Torpor, a controlled rapid drop in metabolic rate and body temperature (Tb), is a hypometabolic adaptation to stressful environmental conditions, which occurs in many small mammals, marsupials, and birds. To date, signaling pathways required for torpor have not been identified. We examined the role of the sympathetic nervous system (SNS) in mediating the torpor adaptation to fasting by telemetrically monitoring the Tb of dopamine β-hydroxylase knock-out (Dbh–/–) mice, which lack the ability to produce the SNS transmitters, norepinephrine (NE), and epinephrine. Control (Dbh+/–) mice readily reduced serum leptin levels and entered torpor after a fast in a cool environment. In contrast, Dbh–/– mice failed to reduce serum leptin and enter torpor under fasting conditions, whereas restoration of peripheral but not central NE lowered serum leptin levels and rescued the torpor response. Torpor was expressed in fasted Dbh–/– mice immediately after administration of either the nonselective β-adrenergic receptor agonist isoproterenol or the β3-adrenergic receptor (AR)-specific agonist CL 316243 [disodium (RR)-5-[2-[[2-(3-chlorophenyl)-2-hydroxyethyl]-amino]propyl]-1,3-benzodioxazole-2,2-dicarboxylate], but not after administration of β1, β2, or α1 agonists. Importantly, the β3-specific antagonist SR 59230A [3-(2-ethylphenoxy)-1-[(1,S)-1,2,3,4-tetrahydronapth-1-ylamino]-2S-2-propanol oxalate] severely blunted fasting-induced torpor in control mice, whereas other AR antagonists were ineffective. These results define a critical role of peripheral SNS activity at β3-AR-containing tissues in the torpor adaptation to limited energy availability and cool ambient temperature.
- β-adrenergic receptor
- knock-out mice
- sympathetic nervous system
The scarcity of food in a cool environment can be very difficult for a small mammal. Some small endotherms have the ability to hibernate, using multiday bouts of lowered body temperature (Tb), as low as 5°C, that are separated by brief arousal periods (Lyman et al., 1982; Heldmaier et al., 1993; Geiser and Ruf, 1995; Geiser, 2004). Other species, including Mus musculus, exhibit a shallower minimum Tb (17–31°C) with bouts of shorter duration, typically ranging from 2 to 20 h (Hudson and Scott, 1979; Webb et al., 1982; Himms-Hagen, 1985a; Swoap, 2001; Bae et al., 2003; Bouthegourd et al., 2004). The regulation of Tb is maintained during torpor and not merely a function of heat lost to the environment after thermoregulation has been abandoned (Carey et al., 2003).
Daily caloric restriction, or even just a single overnight fast, can initiate a torpor bout in mice as long as the ambient temperature (Ta) is cool (Hudson and Scott, 1979; Webb et al., 1982; Himms-Hagen, 1985a; Gavrilova et al., 1999). Despite the extensive body of evidence characterizing hibernation and torpor (Wang and Hudson, 1978; Lyman et al., 1982), proximal signals that send an animal into, or arouse an animal from, a bout of torpor have yet to be identified. One of the signals that plays a central role in energy sensing is leptin, a hormone that is produced and secreted by white adipose tissue (WAT) and conveys information about peripheral fat stores to the hypothalamus (Flier, 1998). Leptin has been shown to blunt the depth of torpor, presumably because it increases metabolic rate, and ob/ob mice, which lack the leptin gene, enter torpor much more deeply and readily than wild-type mice (Döring et al., 1998; Geiser et al., 1998; Gavrilova et al., 1999; Freeman et al., 2004).
Heat is generated, in part, through nonshivering thermogenesis (NST). NST is stimulated primarily through activation of brown fat by the sympathetic nervous system (SNS). To determine whether recovery from a torpor bout requires this pathway, we used dopamine β-hydroxylase knock-out (Dbh–/–) mice, which lack the ability to synthesize the two primary neurotransmitters in the SNS, norepinephrine (NE) and epinephrine (Epi) (Thomas et al., 1995, 1998). Surprisingly, we found that these mice failed to enter torpor. Using a battery of agonist and antagonists for the adrenergic receptors (ARs), we found that torpor requires activation of the β3-AR, most likely in WAT.
Materials and Methods
Animals and pharmacological agents. All mice used were 2–3 months of age. Female Dbh–/– mice and their littermate controls, Dbh+/– mice, were bred and raised at Emory University and shipped to Williams College for physiological assessment. Dbh+/– mice were used as the controls because they have normal NE and Epi levels and are indistinguishable from wild-type mice for all aspects of behavior and physiology tested (Thomas et al., 1998; Murchison et al., 2004; Swoap et al., 2004). NIH Swiss and C57BL/6J mice were purchased from Harlan Sprague Dawley (Indianapolis, IN). Animals were maintained on a reverse 12 h light/dark cycle (lights on at 11:00 P.M., lights off at 11:00 A.M.). All animal studies were approved by the Williams College Institutional Animal Care and Use Committee. All pharmacological agents were purchased from Sigma (St. Louis, MO) with the exception of l-threo-3,4-dihydroxyphenyserine (DOPS), which was kindly provided by Sumitomo Pharmaceuticals (Osaka, Japan).
Implantation of temperature telemeters. Mice were anesthetized initially with 5% isoflurane in an oxygen stream, and maintained on 1–2% isoflurane. Mice were kept on a heating pad (38°C) throughout implantation of the body temperature (Tb) telemeter (TA10TAF20; Data Sciences International, Arden Hills, MN) into the peritoneal cavity. Mice were maintained on a heating pad for 48 h after the surgery and then housed individually at 28–30°C.
Fasting and pharmacology. After 1 week of recovery, mice were moved from the 28–30°C room into temperature-controlled (20 ± 0.25°C) custom-built cages containing telemetry receivers. Data from the Tb telemeters were recorded in 1 s streams at 500 Hz once per minute. In some instances, the rate of change of Tb was calculated over a 24 h period using a 30 min sliding window [i.e., (Tb at t = 0 min – Tb at t = 30 min)/30 min]. The maximum rate of heat loss was then generated from this data set. The mice were allowed to acclimate to 20°C housing for 2 d. Data collected from the third 24 h period of 20°C housing was considered the “fed” period. Mice were then fasted at the onset of the dark cycle, when mice typically initiate food consumption. Mice were allowed ad libitum access to water throughout all experiments. DOPS-treated Dbh–/– mice received three subcutaneous injections at 1 g/kg of body weight (24 h preceding the fast, 12 h preceding the fast, and at the onset of the fast). DOPS was put into solution at 20 mg/ml containing 2 mg/ml vitamin C. Some Dbh–/– mice received three subcutaneous injections of a solution containing 1 g of DOPS/kg and 0.25 mg of benzerazide/kg, three times at the same times indicated for DOPS treatment. Agonists/antagonists of the adrenergic receptors were dissolved in PBS, unless otherwise noted, and administered intraperitoneally 4.5 h after fast in 0.1 cc of solution. Dosages used were as follows (in mg/kg): 1 phenylephrine, 10 isoproterenol, 1 dobutamine, 2.5 salbutamol, 1 disodium (RR)-5-[2-[[2-(3-chloro-phenyl)-2-hydroxyethyl]-amino]propyl]-1,3-benzodioxazole-2,2-dicarboxylate (CL 316243), 1 prasozin in 3% DMSO, 1 atenolol, 1 erythro-1-(7-methylindan-4-yloxy)-3-(isopropylamino)-but an-2-ol (ICI 118,511), and 1 3-(2-ethylphenoxy)-1-[(1,S)-1,2,3,4-tetrahydronapth-1-ylamino]-2S-2-propanol oxalate (SR 59230A) in 3% DMSO.
Serum collection and analysis. Serum was collected from a new set of fed and fasted Dbh+/– and Dbh–/– mice 5.5 h after initiation of the fast. A commercially available kit was used to analyze serum leptin (Linco Research, St. Charles, MO).
Statistics. Data are reported as mean and SE. After ANOVA, Student's t tests were used to compare between genotypes, and paired t tests were used to compare Dbh–/– before and after pharmacological treatment. The 0.05 level of confidence was accepted for statistical significance.
Dbh–/– mice fail to enter torpor
Because brown fat thermogenesis can defend Tb in mice exposed to cold (Collins and Surwit, 2001; Nicholls, 2001), we hypothesized that the rapid rise in Tb during arousal from fasting-induced torpor requires activation of brown fat thermogenesis. To test this idea, we telemetrically monitored the Tb of Dbh–/– mice (n = 8) in response to fasting. We were unable to test the hypothesis because not one Dbh–/– mice initiated torpor in response to fasting (Fig. 1A,B). However, every fasted Dbh+/– littermate control mouse (n = 8) readily entered torpor (Fig. 1A). The minimum Tb of Dbh+/– mice during the torpor bout reached 22.6 ± 0.4°C, whereas Dbh–/– reached a minimum Tb at the end of the 24 h period of 31.1 ± 0.4°C. Although this hypothermic Tb of fasted Dbh–/– may at first glance suggest a torpor bout, none of these mice exhibited the distinctive steep drop in Tb of torpor (Fig. 1). To assess this aspect of torpor, we calculated the maximum rate of Tb drop over a 30 min period within the fed or fasted state. Dbh+/– that entered torpor lost Tb at a significantly higher rate as in the fed state (0.158 ± 0.004 vs 0.077 ± 0.004°C per minute, respectively). Dbh–/– mice, however, showed no significant difference in the maximum drop of Tb in fed or fasted states (0.085 ± 0.004 vs 0.072 ± 0.005°C per minute, respectively).
Peripheral, but not central, restoration of norepinephrine rescues the torpor phenotype
To determine whether the torpor phenotype can be rescued in the Dbh–/– mice, we implanted another nine Dbh–/– mice with Tb telemeters and administered the synthetic amino acid DOPS, which restores NE levels through the action of aromatic l-amino acid decarboxylase (AAD) to the brain and sympathetic nerves, but not to adrenal tissue (Thomas et al., 1998). When fasted, eight of the nine DOPS-treated Dbh–/– mice entered torpor normally (Fig. 1C), with a maximum rate of Tb drop of the entire group of 0.147 ± 0.007°C per minute. This suggests that Dbh–/– mice did not enter torpor because they specifically lack NE. To determine whether central or peripheral NE is required for the torpor, we administered DOPS alone (n = 11) or DOPS with benzerazide (n = 12), an inhibitor of AAD that does not cross the blood– brain barrier and restricts NE restoration to the brains of Dbh–/– mice (Murchison et al., 2004). As seen previously, DOPS rescued the torpor phenotype in 10 of the 11 mice (maximum rate of Tb drop, 0.163 ± 0.006°C per minute). However, none of the DOPS plus benzerazide-treated Dbh–/– mice initiated torpor (for a typical tracing, see Fig. 1C). The maximum rate of Tb drop in this group was 0.045 ± 0.003°C per minute. Thus, sympathetic NE is specifically required for the ability to enter torpor.
Torpor in fasted Dbh–/– mice requires activation of the β3-adrenergic receptor
To determine which class of AR mediates the regulation of torpor by NE, Dbh–/– mice (n = 10) were fasted and administered subtype-selective adrenergic agonists 4.5 h after the initiation of fasting. Activation of α1-ARs with phenylephrine during the fast had no effect on the Tb of Dbh–/– mice. After 7 d of recovery from the fast at 30°C, the same animals were rehoused at 20°C for 3 more days, fasted, and injected 4.5 h into the fast with isoproterenol, a nonspecific β-AR agonist. This resulted in an immediate drop in Tb, indicative of torpor in all 10 mice (minimum Tb was 24.6 ± 0.7°C; maximum rate of Tb drop was 0.140 ± 0.006°C per minute) (Fig. 2 A). After another 7 d recovery period at 30°C, and 3 d at 20°C, these same mice were fasted and administered saline, which had no effect on the Tb of these mice (Fig. 2).
To determine which subtype of the β-ARs was responsible for mediating the torpor response, a new cohort of nine Dbh–/– mice were taken through the following protocol: fasting with the β1-AR agonist dobutamine; 10 d recovery period; fasting with the β2-AR agonist salbutamol; 10 d recovery period; fasting with the β3-AR agonist CL 316243. Torpor was not observed in any of the fasted Dbh–/– mice injected with either dobutamine or salbutamol. Only activation of the β3-AR in fasted Dbh–/– mice permitted the expression of torpor in all nine mice within minutes of the injection, from which the mice did not spontaneously recover (minimum Tb was 21.8 ± 0.5°C; maximum rate of Tb drop was 0.140 ± 0.006°C per minute) (Fig. 2B).
Blocking the β3 adrenergic receptor blunts torpor in Dbh+/– mice
To determine whether the β3 pathway is important for torpor in normal mice, Dbh+/– mice (n = 24) were fasted and administered saline. After a 10 d recovery, these mice were fasted again and administered either the α1 antagonist prazosin (n = 6), the β1 antagonist atenolol (n = 6), the β2 antagonist ICI 118,511 (n = 6), or the β3 antagonist SR 59230A (n = 6). Blocking the β3-AR, and only this subtype, severely blunted fasting-induced torpor (Fig. 3). The length of the torpor bout, as defined as the amount of time spent at a Tb <31°C (Hudson and Scott, 1979), was much shorter when mice were treated with the β3 antagonist as compared with treated with vehicle (111.6 ± 12.6 vs 334.5 ± 6.8 min, respectively). A 10-fold higher dose (10 mg/kg) of SR58230A in a new set of seven Dbh+/– mice did not further blunt the torpor response, as all entered abbreviated bouts of torpor (minimum Tb = 25.9 ± 1.5; time in torpor, 122.1 ± 32.5 min). This suggests that the residual torpor response in SR 59230A-treated mice is mediated by a different pathway. Two wild-type mouse strains, NIH Swiss (n = 5) and C57BL6/J (n = 6) entered torpor normally after receiving SR 59230A but exhibited significantly shallower bouts of torpor than when administered vehicle (26.9 ± 0.8 vs 21.3 ± 0.5°C, respectively).
Leptin, torpor, and Dbh–/– mice
To explore the possibility that an inability to decrease circulating leptin prevents Dbh–/– mice from entering torpor, leptin was measure from serum collected 5.5 h after the initiation of a fast, just before entrance into torpor for control animals. In Dbh+/– mice (n = 7 per group), fasting lowered serum leptin (Fig. 4), as has been observed previously (Frederich et al., 1995). Fed Dbh–/– mice (n = 5) were hyperleptinemic, suggesting that normal mice have a chronic inhibition of leptin release that is mediated through the SNS. After fasting, serum leptin failed to appropriately fall in Dbh–/– mice (n = 7). Treatment of Dbh–/– mice with DOPS (n = 6), which allows for a normal bout of fasting-induced torpor (Fig. 1), caused a significant drop in circulating leptin, mimicking that of fasted Dbh+/– mice (Fig. 4). Similarly, injection of the β3-AR agonist 1 h before serum collection in fasted Dbh–/– mice (n = 7) caused a dramatic fall in serum leptin.
The data presented herein strongly support a critical role of the peripheral SNS activity at the β3-AR in the torpor adaptation of mice to limited energy availability and cool ambient temperature. All wild-type mice we tested initiated a bout of torpor in response to fasting in a cool environment (ambient temperature, <20°C). We found that Dbh–/– mice failed to enter torpor under such conditions. In contrast to the controlled and well regulated bout of hypothermia in wild-type mice, the Dbh–/– mice appeared to enter a hypothermic state in a slow, uncontrolled and unregulated manner. Dbh–/– mice with restored peripheral NE by acute treatment with DOPS or a β3 agonist acquired the ability to enter torpor, suggesting that their failure to enter torpor with fasting is not because of developmental differences and/or compensatory changes. These mice are simply missing the transmitter that allows for the initiation of the torpor response.
Many animals in the wild will enter torpor bouts spontaneously, often in preparation for an anticipated energetic stress (Speakman and Rowland, 1999; Geiser, 2004). However, mice require two environmental insults: (1) relatively cool ambient temperature and (2) caloric deprivation (Himms-Hagen, 1985a). Cool ambient temperature engages thermogenic pathways for the production of heat to offset heat loss. As such, cool ambient temperature activates the SNS particularly to thermogenic organs, such as brown fat (Himms-Hagen, 1985b). Fasting and/or caloric restriction, however, represses activity of the SNS to most tissues (Young and Landsberg, 1976), except white adipose tissue (WAT), in which SNS activity is activated (Migliorini et al., 1997; Rayner, 2001). NE released from the SNS synapses in WAT bind primarily the β3-AR.
In WAT, activation of the β3-AR induces lipolysis and leads to inhibition of leptin release (Young and Landsberg, 1976; Mantzoros et al., 1996; Giacobino, 1997). Importantly, the initiation of torpor requires leptin levels to be low (Geiser et al., 1998; Gavrilova et al., 1999; Freeman et al., 2004). In fact, both ob/ob mice missing the leptin gene and db/db mice missing the leptin receptor can spontaneously enter torpor under fed conditions (Webb et al., 1982; Himms-Hagen, 1985a). Our data then suggest that Dbh–/– mice cannot enter torpor during fasting because they cannot suppress leptin secretion in response to the fast (Fig. 4), because they lack stimulation of WAT by NE from the SNS. By re-establishing activation of the β3-AR during a fast in Dbh–/– mice with CL 316243, suppression of circulating leptin and initiation of torpor were both rescued. Interestingly, whereas injection of either isoproterenol or CL 316243 allowed fasted Dbh–/– to enter torpor, only isoproterenol allowed rewarming from the torpor bout (Fig. 2). This is particularly puzzling given that β3-mediated NST in brown adipose tissue (BAT) generates heat used to exit the torpid state (Cannon and Nedergaard, 2004). We propose three possible explanations for this phenomenon. First, the effects of isoproterenol on β3 receptors may persist long enough to rewarm from torpor hours after administration, whereas CL 316243 may be metabolized too quickly. Second, β1- and/or β2-mediated mechanisms (i.e., shivering or NST at other organs) may act in concert with BAT activation to end a torpor bout. Third, the activation of β3 receptors have effects other than leptin suppression, and sole activation of these receptors in Dbh–/– mice may upset other homeostatic mechanisms necessary for rewarming.
In summary, although low leptin signaling is only permissive for torpor (for example, ob/ob and db/db mice are not always in torpor) (Freeman et al., 2004), our data suggest that the specific activation of WAT by the SNS mediated through the β3-AR during fasting generates an environment (suppressed circulating leptin) conducive for the initiation of torpor in mice.
This work was supported by National Science Foundation Grant IBN 9984170 (S.J.S.) and in part by a Howard Hughes Medical Institute grant to Williams College. We thank Sumitomo Pharmaceuticals (Osaka, Japan) for their generous donation of l-threo-3,4-dihydroxyphenyserine.
Correspondence should be addressed to Steven J. Swoap, Department of Biology, Williams College, Williamstown, MA 01267. E-mail:.
Copyright © 2006 Society for Neuroscience 0270-6474/06/260241-05$15.00/0