Heart rate dynamics in a marsupial hibernator

The physiological processes that are engaged during bouts of torpor in mammalian hibernation are numerous and complex. These processes include, but are certainly not limited to, abandonment of euthermic thermoregulation (Heller, 1983), apnea which can last seconds to minutes (Lyman, 1965; Milsom and Jackson, 2011), and active suppression of metabolic rate (Geiser, 2004). Further, shared cardiovascular changes in animals that use torpor include bradycardia (Zosky and Larcombe, 2003; Swoap and Gutilla, 2009; Morhardt, 1970; Harris and Milsom, 1995) mediated by the autonomic nervous system (ANS) and peripheral vasoconstriction (Swoap and Gutilla, 2009; Osborne et al., 2005) mediated by the sympathetic nervous system (SNS). Other indicators of altered autonomic function during a bout of torpor include SNS activation of white adipose tissue (Swoap and Weinshenker, 2008) with the resultant drop in leptin and elevation in free fatty acid release, and withdrawal of SNS activity from brown fat (Cannon and Nedergard, 2004) in those organisms that have this heat-generating organ.

Heart rate (f_H) during torpor has been examined in several placental mammals such as the 13-lined ground squirrel, the marmot, the black bear and long-eared bats (Tøien et al., 2011; Currie et al., 2014; Eagles et al., 1988; Hampton et al., 2010). Placental mammals show three distinguishing features of f_H during hibernation (Milsom et al., 1999). First, all display low f_H (between ∼8 and 12 beats min⁻¹) during bouts of deep torpor. Second, f_H slows during entrance into torpor in part as a result of asystoles, or skipped beats. Third, f_H during deep torpor is elevated periodically, and that elevation is associated with ventilation (ventilatory tachycardia). Ventilatory tachycardia is also seen in deep torpor in a marsupial (Cercartetus concinnus: Zosky and Larcombe, 2003); however, little is known about the variability in f_H during entrance into torpor in a marsupial hibernator. Further, in animals such as mice and Djungarian hamsters that utilize exclusively daily torpor during which the minimum body temperature (T_b) is around 20°C and the length of the torpor bout is measured in hours and not days or weeks, the minimum f_H is much higher, around 70 beats min⁻¹ (Hudson and Scott, 1979; Swoap and Gutilla, 2009). An increase in f_H variability during entrance into torpor is also seen in these mammals that use daily torpor (Mertens et al., 2008; Vicent et al., 2017).

Marsupials diverged from placental mammals about 160 million years ago (Luo et al., 2011) and are interesting from an evolutionary point of view because they have lost functional brown fat (Oelkrug et al., 2015), which is widely considered crucial for thermogenesis during rewarming from hibernation. The objectives of the current experiment therefore were to measure the f_H and f_H variability characteristics in the marsupial eastern pygmy possum (Cercartetus nanus) at different depths of torpor to examine whether cardiac function especially during deep torpor shows any marsupial idiosyncrasies related to their different form of thermogenesis. This arboreal mammal is found in southeastern Australia and, from an experimental point of view, has an advantage because it enters torpor at a wide range of ambient temperature (T_a), resulting in shallow and short bouts of torpor at high T_a, as well as deep and multi-day bouts of torpor at low T_a (Geiser, 1993; Bartholomew and Hudson, 1962). Because the eastern pygmy possum can undergo both shallow brief torpor and deep hibernation, this organism allows for the examination of f_H control in both of these situations. The eastern pygmy possum can hibernate without food for up to an entire year, longer than known for any other species, and also expresses some of the longest torpor bouts measured to date in the laboratory of up to 1 month (Geiser, 2007). This attribute of varied torpor bout duration and depth allows for examination of the full dynamic range of f_H control in a mammal.

Five eastern pygmy possums, Cercartetus nanus (Desmarest 1818), were caught using nest boxes near Dorrigo, NSW, Australia (30°22′S, 152°34′E). The mean minimum monthly T_a in Dorrigo is 4.4–15.0°C and the mean maximum T_a is 14.4 –24.1°C. The highest recorded maximum T_a is 36.3°C and the lowest recorded minimum T_a is −3.5°C. The animals were housed singly in cages with sawdust bedding and nest boxes containing bedding material on a 12 h:12 h light:dark schedule at a T_a of approximately 22°C. Animals were fed apples, walnuts, sunflower seeds, rolled oats and a mixture of high protein baby cereal, honey, a vitamin supplement, boiled eggs or protein powder and pureed fruit. Food was not provided ad libitum because pygmy possums tend to become obese in captivity. Water was freely available. Throughout the experiments, body mass ranged from 25 to 44 g (mean 35 g). All experiments were approved by the University of New England Animal Ethics Committee and were performed in accordance with the guidelines described by the National Health and Medical Research Council.

Each animal was implanted with an electrocardiogram (ECG) telemeter that (1) detects electrical signals across the heart, (2) measures T_b and (3) measures locomotor activity (ETA-F10, Data Sciences International, St Paul, MN, USA). For implantation, animals were anesthetized using 5% isoflurane in oxygen gas and maintained with 2.5–3% isoflurane for the duration of the implantation procedure. The telemeter was implanted in the abdominal cavity and ECG leads were placed subcutaneously, approximating a lead II configuration, and held in place by sutures used to close the body wall. Wound clips (7 mm size reflex clips, Fine Science Tools, Foster City, CA, USA) were used to close the abdominal incision. During the post-operative recovery period, animals were housed individually at 30°C in cages placed half atop a heating pad. The pygmy possums were allowed to recover for 10 days before any experimentation began.

Telemeter data (ECG recordings and T_b) were monitored for a period of 10 s, once per minute, using receivers beneath the home cage (RPC-1, Data Sciences International). After approximately 4 weeks of sampling, each possum was placed individually into a 0.5 l metabolic chamber at an initial flow rate of ∼300 ml air min⁻¹ without food and water. These measurements began in the morning and were conducted during the daytime, the period of rest of pygmy possums. Air from the chamber was dried and directed into a Sable FC-1B oxygen analyzer for determination of oxygen content. Flow rate was measured with a mass flowmeter (Omega FMA-5606, Stamford, CT, USA). The T_a of the chamber, measured to the nearest 0.1°C with a calibrated thermocouple inserted ∼1 cm into the respirometry chamber, was initially set to 25°C to measure f_H and oxygen consumption of euthermic possums. After 3 h, the T_a of the chamber was raised to approximately 32°C, and lowered again to 25°C at a rate of 1.6°C h⁻¹ to measure f_H and O₂ consumption as a function of T_a and to determine basal values. The possum was moved back to its home cage after completion of the T_a ramp in the metabolic chamber. To examine deeper torpor bouts, the possums were each placed again in the metabolic chamber in the late afternoon and left there overnight with the T_a of the chamber set initially to 13–14°C. After the possum had entered torpor in the morning, as assessed in real time by the T_b of the animal, the air flow rate was lowered to approximately 150 ml min⁻¹ to increase the O₂ differential, and the T_a was lowered by 2–3°C approximately every hour to a minimum of 5–8°C. Sampling from the telemeters was continuous in this condition. Once the possum aroused, it was placed back into its home cage at 22°C. To assess the link between ventilation and f_H control during torpor, the possums were placed in the metabolic chamber once again and held at 15°C overnight in the absence of water and food. A pressure transducer (MPX2010, Motorola, Denver, CO, USA) was used to monitor chamber pressure, from which ventilation frequency was determined (Cooper and Withers, 2010). For f_H, f_H variability and ECG noise analysis, raw data files of ECG recordings from the telemeters were imported into Ponemah Physiology Platform software (Data Sciences International). Noise detection was enabled and waveforms were analyzed.

All results are reported as means±s.e.m. Statistical analyses were performed in SPSS 15.0 (IBM Corp., Armonk, NY, USA). Repeated measures ANOVAs were performed and, when significance was shown, were followed with a post hoc Tukey test. P<0.05 was considered statistically significant.

When pygmy possums were housed in their cages in a room maintained at a T_a of 22±1°C and T_b and ECG tracings were monitored, three of the five possums periodically entered a torpor bout despite the presence of food and water (Fig. 1). At this T_a and on days when the possums did not enter torpor, the average 24 h f_H for the group was 623±17 beats min⁻¹ (n=5), with an average T_b over the same time frame of 35.5±0.4°C. On those days that the possums entered torpor spontaneously (Fig. 1), f_H fell significantly to a minimum of 34±3 beats min⁻¹ (n=3) with a minimum T_b of 24.1±0.1°C. The relationship between f_H and T_b in the animals that entered torpor was complex (Fig. 1). f_H significantly dropped before the onset of the reduction in T_b. This was followed by a more gradual decrease in f_H as T_b declined. f_H increased dramatically during the arousal period, which typically took less than 60 min.

Over 10 h, the T_a of the metabolic chamber housing an implanted pygmy possum was ramped from 25°C to 31–32°C and back down to 25°C (Fig. 2A). Fig. 2E,F shows typical ECG tracings from a pygmy possum at a T_a of 25 and 31°C. From these ECG tracings, both f_H and noise from the ECG tracing were quantified together with the rate of O₂ consumption (V̇_O2). A typical response to the T_a ramp in V̇_O2, f_H and noise on the ECG tracing is shown in Fig. 2B–D. The calculated lower critical temperature of the thermoneutral zone from V̇_O2 measurements was 30.0±0.3°C, similar to that found earlier for this same species (Song et al., 1997). The lowest V̇_O2 in the euthermic possum was 0.705±0.048 ml O₂ g⁻¹ h⁻¹ (n=5) at a T_a of 31.0±0.2°C and a T_b of 35.4±0.3°C. Similarly, the minimum f_H at a T_a of 31°C was 294±31 beats min⁻¹, and the noise in the ECG tracing dropped to 5.1±1.0% of the noise value at a T_a of 25°C.

When housed individually in a metabolic chamber at 13–14°C without food and water overnight for monitoring T_b, ECG tracings and V̇_O2, all possums entered a bout of torpor within 24 h of residence in the metabolic chamber. Before the bout of torpor (i.e. euthermic, with a T_b of 35.5±0.3°C), the average f_H of the possums was 624±11 beats min⁻¹, not significantly different from the f_H when housed at 22°C while not in the metabolic chamber. The average V̇_O2 of the non-torpid pygmy possums at this T_a of 13–14°C was 3.76±0.29 ml O₂ g⁻¹ h⁻¹. Once the possums entered torpor, T_b dropped to approximately 15°C (14.9±0.4°C), which was 1.3±0.3°C above T_a. The minimum V̇_O2 at this T_a fell to 0.110±0.021 ml O₂ g⁻¹ h⁻¹, and the f_H dropped to 25±2 beats min⁻¹ (n=5). When the T_a of the chamber was then lowered to 5–8°C over a period of several hours (see Fig. 3), T_b fell to an average of 8.0±1.0°C, which was 1.2±0.4°C above T_a. The V̇_O2 at this T_a was 0.029±0.007 ml O₂ g⁻¹ h⁻¹ and the minimum f_H dropped to 9±1 beats min⁻¹ (n=5). The lowest f_H sustained in any possum during periods of a stable f_H was 8 beats min⁻¹ in a possum with a T_b of 6.0°C. Attempts to lower the T_a below 5°C evoked arousals in the possums (data not shown).

ECGs were examined during three phases of these deep torpor bouts shown in Fig. 3: pre-torpor, entrance into torpor, deep torpor. Before the possums entered torpor, f_H was fast and had little variability (see ECG tracing in Fig. 4A and the accompanying Poincare plot). During entrance into torpor (Fig. 4B), the ECG displayed skipped beats, with the heart beats occurring episodically in sets of 3 or 4. The variability of f_H is easily seen on the accompanying Poincare plot. During deep torpor, a consistent pattern was observed as shown in Fig. 4C,D. The f_H was slow and steady for minutes at a time, followed by a slowly increasing f_H, then a burst of ECG noise and an associated elevation of f_H, typically to about 40 beats min⁻¹. This elevation lasted 20–30 s, at which point the f_H of the possum returned to a slow and steady level of 8–12 beats min⁻¹. The QRS complex was also examined at three T_b throughout these deep bouts of torpor. The duration of the QRS complex within the ECG changed as a function of T_b. As Fig. 5 shows, the duration of the QRS increased 5.8-fold at a T_b of 8°C (69±9 ms) relative to the QRS duration in euthermia (12±1 ms). During arousal, f_H was quickly obscured by the noise on the ECG tracing, presumably a result of the massive shivering that occurs during emergence from torpor in this species.

To examine whether the short bouts of tachycardia during deep torpor (Fig. 4D) were related to ventilation patterns, the implanted possums were placed in a metabolic cage with a pressure transducer for determination of breathing activities with simultaneous acquisition of T_b and f_H at a T_a of 14°C. A typical tracing during euthermia is shown in Fig. 6A. During deep torpor, periods of apnea were clearly discernible between bouts of breathing. A typical bout of breathing during torpor is shown in Fig. 6B, with the region between 120 and 180 s shown on an expanded scale in Fig. 6C. The bouts of increased f_H during deep bouts of torpor in the pygmy possum occurred as the animal began to breathe. f_H slowed again during the subsequent apnea bout.

Mammalian f_H can be tremendously flexible, varying greatly throughout the day, even on a beat-to-beat basis. This flexibility in f_H is of particular interest in hibernators, where f_H falls as low as a few beats per minute in a deep bout of torpor. The maximum f_H for the eastern pygmy possum appears to be about 625 beats min⁻¹, similar to an earlier estimate (Bartholomew and Hudson, 1962). The f_H was very responsive to T_a above 25°C, falling to 325 beats min⁻¹ at thermoneutrality at a T_a of 31°C, as has been shown before (Bartholomew and Hudson, 1962). This same relationship between f_H and T_a exists in rats and mice. Importantly, the slope of the f_H/T_a relationship is much less steep in rats and mice (8 and 15 beats min⁻¹ °C⁻¹, respectively) than it is with the possum (∼50 beats min⁻¹ °C⁻¹) between 25 and 31°C (Swoap et al., 2004). However, at T_a below 22°C when the possums were euthermic, the f_H stayed unchanged, measured herein at 624 and 623 beats min⁻¹ at 23 and 15°C, respectively. This lack of change of f_H at cooler temperatures has been seen previously, including a measured f_H of euthermic possums of 600–650 beats min⁻¹ at a T_a of 5°C (Bartholomew and Hudson, 1962). This suggests that f_H is at its maximum at 22°C, and that the SNS has a much greater influence than the parasympathetic nervous system (PNS) over the possums' heart period at T_a below 22°C. It is unclear whether cardiac output continues to increase as T_a drops through elevation in stroke volume in response to the elevated metabolic demand at cooler temperatures.

All animals that use torpor, either hibernation or daily torpor, exhibit a large drop in f_H during entrance into torpor. This is true for eastern pygmy possum as well (present study and Bartholomew and Hudson, 1962). However, the extent to which f_H falls differs between animals that use daily torpor and hibernators. With the lower absolute T_b in hibernators, it is not surprising that the absolute minimum f_H in hibernators is substantially lower (5–10 beats min⁻¹) compared with the minimum f_H (70–150 beats min⁻¹) in animals that use daily torpor (Milsom et al., 1999; Morhardt, 1970; Swoap and Gutilla, 2009; Zosky, 2002). However, even for any given T_b that both groups of animals can achieve, the hibernators have a substantially lower f_H. For example, during entrance into torpor, the eastern pygmy possum had a f_H of ∼35 beats min⁻¹ at 24°C, similar to that seen in placental hibernators, woodchucks, hedgehogs and ground squirrels (Lyman, 1958; Harris and Milsom, 1995). This f_H for hibernators at a T_b of 24°C is only a small fraction (20–45%) of what is observed in animals that use daily torpor at the same T_b, including placental mammals (the mouse and Djungarian hamster) and in the fat-tailed dunnart, a marsupial (Swoap and Gutilla, 2009; Mertens et al., 2008; Zosky, 2002), supporting the view that hibernators and daily heterotherms differ not only ecologically but also functionally (Ruf and Geiser, 2015).

The relationship between T_b and f_H showed marked hysteresis (Fig. 1), suggesting that the control of f_H during entrance into torpor and emergence from torpor are the result of different phenomena. Hysteresis loops are also observed with f_H and T_b in the mouse (Swoap and Gutilla, 2009; Morhardt, 1970), and when examining metabolic rate as a function of T_b in dunnarts (Geiser et al., 2014), and in the QT interval of the ECG as a function of T_b in the Djungarian hamster (Mertens et al., 2008). Interestingly, the breadth of the f_H/T_b loop is much greater in the eastern pygmy possum than in other hibernators as well as in animals that use daily torpor. For example, at a T_b of 30°C, the difference in f_H between entrance and exit from torpor is about 400 beats min⁻¹ in the eastern pygmy possum. For hibernators, such as the ground squirrel, and in animals that use daily torpor, such as the mouse and Djungarian hamster, f_H breadth is approximately 200–275 beats min⁻¹ (Lyman, 1965; Swoap and Gutilla, 2009; Mertens et al., 2008).

The variability of f_H during entrance into torpor is common to all mammals that enter torpor (Milsom et al., 1999), and the eastern pygmy possum is no different (Fig. 4). There appear to be at least two sources of f_H variability during torpor and entrance into torpor that are both linked to the ANS. First, the appearance of asystoles (skipped beats) during entrance into torpor (Fig. 4) is seen in many hibernators and is a consequence of elevated PNS activity during entrance into torpor (see Milsom et al., 1999, for a review). While we did not perform any pharmacological experiments in the current study, administration of atropine, a muscarinic receptor antagonist, into the closely related western pygmy possum during torpor significantly elevates f_H and eliminates asystoles (Zosky and Larcombe, 2003). The second source of variability in f_H (Fig. 4) occurred during the periods between bouts of apnea in deep torpor in the eastern pygmy possum. f_H was elevated from 8 to 12 beats min⁻¹ during apnea to approximately 40 beats min⁻¹ during ventilation in deep torpor (Fig. 6). This ventilation-associated tachycardia is also seen in several other hibernating species, such as bears and ground squirrels, and is indicative of altered ANS activity (Dawe and Morrison, 1955; Tøien et al., 2011; Milsom et al., 1999; Harris and Milsom, 1995). Indeed, atropine administration during torpor in the western pygmy possum eliminates ventilation-associated tachycardia (Zosky and Larcombe, 2003). The low f_H of 8 beats min⁻¹ we measured here is in contrast with a previous study where this same species had a f_H of 28 beats min⁻¹ during torpor (Bartholomew and Hudson, 1962). While it is difficult to know the source of the difference between these studies, Bartholomew and Hudson (1962) were surprised at the relatively high torpid f_H and concluded the animals were likely disturbed when instrumented during the bout of torpor. Notably, we used radiotelemetry here with indwelling lines that minimize disturbance of the possums.

We have shown here that the eastern pygmy possum has an enormous dynamic range in f_H of 600 beats min⁻¹, from 625 to 8 beats min⁻¹ (an 80-fold difference), to match the metabolic flexibility (greater than 100-fold) in torpor and euthermia. The animal's T_b during torpor plays a role in this flexibility, which can be seen by the Q₁₀ of the QRS duration of approximately 2 (Fig. 5), similar to that seen in the Djungarian hamster (Mertens et al., 2008). Two other much smaller mammals have been reported to have a greater dynamic range in f_H, the Etruscan shrew (Fons et al., 1997) and the Gould's long-eared bat (Currie et al., 2014). The 2.4 g shrew has a dynamic range of 900 beats min⁻¹ (euthermic rate of 1000 beats min⁻¹ and torpid rate of 100 beats min⁻¹) whereas the 9 g bat has a range of 800 beats min⁻¹ (euthermic rate of 800 beats min⁻¹ and torpid of 8 beats min⁻¹). The rapid 600 beats min⁻¹ increase in f_H (and likely cardiac output) during recovery from torpor in the pygmy possum is probably important for meeting O₂ demand for metabolic rate elevation and heat production, such as shivering in skeletal muscle, during rewarming as has been suggested for the shrew (Fons et al., 1997). To sum, despite the lack of functional brown fat, f_H control in this marsupial hibernator throughout a bout of deep torpor appears indistinguishable from that of a placental hibernator in terms of depth, appearance of skipped beats and ventilation-associated tachycardia. However, there appears to be a large difference in minimum f_H between hibernators and mammals that use daily torpor, even when measured at the same T_b.

We thank Christine Cooper and Phil Withers for the loan and instructions for the use of the pressure transducer.