The mammalian response to hypothermia is increased metabolic heat production, usually by way of muscular activity, such as shivering. Seals, however, have been reported to respond to diving with hypothermia, which in other mammals under other circumstances would have elicited vigorous shivering. In the diving situation, shivering could be counterproductive, because it obviously would increase oxygen consumption and therefore reduce diving capacity. We have measured the electromyographic (EMG) activity of three different muscles and the rectal and brain temperature of hooded seals (Cystophora cristata) while they were exposed to low ambient temperatures in a climatic chamber and while they performed a series of experimental dives in cold water. In air, the seals had a normal mammalian shivering response to cold. Muscles were recruited in a sequential manner until body temperature stopped dropping. Shivering was initiated when rectal temperature fell below 35.3 ± 0.6°C (n = 6). In the hypothermic diving seal, however, the EMG activity in all of the muscles that had been shivering vigorously before submergence was much reduced, or stopped altogether, whereas it increased again upon emergence but was again reduced if diving was repeated. We conclude that shivering is inhibited during diving to allow a decrease in body temperature whereby oxygen consumption is decreased and diving capacity is extended.
- hooded seal Cystophora cristata
- shivering electromyogram
- body core temperature
scholander et al. (23) first reported that body core temperature decreases during experimental diving in seals, and some 40 years later, Kooyman et al. (17) showed that central arterial temperature of a freely diving Weddell seal (Leptonychotes weddelli) decreased by ∼3°C during a 53-min dive. Moreover, Hill et al. (12) found that the aortic temperature of freely diving Weddell seals was reduced by ∼2°C in periods of active diving compared with resting periods without diving. Finally, Odden et al. (21) reported that brain temperature in hooded seals (Cystophora cristata) and harp seals (Pagophilus groenlandicus) may be reduced by as much as 3°C during relatively short (9–15 min) experimental dives. Scholander et al. (23) suggested that the reduction in body core temperature is caused by a reduction in metabolic heat production, which typically occurs in diving seals (6, 15, 24). Odden et al. (21), on the other hand, suggested that body cooling during diving is instead caused by a physiologically controlled increase in heat loss, which in turn will cause a metabolic depression due to Q10 effects. In any case, a reduction of deep body temperature, and particularly brain temperature, by as much as 3°C within minutes will normally lead to vigorous shivering in mammals (25). If body cooling is an effect of reduced metabolism, as suggested by Scholander et al. (23), shivering will compromise metabolic depression and, hence, the body cooling. If, instead, any cooling of the body is caused by a controlled increase in heat loss to enhance the metabolic depression through Q10 effects, as suggested by Odden et al. (21), shivering will again be counterproductive.
In the present study we have tested the hypothesis that shivering is inhibited in hypothermic diving seals. First, we exposed two seals to low ambient temperatures in a climatic chamber while we recorded rectal temperature and muscular activity in three different muscles to determine whether they responded in the proper mammalian way to a drop in deep body temperature when they were not diving. Second, we made three seals perform a series of experimental dives while we recorded muscular activity, brain and rectal temperature, and heart rate.
Three 1- to 2-yr-old hooded seals (C. cristata) (K-3/00: 1 yr, male, 62 kg; K-5/99: 2 yr, male, 81 kg; K-6/99: 2 yr, male, 92 kg) were caught as pups in the pack ice of the Greenland Sea and raised in captivity at Tromsø in 45,000-liter seawater pools with wooden ledges, where they were offered capelin (Mallotus villosus) or herring (Clupea harrengus) supplemented with a vitamin complex (4). The animals were collected under permits issued by The Royal Norwegian Ministry of Fisheries, and experiments were carried out under permit from the Norwegian Animal Research Authority.
Shivering in cold air.
Two of the experimental animals (K-3/00 and K-6/99) were gradually trained to accept being restrained on a specially designed board and left for ∼4 h inside a climatic chamber (type 24/50 DU; Weiss Technik, Giessen, Germany). Before experiments were performed, the animals were instrumented with electromyographic (EMG) electrodes in the humerotrapezius, spinotrapezius, and latissimus dorsi muscles (14) and with a thermocouple to record rectal temperature. The temperature inside the climatic chamber was then gradually reduced from +20°C to −10°C in 30 min, whereafter it was further reduced to a stable −35°C in 70 min. The experiment was terminated after ∼3 h of exposure to −35°C or earlier if rectal temperature fell below 33°C. At all times the animal was observed with the use of a video camera to enable us to distinguish between EMG activity caused by shivering and that caused by movements of the head and flippers. This experiment was repeated four times in each animal.
Shivering during diving.
In preparation for the experiments the animal was placed on a specially designed restraining board of classic design, and all three animals gradually became accustomed to experimental dives of a 10- to 15-min duration in water of 3–4°C. Before each experiment was performed, the animal was instrumented with EMG electrodes in the humerotrapezius, spinotrapezius, and latissimus dorsi muscles as well as with electrodes for measurement of heart rate and a thermocouple for measurement of rectal temperature, and in some of the experiments, brain temperature was measured as well (see Brain temperature). Before diving, the animals were allowed to equilibrate in the water for at least 1 h, whereafter they were submerged three times for periods of 10 or 15 min separated by recovery periods of 40–60 min. If the animal was shivering after the third dive, then it was submerged again for 3–5 min after 3–5 min of recovery. During these experiments the animals were always directly observed so that EMG activity caused by shivering could be distinguished from that caused by movements of the head and flippers. This series of experiments was repeated three times in animals K-3/00 and K-6/00, with the latter animal subsequently instrumented with a probe for measurement of brain temperature, whereafter the series of experiments was repeated three more times. Animal K-5/99 was instrumented with a brain probe from the very beginning, and the series of dive experiments was carried out twice in this animal.
Shivering was recorded as EMG activity in three large skeletal muscles along the dorsal side of the animal. We used the anatomical description of Howell (14) in combination with information obtained from dissection of a hooded seal of the same age as our experimental animals to determine the correct placement of the EMG electrodes in the humerotrapezius, spinotrapezius, and latissimus dorsi muscles. The EMG electrodes were homemade intramuscular electrodes similar to those described by Hohtola et al. (13): two 0.55-mm-thick needle electrodes were mounted 3 mm apart on a polyvinyl chloride disk and insulated, except for a 3-mm tip, with 0.1 mm of plastic spray PRF 202 (Taerosol, Kangasala, Finland). The other ends of the needles were soldered to a cable and connected to a differential amplifier and a band-pass filter with high-pass and low-pass cutoff frequencies at 10 and 500 Hz, respectively (Myosystem 2000; Noraxon, Oulu, Finland). The length of the needles was adjusted so that the tips penetrated the blubber layer and reached 1–5 mm into the underlying muscle. Blubber thickness was measured before instrumentation using an ultrasound apparatus (SDR 1200; Philips Ultrasound, Santa Ana, CA). A subcutaneous electrode was placed midlaterally on the animal as a grounding electrode. The needles were inserted under local anesthesia [subcutaneous injection of 2–3 ml Xylocaine (10 mg/ml); AstraZeneca, Södertälje, Sweden]. The amplified and filtered EMG signal was then rectified and averaged using a resistor-capacitor filter with a time constant of 10 s before it was conveyed to an analog-to-digital (A/D) converter and data acquisition system that stored the data every 20 s (Lab-Acq Pro and Insta-Trend Pro; Dianachart, Oak Ridge, NJ). Unrectified signals were also recorded, using a universal amplifier with a second band-pass filter (high- and low-pass cutoff frequencies at 10 and 100 Hz, respectively) and a printer sampling at 250 kHz (TA 4000; Gould, Valley View, OH). In other words, we recorded the mean rectified EMG values every 20 s in addition to continuous recording of the unrectified (raw data) signals. Baseline nonbiological noise levels for the electrodes were determined after each experiment by submerging them in physiological saline. Baseline noise was subsequently subtracted from the mean rectified values.
Brain temperature was measured as described by Odden et al. (21): a 30-mm-long and 1.5-mm-thick blind-ended stainless steel pipe was surgically implanted into the brain of the animals under full isoflurane anesthesia. The probe was placed 10 mm laterally to the midline into the left cerebral hemisphere, near the third ventricle (as verified postmortem). Instrumented animals were not used in any experiments until at least 48 h after implantation of the probe. Before each experiment, a copper-constantan thermocouple was introduced into the probe under light sedation [0.6 mg/kg im injection of Zoletil forte vet (tiletamin-zolazepam); Virbac, Carros Cedex, France]. The thermocouple was connected through a thermocouple amplifier with internal temperature reference (AD 595 CD: Analog Devices, Norwood, MA) to an A/D converter and data acquisition system as described for the EMG signals. After sedation, the animals were monitored in air for at least 1 h to allow complete recovery before the series of diving experiments commenced.
Rectal temperature was recorded with a copper-constantan thermocouple that was inserted 20 cm into the rectum of the animal. The thermocouple was connected through a thermocouple amplifier with internal temperature reference to an A/D converter and data acquisition system as described for brain temperature. All thermocouples were calibrated at 0 and 40°C, and the measured temperatures were linearized according to Tøien (26). All temperatures were recorded every 20 s as the mean value of the preceding 20 s.
During the diving experiments, heart rate was recorded using two subcutaneous electrodes placed anterior and posterior to the heart along the dorsal midline, and a third electrode was placed laterally on the animal. The electrodes were connected to a monitoring unit (CM-4008; Medi-Stim, Oslo, Norway) and recorded on a printer (TA 4000; Gould).
In Figs. 1, 3, and 4, EMG data are presented as medians of the sampled mean rectified values for extended periods (minutes). Medians better represent shivering activity than means because the muscle contractions involved in even small movements of the animal generate larger EMG signals than the signals generated by shivering. Bouts of EMG activity due to animal movements would have an unproportional influence on EMG data if means were used instead of medians. Moreover, given the dual source of EMG signals (shivering or physical movements), EMG data are hardly normally distributed, which further warrants the use of medians instead of means. When using median values, the interquartile range is used as a measure of variance (see Fig. 4). Because the absolute magnitude of an EMG signal picked up by an electrode will vary greatly from muscle to muscle and from experiment to experiment (26), a relative change in EMG signal (relative to the starting point of each experiment) was used when comparing different muscles and different experiments (see Fig. 5), and the results are presented as means ± SE. Body temperature data are presented as means ± SD. Comparisons of body core temperatures of shivering and nonshivering animals were made using two-tailed Student's t-test for unpaired samples. Differences in relative mean rectified EMG levels recorded under various experimental conditions were compared using a one-sample t-test. A P value <0.05 was taken to indicate a statistically significant difference.
Shivering in Cold Air
When the seals were exposed to ambient air temperatures of −35°C for extended periods, rectal temperature decreased and shivering ensued when rectal temperature dropped below 35.3 ± 0.6°C (n = 6) (Fig. 1). The three large muscles from which EMG signals were recorded were typically recruited in a sequential manner if rectal temperature continued to decrease. The sequence of recruitment was consistent within but not between animals. In animal K-3/00, shivering always started in the humerotrapezius muscle, and if necessary, spread to the latissimus dorsi and spinotrapezius muscles (Fig. 1), whereas in animal K-6/99, shivering usually started in the latissimus dorsi and only occasionally spread to the other two muscles (Fig. 1).
Shivering During Diving
Of a total of 33 dives in our three animals, 20 were discarded because the animal did not shiver before the dive. These data could therefore not be used to address the hypothesis. On average, rectal temperature dropped 0.5 ± 0.3°C (n = 33) and brain temperature dropped 0.7 ± 0.5°C (n = 13) in response to the 10- to 15-min dives. The average rectal temperature in animals that shivered before a dive was 35.7 ± 0.6°C (n = 19) compared with 36.3 ± 0.5°C (n = 19) in animals that did not shiver before a dive. The average brain temperature in animals that shivered before a dive was 37.3 ± 0.6°C (n = 7) compared with 37.6 ± 0.6°C (n = 9) in animals that did not shiver. These differences in body temperature were statistically significant for rectal temperature (P = 0.004, two-tailed Student's t-test, unpaired samples) but not for brain temperature (P = 0.24). In water, as in cold air, shivering was not necessarily activated in all muscles simultaneously. Figure 2 shows unrectified (raw data) EMG signals from one muscle and heart rate before, during, and after a 10-min dive, followed by a recovery period of 3.5 min and a second dive of a 3-min duration in animal K-5/99. In this case, heart rate dropped from ∼130 beats/min before the dives to ∼20 beats/min during the dives and increased again to 130 beats/min upon emergence. This was more or less the pattern in all dives in all three animals. In the experiment illustrated in Fig. 2, only the latissimus dorsi muscle was activated. The shivering activity of this muscle before the series of dives was clearly suppressed during the dive. After the dive, shivering was quickly reactivated but was again suppressed upon resubmergence, whereafter it was reactivated upon reemergence. This pattern of inhibition and activation of shivering during and after dives, respectively, is also illustrated in Figs. 3 and 4, which show the median values of mean rectified EMG after transformation of the unrectified EMG signal from Fig. 2 before, during, and after dives.
When all data from all muscles in all experiments in which shivering occurred before a dive were combined (Fig. 5), relative changes in median values of mean rectified EMG were employed. Thus 100% was chosen as the value before the dive, and the values obtained during and after the dive were converted to percentages relative to that value in each experiment (Fig. 5). In so doing, we found that there was a noticeable reduction in EMG activity during the dive in all muscles that were shivering before the dive. After the dive, EMG activity always increased, but it was reduced again if diving was repeated (Fig. 5). The reduction in relative EMG activity during dives was statistically significant (P < 0.05, one-sample t-test, hypothesized mean = 100) in all muscles, except for the spinotrapezius muscle during dives and the humerotrapezius muscle during redives (Fig. 5). When data from all muscles were pooled, however, the reduction in EMG activity during diving was significant during both dives and redives.
Shivering in Cold Air
In this study we have shown that seals exposed to ambient temperatures below their lower critical temperature start shivering to increase heat production in the normal mammalian way (Fig. 1). The two largest muscles from which EMGs were recorded, the humerotrapezius and latissimus dorsi, were usually the first to be called upon to shiver, whereas the smallest muscle, the spinotrapezius, was usually the last. Previous studies have shown an increase in oxygen consumption when ambient air temperature drops below −11°C in subadult gray seals (Halichoerus grypus) (7) and −13°C in subadult harbor seals (Phoca vitulina) (11). In the present study, subadult hooded seals were exposed to an air temperature of −35°C, and although we did not exactly determine the lower critical temperature of our animals, we can safely assume that −35°C is far outside their thermoneutral zone. This is also supported by the fact that even though the largest animal (K-6/99) seemed to tolerate a greater hypothermia before shivering was activated than the smaller one (K-3/00; Fig. 1), they both invariably shivered when exposed to that temperature. The mean rectal temperature when shivering first started was 35.3°C, which is between 1 and 2°C below a normal rectal temperature for these animals. Rectal temperature therefore seems to be an acceptable indication of thermal status of seals in air. This also seems to be true in water, because rectal temperature of seals that were shivering just before a dive was significantly lower than in seals that did not shiver. However, rectal temperature in seals that were shivering in water just before a dive was higher, but not significantly higher (P = 0.1), than the rectal temperature at the onset of shivering in air. This apparent difference in the thermal sensitivity of seals in air and water may be misleading, because body temperatures are represented by rectal temperature, which is not always a good, albeit commonly used, indicator of thermal status (see e.g., Ref. 20). If authentic, it is probably caused by differences in the input from central thermoreceptors. Seals have a high hypothalamic thermosensitivity similar to terrestrial mammals of similar size, but they have the lowest sensitivity to thermal stimulation of the skin of all mammals studied (Ref. 25, based on data from Ref. 10). It is therefore likely that differences in brain temperature, more than differences in skin temperature, explain why the seals were shivering in cold water already when rectal temperature was 35.7°C, whereas rectal temperature fell below 35.3°C before they started to shiver at an ambient temperature of −35°C in air.
The reason why the shivering response in the nondiving situation was studied in air instead of in water is that we knew from previous studies of similarly sized arctic seals (8, 18) that water temperatures of 2–4°C would not be below the lower critical temperatures of our seals and, hence, that they would not shiver in such water unless they were diving.
Shivering During Diving
It is well documented that seals have the ability to cool their body core during prolonged diving (3, 12, 17, 21, 22). The present study has shown that diving also inhibits shivering in seals. This may well be rather important, because if shivering were not inhibited, the animal's ability to cool during diving, and hence its ability to save oxygen and thereby its ability to extend the duration of the dives, would be compromised. We suggest that inhibition of shivering is an integral part of the complex “package” of reflexes, which in common often are referred to as the “diving responses” of seals (2). Previous studies have shown that shivering in mammals is inhibited by hypoxia (see e.g., Ref. 9, 16). However, arctic seals are well equipped with both blood and muscle oxygen stores (see e.g., Ref. 5, 19) and are therefore fairly well oxygenated for several minutes into a dive, and because shivering in our animals was inhibited instantly upon submersion, the inhibition of shivering was hardly caused by hypoxia.
Our finding that diving overrides thermoregulation-induced shivering might at first seem contradictory to the findings of Hammel et al. (10), showing that during experimental hypothalamic warming in harbor seals, the diving response did not override thermoregulation-induced vasodilatation in the flippers. However, if we accept the proposal by Blix et al. (3) that body temperature is actively downregulated by means of controlled perfusion of primarily the front flippers, this all makes much more sense. During diving, it is simply advantageous for the animals to get cold. In the experiments of Hammel et al. (10), the animal had its brain heated when it already was striving to get cold, hence the lack of vasoconstriction in the flipper during diving, whereas in our experiments, inhibition of shivering was necessary to ensure that body cooling persisted.
Oxygen consumption during diving can at present not be measured directly, and instead the indirect approach has been to use respirometry to determine oxygen extraction rates before and after a dive (6, 15, 22, 24). During a dive, the extraction of atmospheric oxygen is of course zero, because the animal is not breathing, whereas oxygen consumption by the tissues is still going on. After a dive, oxygen extraction rate is therefore, for a while, well above the predive level, because the animal is “repaying” its oxygen debt. Such studies clearly show that seals that dive for prolonged periods lower their oxygen consumption during diving. This may be a result of the widespread peripheral vasoconstriction that renders most internal organs uncirculated and, hence, partly anaerobic during long dives and/or a Q10 effect caused by body core cooling (3). In estimating the amount of oxygen consumed during the dive, it is not correct to assume that the extra consumption after the dive equals consumption during the dive (15, 22). In fact, our finding of a reduced body core temperature together with inhibition of shivering during the dives implies that a presently unknown fraction of the “apparent oxygen debt” after the dive will consist of oxygen spent on shivering to bring body core temperature back to normal. The dive consumption rates for oxygen obtained so far may therefore be overestimated.
In conclusion, this study has shown that seals have a normal thermoregulatory shivering response to hypothermia in air, whereas diving inhibits shivering and thereby allows body temperature and oxygen consumption to drop. This adaptation may contribute to extend the diving capability of the animal and explain how seals repeatedly are reported to exceed their calculated aerobic dive limit.
This study was supported by Norwegian Research Council Grant 138699/432 and by the Roald Amundsen Centre for Arctic Research at The University of Tromsø.
We thank Dr. Øivind Tøien (Institute of Arctic Biology, University of Alaska, Fairbanks, AK) for advice in setting up EMG measurements and Dr. Esa Hohtola (Dept. of Biology, University of Oulu, Oulu, Finland) for advice on EMG-electrode design.
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 © 2005 the American Physiological Society