly troubling for this hypothesis was that the high-energy molecule ATP was equally effective in inducing hypothermia (Fig. 1C). In fact, ATP seemed to have a longer lasting effect on the duration of hypothermia than either ADP or AMP (Fig. 1). We now support the hypothesis that peripheral administration of adenine nucleotides induces hypothermia via conversion of these nucleotides to adenosine, with subsequent activation of the adenosine receptor. Indeed, blockade of the adenosine receptor with aminophylline nearly completely blunted the hypothermic effect of AMP (Fig. 5). The similarities in depth and duration of hypothermia induced by AMP and adenosine, as well as the extent of the blunting of the hypothermia by aminophylline, were simply striking (Fig. 5).
Endothelial cells express ecto-5′-nucleotidases that can dephosphorylate ATP, ADP, and AMP to adenosine (4, 12, 26, 27, 30, 32). Adenosine can have a profound effect on the cardiovascular system, lowering heart rate, cardiac output, peripheral resistance, and blood pressure (3, 4, 10, 11, 28, 42, 45). Adenosine binds specific G-protein linked cell surface receptors, of which there are multiple forms (12). When activated by adenosine, the A1 receptor, which is highly expressed in the sino-atrial and atrial-ventricular nodes of the heart, activates adenosine-sensitive potassium channels, leading to hyperpolarization of the cell with the resultant slowing of nodal firing (25, 35). The slowing of heart rate induced by adenosine (10, 11, 28) is also observed with administration of AMP (Fig. 4) and ATP (28). Importantly, hydrolysis of ATP to adenosine by nucleotidases is required for the bradycardic action of ATP (29). The observations that 1) ATP has a more potent effect on heart rate than adenosine (28), and 2) a more potent hypothermic effect than AMP (Fig. 1) can perhaps be explained by a longer elevation of circulating adenosine due to the kinetics of nucleotide hydrolysis. This remains to be tested. In addition to conversion to adenosine, another fate of AMP is the conversion to IMP, catalyzed by adenylic deaminase. Peripheral administration of IMP had a small, but significant impact on the Tb of mice (see results). Because AMP induces such a large decrease in Tb compared with IMP, it is not likely that degradation of AMP to IMP is the mechanism for AMP-induced hypothermia. The mechanism of IMP-induced hypothermia remains to be determined.
It does not appear that AMP requires central nervous system mechanisms to induce hypothermia. The arcuate nucleus within the hypothalamus regulates energy homeostasis, through alterations in both hunger and energy expenditure (34). Perinatal treatment of mice with MSG ablates this region (39). MSG-treated mice did not enter torpor upon fasting (Fig. 3 and Ref 18), indicating the importance of this region in mediating the metabolic response to fasting. Similarly, chemical ablation of the paraventricular nucleus, which receives much input from the arcuate nucleus, prevents daily torpor in Siberian hamsters (31). Given that AMP caused hypothermia in MSG-treated mice (Fig. 3), the arcuate nucleus likely plays little role in mediating the AMP-induced hypothermia. However, it should be noted that we cannot exclude the possibility that the role of the arcuate nucleus in fasting-induced torpor is to produce and release AMP and/or adenosine in response to fasting and that these molecules then mediate the decrease in Tb. Our data only show that exogenous AMP does not require the arcuate nucleus to exert its hypothermic effect in mice.
The extent and time course of the bradycardia induced by AMP suggest the ensuing hypothermia is a result of the cardiovascular effects, and not central nervous system effects, of peripheral AMP delivery. Before injection of AMP, the heart rate of these mice was near maximal, at 750 bpm, which was a function of the cool ambient temperature of the housing in this study (38). The heart rate of fasted mice just before torpor was significantly lower at 650 bpm because these mice had been fasted for 7 h (Fig. 4). The injected AMP caused heart rate to fall to ∼250 bpm within a very short time frame (Fig. 4), likely due to activation of the adenosine receptor. While the kinetics of heart rate change during AMP-induced hypothermia and fasting-induced hypothermia are clearly different (Fig. 4), the possibility exists that fasting-induced torpor elevates circulating adenosine or AMP, and that this compound lowers heart rate during torpor. This possibility remains to be tested. The decrease in heart rate with AMP administration (Fig. 4) likely contributes toward the large decrease in blood pressure caused by adenosine (3, 42, 45). With a depressed perfusion pressure, we hypothesize that the ability to deliver glucose and oxygen is severely limited, forcing cell/organ metabolic rate to fall. The diminished capacity to produce heat via metabolism after AMP injection would result in a fall in core Tb. Hence, rather than directly influencing the metabolism of organs as suggested by others (44), exogenous AMP may evoke hypothermia through a bradycardia-related mechanism.
Adenosine and adenine nucleotides may be used therapeutically to induce hypothermia, in addition to the therapeutic use of adenosine to treat paroxysmal supraventricular tachycardia (9). While it remains to be determined whether AMP or adenosine plays a physiological role in mediating the bradycardia that occurs during fasting-induced torpor, our data suggest that peripheral administration of AMP does not mimic a true state of torpor but rather evokes a hypothermic state due to its cardiovascular effects.
This work was supported by National Institutes of Health Grant R15 HL081101-01 (to S. J. Swoap).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society