Introduction

Asphyxia is a life-threatening condition of reduced or interrupted oxygen absorption from the environment through normal gas exchange organs (e.g., skin, gills, and lungs). For air breathing vertebrates, it can result from being in environments where oxygen is not readily accessible such as underwater or in a low-oxygen atmosphere (e.g., high altitude). Asphyxia causes generalized hypoxia, which primarily affects the tissues and organs most sensitive to reduced convective oxygen transport (brain and heart) and disrupts metabolic and physiological homeostasis (i.e., constancy of the internal environment necessary for the survival of a biological system within its environment). Unlike their terrestrial counterparts, diving mammals and birds are adapted to long interruptions in breathing that enable them to maintain an oxygen-based metabolism and physiological homeostasis while underwater. The latter point is especially important; they not only avoid hypoxic tissue damage and death but also maintain behavioral, physical (e.g., exercise), and physiological functions (e.g., digestion and renal activity) similar to terrestrial animals. The adaptations that enable this remarkable ability occur at many levels, from biochemistry to behavior. There have been many reviews published on adaptations in aquatic mammals and birds for prolonged breath-hold diving (Andersen 1966; Butler and Jones 1982; Blix and Folkow 1983; Elsner and Gooden 1983; Kooyman 1989; Butler and Jones 1997; Ponganis 2011; Castellini 2012). Here, I will focus on what is known about the multi-level integration of these adaptations in marine mammals (primarily seals, sea lions, and dolphins) that maximize the duration of voluntary, aerobic dives.

The dive response

A historical perspective

The concept of a dive response originated in the 1930s, but it was based in part on earlier observations by Bert (1870) and Richet (1899) [for historical reviews, see Andersen (1966), Butler and Jones (1997) and Castellini (2012)]. To explain the prolonged breath-hold ability of aquatic mammals and birds, Irving (1933) hypothesized circulatory adjustments that confined oxygen-rich blood to the brain and heart. He and Scholander then validated this hypothesis experimentally with a series of studies in which aquatic mammals and birds were forcibly submerged (Irving 1938, 1939a; Scholander 1940; Irving et al. 1935, 1941a, b, 1942; Scholander et al. 1942). Their experimental approach was more similar to asphyxiation by drowning than to voluntary apnea during dives made by animals in the wild, but it was the only way to study diving physiology at the time. These studies showed that the pulmonary and cardiovascular changes associated with the dive response include cessation of breathing (apnea), a decrease in heart rate (bradycardia), a concurrent reduction in cardiac output, and peripheral vasoconstriction that maintains central arterial blood pressure while maintaining blood flow primarily to the heart, lungs, and brain. During forced submersion, the bradycardia can be profound (i.e., sometimes less than 10 % of resting heart rate) and organ perfusion is reduced to negligible levels except for the brain, heart, lungs, and adrenal glands (Bron et al. 1966; Kerem and Elsner 1973; Zapol et al. 1979; Blix and Folkow 1983; Hill et al. 1987). The major effect of these cardiovascular changes is to decrease convective oxygen transport to most tissues to conserve oxygen for the brain and heart, two organs that are sensitive to oxygen deficiency and essential for survival when faced with the threat of asphyxia (Elsner and Gooden 1983). A similar asphyxial response occurs in all vertebrates, including fish when they experience hypoxic water or are removed from water and unable to oxygenate their blood (Irving 1939b; Leivestad et al. 1957; Scholander 1964; Andersen 1966). Thus, these cardiovascular adjustments appear to be a primitive response designed to prolong life when oxygen absorption is interrupted and were probably involved in the early stages of the evolution of terrestrial life in the vertebrates (Gordon et al. 1969). Indeed, Scholander (1963) called this striking cardiovascular defense against asphyxiation in vertebrate animals the “Master Switch of Life.” Although it can prolong life, it causes peripheral tissues to switch to anaerobic glycolysis with a simultaneous increase in lactate, reduces or stops abdominal organ (e.g., liver and kidney) function, and disrupts physiological homeostasis (Scholander 1940; Bradley and Bing 1942; Ladd et al. 1951; Bradley et al. 1954; Butler and Jones 1997; Davis et al. 1983). Hence, these circulatory and metabolic adjustments cannot be sustained indefinitely and are stressful to the animal (Soini et al. 1992). However, even during their early work, Irving and Scholander realized that there was probably a difference in the dive response between forced submergence and voluntary dives, but it would be the 1960s before new technology and a little known species of Antarctic seal would enable the study of diving animals in the wild.

Kooyman and Campbell (1972) were the first to show that the dive response during most voluntary dives is less profound than during forced submersion. Kooyman pioneered the use of the isolated ice-hole paradigm to study voluntary dives by Weddell seals (Leptonychotes weddellii) in Antarctica (Kooyman 1965, 1966). The natural habitat for Weddell seals is beneath the shore-fast ice that surrounds the Antarctic continent. By placing a portable hut over a man-made ice hole far from naturally occurring holes and cracks, it was possible to instrument seals and release them into the hole. Seals could make voluntary dives as deep (McMurdo Sound is about 600 m deep) and as long as they desired, but they had to return to the isolated ice-hole to breathe. This behavioral constraint on where the animal surfaced to breathe enabled the deployment of recording instruments, measurement of metabolic rate, and blood sampling before and after voluntary dives, an experimental protocol that was impossible for free-ranging marine mammals at sea. For the first time, heart rate was recorded during the first 45 s of voluntary dives using breakaway electrocardiogram leads. Two important conclusions came from this landmark study. First, bradycardia during dives shorter than 16 min was less severe than during forced submersion and was also more variable (Fig. 1). Second, dives longer than 16 min had a lower heart rate that commenced at the beginning of the dive. Hence, heart rate during most voluntary dives was more variable, but it was graded (instead of Scholander’s “Master Switch” concept) and appeared to be set in anticipation of dive duration. Kooyman and Campbell never gave an explanation for the variability in heart rate for dives shorter than 16 min, but they concluded that only small blood flow changes occur and aerobic metabolism probably continues in most tissues. This prescient conclusion would eventually be validated and change our entire understanding of the dive response.

Fig. 1
figure 1

Relationship of heart rate (first 45 s of dives) to dive duration in adult Weddell seals. Regression line calculated using Bartlett’s regression analysis, P = 0.02 that b < 0. Adapted from Kooyman and Campbell (1972)

Full size image

In addition to new insights into the dive response arising from studies of Weddell seals making voluntary dives, the next significant breakthrough focused on diving metabolism and whether it was primarily anaerobic or aerobic. Because forced submergence was associated with a transition to anaerobic metabolism in peripheral tissues, especially skeletal muscle, it was thought by some researchers that the same metabolic transition occurred during voluntary dives. If this were true, then marine mammals might have an enhanced glycolytic potential due to high activities of glycolytic enzymes and isozymes specifically adapted to function in the anaerobic direction that could quickly switch from low to high activity (for a historical review, see Castellini and Castellini 2004). Studies to validate these ideas were prevalent in the 1970s, but they failed to find consistent and convincing evidence for enhanced glycolytic potential. For a time, the apparent absence of enhanced anaerobic capacity appeared paradoxical: How could marine mammals tolerate such long breath-holds with glycolytic enzyme activities similar to terrestrial mammals? Ultimately, these studies were superseded by new research on Weddell seals, and the paradox was resolved. Using the same experimental paradigm as the heart rate study, Kooyman et al. (1980) showed that there was no post-dive increase in blood lactate after most voluntary dives and that seals normally dive within their aerobic dive limit (ADL), which ranges from 16 to 18 min for this species (Fig. 2). Combined with data showing that the dive response is less intense during voluntary dives, it became apparent that peripheral tissues and organs receive sufficient oxygen to remain aerobic, a result quite opposite from forced submergence where there is a transition to anaerobic metabolism and a cessation of organ function (e.g., kidneys). Additional studies on Weddell seals making voluntary dives showed that the liver, kidneys and gastrointestinal tract continued to function (Davis et al. 1983) and that physiological homeostasis was maintained (Castellini et al. 1988). Taken together, these studies of Weddell seals fundamentally changed our understanding of diving physiology and demonstrated that to understand the significance of biological adaptations (e.g., the dive response), one must study animals in their natural environment. This was the fundamental limitation of earlier laboratory studies. Researchers thought that the dive response during forced submergence was identical to that in the wild, but it turned out to be a primitive response to the threat of asphyxiation.

Fig. 2
figure 2

A summary of the peak arterial lactate concentrations obtained during recovery from various dive durations in Weddell seals making voluntary dives. The diamond at the abscissa zero is the average resting lactate concentration. The term mg % is equivalent to mg 100 ml−1. Adapted from Kooyman et al. (1980)

Full size image

Early research on the dive response makes little mention of the potential influence of exercise (Scholander 1964). This is understandable since forced submergence studies did not allow the animals to swim, although they sometimes struggled against their restraint board (Scholander 1940, p 48). Kooyman and Campbell (1972) briefly mentioned that the heart rate of a captive, 7-month-old Weddell seal making short dives with vigorous swimming was similar to adults making voluntary dives. However, the swim speed and flipper stroke frequency of the adult seals were unknown, so the influence of exertion on the dive response could not be assessed. Jones et al. (1973) measured heart rate in harbor seals (Phoca vitulina) swimming freely in shallow tanks and found that the bradycardia during voluntary dives was only 40–50 % of the predive level. Nevertheless, as Castellini (1985) pointed out, even the relatively moderate bradycardia and associated reduction in cardiac output observed during the majority of voluntary dives seemed diametrically opposed to the normal cardiovascular response (i.e., hyperventilation, tachycardia, an increase in cardiac output and peripheral vasodilation to active muscles) exhibited by terrestrial mammals and birds during exercise.

Measurements of heart rate and total oxygen consumption while swimming submerged and at the surface became possible in the late 1970s with the availability of large water flumes, the equivalent of a treadmill for aquatic mammals and birds. Studies of captive harbor seals, grey seals (Halichoerus grypus) and California sea lions (Zalophus californianus) swimming in a water flume showed that heart rate oscillated between surface tachycardia and submerged bradycardia (Fedak 1986; Williams et al. 1991; Butler et al. 1992). However, the interval between breathing bouts often decreased with increasing swim speed until the animal was swimming at the surface, a behavior that made flume studies difficult to interpret. Although average heart rate over complete dive cycles (i.e., surface and submerged swimming) increased linearly with oxygen uptake (Fig. 3; Butler et al. 1992), there was no clear relationship between submerged heart rate and workload. Studies of heart rate in Weddell seals making voluntary dives and ocean-trained California sea lions were also inconclusive about the effects of exercise because there was no measure of physical exertion other than swim speed, and this turned out to be an unreliable indicator of exertion during voluntary dives (Hill et al. 1987; Ponganis et al. 1997). Nevertheless, the results from these studies indicated that heart rate variation of the typical terrestrial exercise response was being used which prompted Butler (1988) to conclude that a modified form of cardiovascular response to exercise appears to be that most typically employed during the underwater activities of aquatic birds and mammals.

Fig. 3
figure 3

Relationship between heart rate and rate of oxygen consumption in a sub-adult, female California sea lion at rest and while swimming at different speeds and carrying different loads in a water flume. Adapted from Butler et al. (1992)

Full size image

Andrews et al. (1997) studied the relationship between dive depth and heart rate in northern elephant seals (Mirounga angustirostris) making voluntary dives at sea. Starting with predive hyperventilation and tachycardia, there was a rapid decrease in heart rate during descent that plateaued at 20–50 beats min−1 and remained steady until the beginning of ascent. Heart rate began to increase gradually during ascent and then rapidly during the last 15 s before surfacing. The authors concluded that elephant seals do not anticipate dive duration but instead adjust their heart rate throughout a dive. However, since flipper stroke frequency was not recorded, they did not know that elephant seals use different modes of locomotion during voluntary dives: continuous stroking, stroke-and-glide swimming, and prolonged gliding. By attaching a video and data recorder (VDR) to an elephant seal, Davis et al. (2001) showed that they use cost efficient modes of locomotion that depend on buoyancy. A seal uses continuous stroking from the surface to a mean depth of ca. 20 m followed by stroke-and-glide swimming as its lungs collapse, and it becomes negatively buoyant. Prolonged gliding starts at a depth of ca. 60 m so that the seal experiences a near-resting metabolic rate for most of descent to a depth of over 300 m. The seal uses stroke-and-glide locomotion during the initial phase of gradual ascent, then continuous stroking during the final, steep phase. The changes in heart rate described by Andrews et al. (1997) correspond well with an exercise response during the different levels of energetic effort associated with the three modes of swimming except for the last 15 s in which the seal exhibits an anticipatory tachycardia unrelated to stroking. A similar change in heart rate was recorded for California sea lions trained to dive off-shore (Ponganis et al. 1997) and young, free-ranging harbor seal pups (Greaves et al. 2004), but stroking was not recorded. Average swimming speed varies little with swimming mode and is not a good indicator of propulsive effort. What had been missing in the analysis of heart rate was an index of instantaneous effort throughout the dive. This index is stroke frequency, and it would play a major role in our understanding of the dive response during voluntary, aerobic dives.

To summarize this section, original research on the dive response used forced submergence, which elicited a profound bradycardia, peripheral vasoconstriction and a transition to anaerobic metabolism in peripheral tissues that were similar to the asphyxial response that occurs, to varying degrees, in all vertebrates. Later research on Weddell seals making voluntary dives showed that: (1) the dive response was less pronounced, (2) most dives were within the ADL, (3) organs continued to function, and (4) physiological homeostasis was maintained. Furthermore, diving heart rate appeared to increase with exercise even though there was still a dive response.

A new perspective

Results from studies of Weddell seals making voluntary dives raised entirely new questions: If blood oxygen is not sequestered for the brain and heart, then what is the purpose of the dive response during aerobic dives? Why reduce tissue perfusion at all, especially during submerged swimming when muscle metabolic rate is elevated? The normal mammalian response to exercise is to increase cardiac output and muscle perfusion, not to reduce it (Rowell 1986; Wagner 1991). Does the dive response create a conflict between convective oxygen transport and aerobic tissue metabolism, especially in active skeletal muscles? I began to ponder these questions in 1987 and, over the next eight years, developed a model to examine the potential importance of the dive response in optimizing the use of blood and muscle oxygen stores during aerobic dives that involve different levels of muscular exertion (Davis and Kanatous 1999). The numerical model was based on Fick’s principle and integrated cardiac output, regional blood flow, convective oxygen transport, muscle oxy-myoglobin (oxy-Mb) desaturation, and regional metabolic rate. It quantified how the optimal matching or mismatching of convective oxygen transport to tissue oxygen consumption would affect the ADL. We chose an adult Weddell seal on which to base our model because of available data on the diving physiology and metabolism of this species. However, it could be used with other marine mammals by substituting the appropriate physiological and metabolic variables. Given the difficulty of measuring the contributions of blood and muscle oxygen stores to whole-body oxygen consumption rates at different levels of tissue perfusion and muscular exertion in vivo, this model was a logical approach to understanding the physiological basis for the ADL.

Based on model results, we concluded that the primary role of the dive response is to regulate the degree of hypoxia in skeletal muscle so that blood and muscle oxygen (oxy-hemoglobin and oxy-Mb, respectively) stores can be efficiently used at different levels of exercise to maximize aerobic dive duration. Theoretically, this is achieved by adjusting cardiac output and convective oxygen transport to muscle according to the rate of muscle oxygen consumption. At higher levels of muscle oxygen consumption during exercise, heart rate, cardiac output, and muscle perfusion must increase to allow the coordinated depletion of available blood and muscle oxygen stores.

Although this may sound similar to a normal exercise response in terrestrial vertebrates, there are important differences in the diving animal. On the one hand, a dive response with bradycardia is still necessary to reduce convective oxygen transport and lower the partial pressure of oxygen in muscle. This promotes the release of Mb-bound oxygen and maximizes aerobic dive duration. On the other hand, heart rate must increase (i.e., bradycardia is less intense) with the level of physical exertion to ensure that active muscle receives additional blood oxygen (ca. 50 % of the muscle’s oxygen comes from the blood over a broad range of muscle metabolism) to supplement the simultaneous use of Mb-bound oxygen to maintain aerobic metabolism (Davis and Kanatous 1999).

In 2009, we tested the model’s predictions during voluntary dives by using animal-borne instruments that simultaneously recorded heart rate as a continuous electrocardiogram, flipper or fluke stroke frequency using two or three-axis accelerometers, and dive depth in Weddell seals and bottlenose dolphins (Tursiops truncatus) during voluntary dives (Davis and Williams 2012). Stroke frequency was chosen as a metric for physical exertion because it is correlated with locomotory performance and oxygen consumption for a wide range of mammals (Taylor et al. 1980; Williams 1983; Williams et al. 2004). The results showed that heart rate does indeed increase with flipper or fluke stroke frequency during voluntary, aerobic dives (Fig. 4). The results were consistent with model predictions and led us to three conclusions: (1) A dive response is needed to efficiently use blood and muscle oxygen stores to maximize aerobic dive duration, (2) The intensity of the dive response varies inversely with the level of muscle metabolism, and (3) There is a minimum heart rate and concurrent convective oxygen transport needed to maintain aerobic metabolism in organs and tissues other than muscle.

Fig. 4
figure 4

Heart rates for adult Weddell seals (a) and bottlenose dolphins (b) as a function of flipper or fluke stroke frequency, respectively. The heart rates while resting submerged (i.e., 0 stroke frequency) are shown as the average ± s.d. and the dashed line. Heart rate increased linearly with stroke frequency for both species, although the range of values differed. Note the difference in the scales for both plots, with slower swimming seals (a) showing lower stroke frequencies and submerged heart rates than the faster swimming dolphins (b). Each point denotes the average heart rate over 40 s and 5 s intervals for seals and dolphins, respectively. Adapted from Davis and Williams (2012)

Full size image

The first conclusion results from the reliance on two types of molecules to transport and store oxygen in the vertebrate body that have very different oxygen affinities. Hemoglobin (Hb) has a moderate binding affinity for oxygen [an oxygen partial pressure at 50 % saturation or P 50 of about 27 mmHg (3.59 kPa)], while Mb has a very high binding affinity for oxygen (P 50 of 3.7 mmHg (0.49 kPa); Wright and Davis, unpub. obs.]. Unlike terrestrial animals, many aquatically adapted vertebrates rely on an elevated (up to 20-times greater) concentration of muscle oxy-Mb to store from one-third to one-half of the total oxygen used during a dive (Kooyman et al. 1999). For Mb-bound oxygen to become available for aerobic metabolism, the intracellular partial pressure of oxygen in the muscle must be less than 10 mmHg (1.3 kPa); in other words, active muscles must become very hypoxic (but not anaerobic). We know from studies of Weddell seals making voluntary dives that desaturation of Mb begins as soon as the seal submerges (Guyton et al. 1995). At the beginning of a dive when the partial pressure of oxygen in arterial blood is around 100 mmHg (13.33 kPa), the most effective way to make the muscle hypoxic is by ischemia. This is where the dive response plays a major role. Peripheral vasoconstriction reduces convective oxygen transport to the muscle resulting in tissue hypoxia. The partial pressure of oxygen in the muscle becomes low enough for oxygen to dissociate from Mb and diffuse into mitochondria. As the dive progresses, hypoxic hypoxia (i.e., a reduction in arterial oxygen partial pressure also known as hypoxemia) also contributes to the level of muscle hypoxia and Mb desaturation (Davis and Kanatous 1999).

The explanation for the second conclusion results from the balance between convective oxygen transport and muscle oxygen consumption. To maximize the use of available body oxygen stores and the aerobic dive duration over a broad range of muscle metabolic rates, the full use of the oxygen available in the blood (arterial oxygen saturation not less than ca. 38 %, Qvist et al. 1986; Davis and Kanatous 1999; Meir et al. 2009) and total muscle oxygen stores must be completed simultaneously. According to our model, this can be achieved by adjusting cardiac output, most of which results from changes in heart rate (Johansen and Aakhus 1963; Elsner et al. 1964; Ponganis et al. 1990b), according to the level of muscular exertion. Thus, as muscle oxygen consumption rises, a concomitant increase in cardiac output and muscle perfusion (i.e., the dive response is less pronounced) will augment muscle oxy-Mb (Fig. 5). At routine levels of exertion (whole body oxygen consumption ca. 4 ml O2 min−1 kg−1; Kooyman et al. 1971; Castellini et al. 1992; Williams et al. 2004) for a Weddell seal, the maximum ADL occurs when heart rate is 45 % of the resting, predive level (ca. 23 beats min−1). If the seal were resting while submerged, the maximum ADL would occur with a heart rate of 24 % of resting level (ca. 12 beats min−1). Conversely, if metabolism were greater than ca. 7.4 ml O2 min−1 kg−1 (ca. 3.5 times resting metabolism), then the maximum ADL occurs with a heart rate that theoretically exceeds the resting level; that is, no dive response. Based on the model for an exercising Weddell seal, the blood provides about 50 % of the muscle’s oxygen consumption over a broad range of exercise levels (i.e., 1–16 times resting muscle oxygen consumption) with the remainder coming from endogenous oxy-Mb (Davis and Kanatous 1999). As a result, the ADL decreases as muscular exertion increases.

Fig. 5
figure 5

Optimum heart rate (expressed as a percentage of the resting, predive level of 52 beats min−1) as a function of skeletal muscle oxygen consumption and whole-body oxygen consumption rate. Optimum refers to the heart rate that gives the maximum aerobic dive limit (ADL). Adapted from Davis and Kanatous 1999

Full size image

Finally, the third conclusion results from the convective oxygen requirements of other tissues and organs to maintain aerobic metabolism. The model predicted that heart rate in Weddell seals would not decrease below 24 % of the predive, resting level (i.e., 12 min−1; Fig. 5) to maintain aerobic metabolism in tissues and organs other than the muscles, and this was almost identical to the measured minimum heart rate (i.e., 11 min−1; Fig. 4) during voluntary dives when the seal stopped swimming at depth and drifted (Davis and Williams 2012). At this level of bradycardia, the theoretical extraction coefficient of oxygen from the arterial blood (arterial − venous O2 difference/arterial O2 content) is less than 60 % for the splanchnic organs, kidneys, heart and brain which indicates sufficient convective oxygen delivery to support aerobic metabolism (Davis and Kanatous 1999). This is critical for the maintenance of physiological homeostasis in Weddell seals and other long duration divers that may spend over 90 % of their time at sea submerged (Butler and Jones 1997; Kooyman et al. 1980; Fedak 1986; Le Boeuf et al. 1989; Hindell et al. 1992) and is consistent with normal renal and hepatic function during such dives (Davis et al. 1983).

The need to match an increase in muscle oxygen consumption with an increase in convective oxygen transport explains why long duration divers such as the Weddell seal, which maintain a low level of metabolism by using cost-efficient modes of locomotion (Williams et al. 2004), have a more pronounced dive response than short duration divers such as the bottlenose dolphin, which have a higher range of stroke frequencies and swim speeds as well as a lower muscle Mb concentration (Noren et al. 2001).

In diving mammals and birds, the concentration of Mb in the muscles represents another important adaptation for exercising underwater. Only those species with a comparatively large muscle oxygen store will benefit from a dive response. Davis and Williams (2012) hypothesized that the evolution of high Mb concentrations in the muscles of diving mammals and birds as a means of increasing oxygen stores and extending aerobic dive duration was only possible because the dive response already existed to protect the animals against asphyxia.

To summarize this section, the physiological model of Davis and Kanatous (1999) indicates that the cardiovascular adjustments known as the dive response enable marine mammals to balance the conflicting demands of: (1) optimizing the distribution and use of blood and muscle oxygen stores to maximize the ADL over the normal range of diving metabolic rates and (2) ensuring that active muscle receives adequate oxygen as exertion increases. Furthermore, the results from the simultaneous measurement of heart rate and flipper or fluke stroke frequency in Weddell seals and bottlenose dolphins during voluntary dives are consistent with model predictions and the results from other studies examining the ability of diving mammals and birds to support aerobic muscle metabolism while swimming submerged (Kooyman et al. 1980; Butler 1988; Bevan and Butler 1992; Guyton et al. 1995; Butler and Jones 1997). During completely aerobic, voluntary dives, the cardiovascular adjustments that maximize dive duration combine elements of both a dive response and an exercise response, with one or the other predominating depending on the level of physical exertion (Fig. 6). These results indicate that the dive response, which is a primitive cardiovascular response to protect animals from asphyxia, is integrated with the equally primitive exercise (i.e., fight or flight) response to regulate blood flow in a manner that maintains aerobic metabolism at different levels of submerged exercise. This cardiovascular adaptation may have enabled the evolution of an elevated muscle Mb concentration as a significant body oxygen store despite its high affinity for oxygen relative to Hb. Furthermore, the dive response, as we now understand its role in animals making voluntary dives, appears to have evolved independently in at least three taxonomic Orders (Cetaceans, Sirenians, and Carnivores) of marine mammals and in diving birds from the ancient, vertebrate response to asphyxia.

Fig. 6
figure 6

A comparison of pulmonary and cardiovascular changes and their influence on convective oxygen transport to muscle during exercise in a terrestrial mammal (a) and Weddell seal during forced submergence (b) or aerobic, voluntary dives (c). For comparison, the terrestrial mammal and Weddell seal are assumed to have a body mass of 450 kg (Davis and Kanatous 1999). The symbol, font size and arrow thickness indicate relative magnitude of the variables. The large lungs in a indicate hyperventilation and enhanced convective oxygen transport, while ventilation ceases and there is also lung compression at depth with little oxygenation of the blood in b and c. The arrow from the lungs to the heart represents the pulmonary vein. The size of the heart symbol indicates heart rate (HR, beats min−1 or bpm), with a tachycardia in a, a profound bradycardia in b, and a variable bradycardia (HR = 11–40 bpm, Fig. 4a) depending on the level of physical exertion in c. The red diamond indicates a surface, resting heart rate for Weddell seals of 52 bpm (Zapol et al. 1979), and the same resting heart rate is assumed for the terrestrial mammal in a. The size of the arrow from the heart symbol indicates relative arterial blood flow, with peripheral vasodilation and increased flow in a, severe peripheral vasoconstriction and reduced flow in b, and variable vasoconstriction and blood flow depending on the level of physical exertion in c. The color of the blood vessels indicate relative hemoglobin (Hb) concentration and that of the muscle myoglobin (Mb) concentration, which are relatively low in a and high in b and c. Finally, the arrows in the muscle indicate the relative increase in oxygen consumption (\( \dot{V}{\text{O}}_{2} \)) and lactate concentration. Adapted from Davis and Williams (2012)

Full size image

The role of globins and body oxygen stores

Globins are small respiratory proteins that reversibly bind oxygen and have been identified in bacteria, plants, fungi, and animals. Most globins facilitate or augment oxygen delivery for aerobic metabolism, and five types have been identified in vertebrates: Hb, Mb, neuroglobin (Ngb), cytoglobin (Cybg) and androglobin (Adgb) (Pesce et al. 2002; Hoogewijs et al. 2012; Storz et al. 2012). Hb is located in red blood cells and transports oxygen in the circulatory system. Mb, which is present in cardiac and skeletal muscle, serves as an oxygen store and facilitates oxygen diffusion. Hb and Mb phylogenetically separated over 550 million years ago and may have evolved in metazoans in response to aerobic metabolic demands, thereby enabling animals to become larger and more active (Pesce et al. 2002). Ngb, Cybg and Adgb occur at much lower concentrations in some cells (e.g., neurons, some endocrine tissues and fibroblasts), but their physiological roles are less clear (Burmester et al. 2002; Storz et al. 2012). In this review, I will focus on the role Hb and Mb, although an adaptive role for Ngb and the other globins may eventually be found in marine mammals (Williams et al. 2008; Mitz et al. 2009; Schneuer et al. 2012).

The oxygen affinity of Hb scales inversely with body mass, but the average P 50 is about 27 mmHg (3.60 kPa) for a 100 kg, low altitude-adapted, terrestrial mammal (Snyder 1983). The P 50 for marine carnivores ranges from 25 to 31 mmHg (3.33–4.13 kPa) with no consistent differences among seals, sea lions, fur seals, walrus (Odobenus rosmarus) and the sea otter (Enhydra lutris) (Lenfant et al. 1970; Qvist et al. 1981; Snyder 1983; Wilford et al. 1990; Meir et al. 2009; McDonald and Ponganis 2013). Compared with humans, dogs and pigs, the P 50 of seal and sea lion Hb shows little temperature sensitivity but a small increase in the Bohr effect (Lenfant et al. 1970; Qvist et al. 1981; Willford et al. 1990). The latter would facilitate oxygen offloading if acidosis resulted during a dive, but this is probably uncommon since most dives are aerobic. Small cetaceans may have a left shifted P 50 [20–25 mmHg (2.67–3.33 kPa)] while larger cetaceans may be right shifted [25–32 mmHg (3.33–4.27 kPa)] (Snyder 1983), and there is evidence for a small increase in the Bohr effect compared to dogs and pigs (Dhindsa et al. 1974). However, different methods for measuring P 50 may account for the range of values for cetaceans. Relative to low altitude-adapted, terrestrial mammals of similar size, marine mammals show no consistent difference in Hb P 50 that could be considered an adaptation for hypoxia during diving. In contrast, high altitude-adapted, terrestrial mammals such as the vicuña (Vicugna vicugna) show a pronounced left shift in Hb P 50 [19.5 mmHg (2.60 kPa)] that is considered an adaptation for maintaining arterial blood oxygen saturation >90 % at altitudes up to 4,800 m [ambient PO2 = 91 mmHg (12.1 kPa)] (Hall et al. 1936; Schmidt-Nielsen 1997, p.74). However, at sea level [ambient PO2 = 159 mmHg (21.2 kPa)], where marine mammals breathe and oxygenate their blood, a Hb P 50 of 25–31 mmHg (3.33–4.13 kPa) results in nearly full (ca. 96 %) saturation of arterial blood at the beginning of a dive, so a left shift in Hb P 50 would not significantly enhance saturation.

Although the P 50 of Mb in mammals (average = 3.7 mmHg [0.49 kPa] ± 0.33 s.d.) is much lower than Hb, there are only small differences among terrestrial and marine species (Wright and Davis, unpub. obs.), and the biological significance of these differences is not understood. However, the molecular structure of Mb relative to oxygen binding is highly conserved to maximize the ADL in marine mammals (Dasmeh et al. 2013). The low P 50 maintains a gradient for oxygen to diffuse from red blood cells in the capillaries into muscles. However, it also means that endogenous oxygen bound to Mb is only available to the muscles (i.e., no reverse diffusion into blood) and not to other tissues. Because of the difference in P 50 between Mb and Hb, the dive response is required optimize the use of blood and muscle oxygen stores over the normal range of diving metabolic rates to maximize the ADL at each level of exertion (see above). In his early review of the dive response, Andersen (1966) explained that the difference in oxygen affinity between Hb and Mb necessitated peripheral vasoconstriction to make Mb-bound oxygen available for muscle metabolism. However, he failed to understand the graded nature of the dive response during exercise and that active muscle relies on both convective oxygen transport and endogenous oxy-Mb to maintain aerobic metabolism.

Although the oxygen affinities of Hb and Mb in marine mammals do not differ from most terrestrial mammals, the amount of these globins is elevated to increase oxygen stores in blood and muscle. This is one of the principal adaptations in marine mammals for extending ADL (the other way is to reduce the metabolic cost of diving; see “Morphological and behavioral adaptations to reduce the energetic cost of diving”). Of all the adaptations that marine mammals exhibit for maintaining aerobic metabolism during dives, they never reacquired the ability to breathe water. As a result, they carry all of the oxygen that will be available during the dive in their blood, muscle, and lungs; this is the equivalent of their SCUBA tank. Compared to terrestrial mammals, marine mammals generally have an increased blood volume, hematocrit (Hct), Hb concentration [due primarily to an increased mean cell volume (MCV) and mean cell Hb (MCH)] and muscle Mb concentration (Lenfant et al. 1970; Kooyman and Ponganis 1998; Dolar et al. 1999; Castellini et al. 2006) (Table 1), although this varies with species and age (Noren et al. 2001, 2002; Burns et al. 2005, 2007; Richmond et al. 2006; Clark et al. 2007; Fowler et al. 2007; Spence-Bailey et al. 2007; Prewitt et al. 2010; Mirceta et al. 2013). Factors that limit the maximum concentration of Hb in the blood and Mb in the muscles are not completely understood (Castellini et al. 2006; Wright and Davis 2006; Mirceta et al. 2013), but the degree to which they are elevated varies among the pinnipeds, cetaceans, sirenians and the sea otter, and this influences their total body oxygen stores (Fig. 7). The domestic cow is as a good example of a sedentary, terrestrial mammal with unlimited access to atmospheric oxygen near sea-level, so large oxygen stores are not necessary. Pinnipeds and cetaceans that make the longest dives [e.g., Weddell seal, northern elephant seal (Mirounga angustirostris) and sperm whale (Physeter macrocephalus)] carry most of their oxygen in the blood (2–3-fold increase in blood volume and a 2-fold increase in Hb concentration) and muscle (>10-fold increase in Mb concentration) with relatively little in the lungs. Pinnipeds and cetaceans that are shorter duration divers [e.g., California sea lion, Northern fur seal (Callorhinus ursinus), walrus and bottlenose dolphin] have smaller oxygen stores that are more equally divided among the lungs, blood and muscle. Harbor seals have body oxygen stores that are intermediate between short and long duration divers. Most of these marine mammals have lung volumes that are similar to terrestrial mammals of the same size (Kooyman et al. 1973). For those species with only modest increases in blood and muscle oxygen stores (sea lion, fur seal, walrus and bottlenose dolphin), lung oxygen contributes significantly (ca. one-third) to total body oxygen stores. In contrast, lung oxygen is a small fraction of the total body oxygen stores for the long-duration divers (Weddell seal, elephant seal, sperm whale) because blood and muscle oxygen stores are so large and these species exhale at the beginning of a dive, thereby reducing diving lung volume. This is not true for the sea otter, which has unusually large lungs (3-times larger than terrestrial mammals of similar mass; Lenfant et al. 1970) that serve as a significant store of oxygen (ca. two-thirds of total) during relatively short (ca. 2 min) duration dives (Wolt et al. 2012). Although sea otters dive on an inhalation, submerged lung volume has not been measured, so the actual contribution of lung oxygen to the total oxygen stores remains uncertain and may be less than shown in Fig. 7. Their large lungs make sea otters positively buoyant throughout the dive, so they must stroke continuously except during ascent which augments their already high (2–3 times predicted) resting metabolic rate. Manatees have a low blood Hb concentration (Alsina-Guerrero 2011) and little muscle Mb (Scholander and Irving 1941) (Table 1). As a result, they have small body oxygen stores more similar to a cow, but their very low resting metabolic rate (0.4-times predicted) and sedentary behavior enable them to dive aerobically for ca. 2 min and longer (Gallivan and Best 1980).

Table 1 Determinants of body oxygen stores for a sedentary terrestrial mammal (cow, Bos, 450 kg), manatee (Trichechus manatus, 350 kg), California sea lion (Zalophus californianus, 35 kg), fur seal (Callorhinus ursinus, 154 kg), walrus (Odobenus rosmarus, 65 kg), harbor seal (Phoca vitulina, 24 kg), Weddell seal (Leptonychotes weddellii, 450 kg), elephant seal (Mirounga angustirostris), Atlantic bottlenose dolphin (Tursiops truncatus, 200 kg), sperm whale (Physeter macrocephalus, 10,000 kg), sea otter (Enhydra lutris, 28 kg)
Full size table
Fig. 7
figure 7

Magnitude (ml O2 kg−1) and distribution (lungs, blood and muscle) of body oxygen stores for a sedentary terrestrial mammal (cow, 450 kg), sea otter (Enhydra lutris, 28 kg). California sea lion (Zalophus californianus, 35 kg), fur seal (Callorhinus ursinus, 154 kg), harbor seal (Phoca vitulina, 24 kg), Weddell seal (Leptonychotes weddellii, 450 kg), elephant seal (Mirounga angustirostris), walrus (Odobenus rosmarus, 65 kg), Atlantic bottlenose dolphin (Tursiops truncatus, 200 kg), sperm whale (Physeter macrocephalus, 10,000 kg), manatee (Trichechus manatus, 350 kg). Adapted from Scholander and Irving (1941), Lenfant et al. (1970), Kooyman (1989) and Kooyman and Ponganis (1999)

Full size image

Tissue level adaptations for maintaining aerobic metabolism in skeletal muscle

Convective oxygen transport and capillarity

During an aerobic dive, skeletal muscle must become hypoxic (but not anoxic) for oxygen to dissociate from Mb, and this begins as soon as the seal submerges (Guyton et al. 1995). To accomplish this, the dive response reduces convective oxygen transport (ischemic hypoxia) to muscle depending on the level of exertion (Fig. 6). As the dive progresses, hypoxic hypoxia (i.e., a reduction in arterial oxygen partial pressure) also contributes to the level of muscle hypoxia and Mb desaturation (Davis and Kanatous 1999). In addition, capillary density around the muscle fibers in harbor and Weddell seals is ca. 60 % or less than in the skeletal muscles of terrestrial mammals (Kanatous et al. 2001, 2002) (Fig. 8). Reduced capillarity in skeletal muscle creates the equivalent of an anatomical shunt that further reduces convective oxygen transport (contributing to muscle hypoxia) while maintaining convective oxygen transport to organs and tissues without myoglobin. A muscle poised for aerobic metabolism with reduced capillarity would be difficult to reconcile in a terrestrial mammal or bird (Conley et al. 1987), but it is consistent with the need to reduce convective oxygen transport during voluntary dives for the efficient use of blood and muscle oxygen stores. Unfortunately, equivalent data are not available for sea lions and dolphins, which make shorter dives, exhibit a less pronounced dive response, and have higher routine swim speeds (Ponganis et al. 1990a, 1997; Davis and Williams 2012).

Fig. 8
figure 8

Light micrographs comparing the capillary density in cross-sections of skeletal muscles with similar volume densities of mitochondria between the a harbor seal (Phoca vitulina) (M. longissimus dorsi) and b dog (Canis familiaris) (M. gastrocnemius). Despite similar volume densities of mitochondria, the seal muscle has 60 % fewer capillaries. Adapted from Kanatous et al. (2001)

Full size image

Myoglobin distribution

Due to reduced convective oxygen transport during a dive and the low P 50 of myoglobin, the partial pressure gradient for oxygen diffusion within the muscle is very low. As a result, the rate of intracellular oxygen diffusion, as described by Fick’s equation for diffusion (Schmidt-Nielsen 1997), is low. Although the high myoglobin concentration facilitates oxygen storage and facilitated diffusion, histological studies have revealed other possible adaptations for maintaining aerobic metabolism. There are significant differences in the Mb concentration along the length of the primary swimming muscles (epaxial musculature) in harbor, harp (Pagophilus groenlandicus) and hooded (Cystophora cristata) seals (Polasek et al. 2006; Lestyk et al. 2009). The middle and caudal regions have significantly higher concentrations of Mb than the cranial region (Fig. 9). In addition, among five species of cetaceans [dusky dolphin (Lagenorhynchus obscurus), false killer whale (Pseudorca crassidens), stripped dolphin (Stenella coeruleoalba), Indo-Pacific humpback dolphin (Sousa chinensis) and Indo-Pacific bottlenose dolphin (Tursiops aduncus)], the interior of the epaxial musculature lying closest to the vertebrae shows a significantly higher myoglobin concentration than the exterior (Polasek and Davis 2001). Dolar et al. (1999) found significantly higher concentrations of Mb in swimming as compared to non-swimming muscles in Fraser’s dolphins (Lagenodelphis hosei), spinner dolphins (Stenella longirostris), and a pygmy killer whale (Feresa attenuata). These results show that myoglobin is not homogeneously distributed among and within the muscles of pinnipeds and cetaceans, and the concentrations may be highest in those areas that produce greater contractile force and consume more oxygen during submerged swimming. Enhancing oxygen stores in those areas of the musculature that work the hardest would enhance the maintenance of aerobic metabolism during submerged swimming.

Fig. 9
figure 9

Contours of myoglobin concentrations (mg g−1) from one harbor seal. The orientations of the contours are from the head looking toward the tail. CR cranial, MID mid-dorsal, CA caudal. Adapted from Polasek et al. (2006)

Full size image

Fiber types

Based on histochemical staining and immunohistochemical analysis of myosin heavy chains, there are three major fiber types in skeletal muscle: slow twitch oxidative (Type I), fast twitch oxidative (Type IIa), and fast twitch glycolytic (Type IIx/b,c,d depending on species) (Staron 1997). Type I fibers have an increased concentration of Mb and number of mitochondria and are used in sustained, low energetic activity. Type IIa fibers also have an increased concentration of Mb and number of mitochondria, but use ATP at a faster rate, have a higher contraction velocity and are used in more energetic activities than Type 1 fibers. Finally, Type IIx/b,c,d fibers have little Mb, few mitochondria, a very high contraction velocity, and are associated with burst-type activities. Individual muscles are a mixture of these three fiber types depending on their normal function. The metabolic poise for these three fiber types is not well characterized for marine mammals, but generally Type I fibers are oxidative and use lipids for ATP production, Type IIx/b fibers use glucose for anaerobic ATP production, and Type IIa fibers are probably oxidative but with an enhanced anaerobic capacity (see below for a discussion of enzyme activities). Generally, the higher Mb concentration is associated with a greater aerobic capacity (Scott et al. 2001).

For Weddell and harbor seals, most muscle fibers are slow twitch oxidative (Type I) and fast twitch oxidative (Type IIa) in both primary (M. longissimus dorsi and hind limb) and secondary (M. pectoralis) swimming muscles, although the percentages of these fiber types vary depending on species and muscle (Kanatous et al. 2002; Watson et al. 2003) (Fig. 10). Most noteworthy is the complete absence of fast twitch glycolytic (Type IIx/b) fibers (Fig. 11). These results indicate that seal skeletal muscle is composed primarily of oxidative fibers that are poised for aerobic metabolism. In contrast, northern fur seals have a higher (ca. 40 %) percentage of fast twitch glycolytic (Type IIx/d) fibers in the primary (M. pectoralis) swimming muscles but a lower percentage (ca. 10 %) in secondary (M. longissimus dorsi) swimming muscles, with the majority being primarily fast twitch oxidative (Type IIa) and fewer slow twitch oxidative (Type I) fibers (Shero et al. 2012). A similar percentage of fast twitch glycolytic fibers may also occur in the epaxial muscle of the Pacific white-sided dolphin and the hind limb muscle of the sea otter, although histochemical staining techniques make it difficult to clearly distinguish between Type IIa and Type IIx/b fibers (Ponganis and Pierce 1978; Ponganis pers. com.). Differences in fiber type composition among marine mammals may result from differences in routine dive duration and swim speed, with Weddell and harbor seals making longer dives with lower flipper stroke frequencies and fur seals and dolphins making shorter dives with higher flipper and fluke stroke frequencies and burst-type activities. Nevertheless, the majority of muscle fibers are oxidative.

Fig. 10
figure 10

Histogram showing fiber-type composition (percentage of total fiber number) in the primary swimming (M. longissimus dorsi and hind limb) and secondary swimming (M. pectoralis) muscles of Weddell seals. Note the lack of fast-twitch glycolytic (Type IIB) fibers in all the muscles measured. Adapted from Kanatous et al. (2002)

Full size image
Fig. 11
figure 11

Serial cross sections from the epaxial muscles of a harbor seal stained for myosin heavy chain isoforms using a series of monoclonal antibodies. a Type I (slow twitch oxidative) fibers, b Type IIa (fast twitch oxidative) fibers and c Type IIb (fast twitch glycolytic) fibers. Lack of stain in c indicates an absence of Type IIb fibers in this cross section of the muscle. Note that the three sections were cut from the same sample, so the individual fibers are the same. Bar 50 μm. Adapted from Watson et al. (2003)

Full size image

Mitochondrial volume density

Mitochondrial volume density in swimming muscles of harbor seals (9.7 %), northern fur seals (8.8 %) and Steller sea lions (6.2 %, Eumetopias jubatus) is 1.7–2.2-fold greater than most terrestrial mammals of comparable size but similar to athletic dogs and ponies (Kanatous et al. 1999; Fig. 12). In contrast, mitochondrial volume density in adult Weddell seals (3.1 %) is similar to most terrestrial mammals (Kanatous et al. 2002) and the emperor penguin (Aptenodytes forsteri), another long-duration diver (Ponganis et al. 1997). In harbor seals, mitochondrial volume density is significantly greater (1.2-fold) in the deep regions of the epaxial muscle closer to the vertebrae, possibly associated with greater sustained force production and oxygen consumption during submerged swimming (Watson et al. 2003). For all of these pinnipeds, 95 % of the mitochondria are interfibrillar and only 5 % subsarcolemmal. As a result, mitochondria are distributed uniformly throughout the muscle fiber and not concentrated near the sarcolemma and close to capillaries (Fig. 13). In contrast, subsarcolemmal mitochondria usually account for ca. 10 % of the total mitochondrial volume in sedentary terrestrial mammals and in excess of 30 % in athletic species (Kayar et al. 1989). Because the mitochondrial uptake of oxygen during aerobic dives is diffusion limited due to low intracellular PO2, increased mitochondrial densities and their uniform distribution throughout the muscle fiber increases the oxygen sink and decreases the diffusion distance for oxy-myoglobin, which is evenly distributed throughout the muscle fiber, thereby increasing mass flux of oxygen into mitochondria based on Fick’s Law of diffusion. This cellular adaptation facilitates the use of the large muscle oxygen store to maintain aerobic metabolism despite the high P 50 of Mb. These results also indicate that increased mitochondrial density in skeletal muscle may be enhanced to a similar degree in mammals adapted to either low muscle PO2 (breath-hold diving), increased oxygen demand (athletic terrestrial mammals) or a combination of both.

Fig. 12
figure 12

Percent volume density of mitochondria of locomotory muscles in relation to body mass. Regression line shows relationship for vastus medialis, a primary locomotory muscle, for a wide variety of terrestrial mammals (circles) ranging in size from 0.04 kg (dwarf mongoose, Helogale pervula) to 450 kg (steer, Bos taurus) (Mathieu et al. 1981; Kayar et al. 1989). Dashed lines, 99 % confidence intervals for the regression.  % volume densities of mitochondria in swimming muscles of sea lions, fur seals, and seals were 1.7, 1.9, and 2.1 times greater, respectively, than values predicted based on their body size.  % volume densities of mitochondria in vastus medialis of dog (Canis familiaris, 10.7 %) and pony (Equus caballus, 6.8 %) were 2.2 and 1.8 times greater, respectively, than predicted values and were similar to those of comparably sized pinnipeds. Adapted from Kanatous et al. (1999)

Full size image
Fig. 13
figure 13

Electron micrographs of swimming muscles of: a harbor seal and b northern fur seal. M mitochondria, F myofibril, L lipid droplet. Adapted from Kanatous et al. (1999)

Full size image

Enzyme activities

The aerobic metabolic poise of skeletal muscle indicated by fiber type composition and mitochondrial volume density is also apparent in enzyme activities. The activity of citrate synthase (CS), an important enzyme in the Kreb’s cycle and an indicator of aerobic capacity, is 1.1–1.5-fold greater in swimming muscles than in non-swimming muscles of harbor seals, northern fur seals and Steller sea lions, and this activity is correlated with mitochondrial volume density (Kanatous et al. 1999). In harbor seals, there are also significant differences in CS activities along the length of the epaxial muscles, with higher activities in the middle and caudal regions compared to the cranial region (Polasek et al. 2006). Average CS activity in the epaxial muscles of harbor seals is similar to that in the locomotory muscles of the athletic dog indicating a high aerobic capacity compared to many terrestrial mammals. This contrasts with CS activity in the epaxial muscle of the Weddell seal, which is about 50 % of that in dog locomotory muscle. Compared with shorter-duration divers, the lower aerobic capacity of Weddell seal muscle, which is similar to sedentary terrestrial mammals of similar size, may reflect their energy-conserving modes of locomotion and longer dive durations (Kanatous et al. 2002).

β-Hydroxyacyl coenzyme A dehydrogenase (HOAD) catalyzes one reaction in the β-oxidation of fatty acids in mitochondria and is a measure of the capacity for fatty acid catabolism. The activities of HOAD are 1.0–1.3-fold greater in the swimming muscles than in the non-swimming muscles of harbor seals, northern fur seals and Steller sea lions, and this activity is correlated with mitochondrial volume density (Kanatous et al. 1999). HOAD activities are similar in the swimming muscles of harbor seals and fur seals, but approximately 2-fold greater than in Steller sea lions and 4–5-fold greater than in the cotton rat (Sigmodon hispidus). Harbor and Weddell seals have significantly higher (3–4-fold) HOAD activities and much smaller CS/HOAD ratios in the epaxial muscles (1.2 and 0.3, respectively) than in the locomotory muscles (1.5–7.2) of dogs (Kanatous et al. 2002; Polasek et al. 2006). When scaled to tissue-specific resting metabolic rate (RMR), HOAD activity shows an even greater (6-fold) elevation in harbor seals compared to dogs (Polasek et al. 2006). These results indicate a higher capacity for lipid metabolism, which is characteristic of fuel homeostasis in seals (Irving et al. 1935; Davis et al. 1991a).

Finally, lactate dehydrogenase (LDH) is a key glycolytic enzyme that reversibly catalyzes the conversion of pyruvate to lactate and is an indicator of anaerobic capacity. In harbor and Weddell seals, LDH activity is elevated (1.4–2-fold) in both locomotory and non-locomotory muscles compared to the dog (Kanatous et al. 2002; Polasek et al. 2006). When scaled to tissue-specific RMR, LDH activity shows an even greater (2.5–3.5-fold) elevation in harbor seals compared to dogs (Polasek et al. 2006). These results differ from those of Castellini et al. (1981) who concluded that the skeletal muscles of marine mammals do not have elevated activities of LDH and that they do not possess unusually high capacities for anaerobic glycolysis. Nevertheless, the LDH/CS ratios are ca. 3-fold higher in the locomotory muscles of harbor and Weddell seals compared to the dog (Kanatous et al. 2002; Polasek et al. 2006). Despite a metabolic poise that is geared primarily for aerobic metabolism, seals appear to have a glycolytic capacity that is greater than that in the dog and which may have survival advantages when they exceed their ADL.

Lipid droplets

The primary source of lipids for muscle metabolism are fatty acids and triglycerides (as lipoproteins) that are transported in the blood or triglycerides that are stored as lipid droplets (intramyocellular lipids) between mitochondria and myofibrils (Fig. 13). Triglycerides in lipid droplets are hydrolyzed by lipase, and the fatty acids enter the β-oxidation pathway in the mitochondria resulting in the production of ATP. The volume density of lipid droplets is less than 1 % in the swimming muscles of harbor seals (0.13–0.20 %), Weddell seals (0.16 %), fur seals (1.0 %), and Steller sea lions (0.10 %) but greater than in locomotory muscles in dogs (0.03–0.10) and ponies (0.01 %) (Hoppeler et al. 1987; Kanatous et al. 1999; Kanatous et al. 2002; Watson et al. 2007). In terrestrial mammals, lipid droplets act as a fuel reservoir when the rate of delivery of plasma free fatty acids and triglycerides is less than the rate of use in the muscles, especially during submaximal exercise (Staron et al. 1989; Horowitz and Klein 2000). The role of lipid droplets as a source of energy for muscle metabolism is discussed in “Fuel Homeostasis”, but they potentially provide enough energy to support several hours or more of muscle metabolism at routine levels of exertion (i.e., 5–6 times the estimated resting muscle metabolic rate of 1.4 ml O2 min−1 kg−1; Fig. 5) during aerobic dives (Davis and Kanatous 1999; Kanatous et al. 1999).

Tissue level adaptations for maintaining aerobic metabolism in other organs

To varying degrees, the heart, splanchnic organs, and kidneys of harbor seals show adaptations for aerobic lipid metabolism, but with a glycolytic capacity that is greater than in dogs (Fuson et al. 2003; Polasek et al. 2006).

Heart

Similar to skeletal muscle, Mb concentration in the seal heart is 10-fold greater than in the dog (Polasek et al. 2006). Although mitochondrial volume density in the mammalian heart is higher (ca. 18 volume %) than in skeletal muscle, there are no significant differences in mitochondrial volume density or CS activity between seals and dogs (Fuson et al. 2003; Polasek et al. 2006). Harbor seals have significantly greater HOAD activity (1.1–2.6-fold), LDH activity (1.1–1.3-fold) and LDH/CS ratios (1.31-fold) compared to dogs. When scaled to tissue-specific RMR, the difference in HOAD activity (3.2-fold) and LDH activity (1.5-fold) between seals and dogs is even greater. Glycogen levels may also be elevated ca. 2-fold compared to terrestrial mammals (Kerem et al. 1973).

Splanchnic organs

When scaled to tissue-specific RMR, the liver, stomach and intestine of harbor seals have significantly higher (1.5–2.7-fold) mitochondrial volume densities compared with dogs (Fuson et al. 2003). Harbors seals also have significantly greater CS activity in the liver (2.5-fold), HOAD activity in the liver and intestine (8.5 and 1.8-fold, respectively) and LDH activity in the liver and intestine (5.5 and 2.4-fold, respectively) compared with dogs. In the liver, the ratios of CS/HOAD are less (1.2 and 4.0, respectively) and LDH/CS greater (85 and 38, respectively) in seals compared with dogs.

Kidneys

When scaled to tissue-specific RMR, the kidneys of harbor seals have a significantly greater (1.5-fold) mitochondrial volume density compared with dogs (Fuson et al. 2003). However, there are no significant differences in the activities of CS, HOAD, and LDH.

Renal and splanchnic organ function

In mammals, the kidneys and splanchnic organs (liver, gastrointestinal tract, pancreas, and spleen) receive ca. 61 % of cardiac output under resting conditions (Sapirstein 1958). The functions of the kidneys (e.g., water and electrolyte balance, nitrogenous waste removal, endocrine hormone production) and liver (e.g., detoxification, protein and lipoprotein synthesis, bile production) serve whole-body, homeostatic functions while the gastrointestinal tract and pancreas are involved in the chemical breakdown and absorption of food and hormone production. The spleen serves primarily as a blood filter, but it also acts as a storage site for red blood cells in some marine mammals.

Ortiz (2001) provided a good review of research on renal anatomy, function, and osmoregulation in marine mammals. There were many studies of renal function in harbor seals during the 1940s and 1950s (Hiatt and Hiatt 1942; Bradley and Bing 1942; Ladd et al. 1951; Bradley et al. 1954; Schmidt-Nielsen et al. 1959; Murdaugh et al. 1961). Beginning in the late 1970s, research shifted to renal function and osmoregulation during prolonged fasting, primarily in adult and newly weaned northern elephant seals (see Ortiz 2001 for a review). However, all of these studies focused on animals resting on land (post-prandial, post-absorptive, prolonged fasting) or during forced submergence. Under resting, post-absorptive conditions, renal plasma flow (RPF; 5.9 ml min−1 kg−1; Bradley and Bing 1942) and glomerular filtration rate (GFR; 1.2 ml min−1 kg−1; Hiatt and Hiatt 1942) in harbor seals are ca. 40 % lower than predicted for terrestrial mammals of similar mass (Edwards 1975), possibly due to their apneustic breathing pattern (i.e., breath-holding between expiration and inspiration similar to short dives). During prolonged (ca. 70 days) fasting in northern elephant seal pups, GFR decreases ca. 50 % depending on the duration of the fast to conserve water (Pernia et al. 1989; Adams and Costa 1993). In contrast, it increases 2-fold for short duration (7–18 days) fasting in lactating female elephant seals (Crocker et al. 1998). For post-absorptive bottlenose dolphins resting in a stretcher, RPF (5.1 ml min−1 kg−1) and GFR (2.0 ml min−1 kg−1) are ca. 70 % and 1.4-fold higher, respectively, than predicted for terrestrial mammals of similar mass (Malvin and Rayner 1968; Ridgway 1972; Edwards 1975). For both harbor seals and bottlenose dolphins, GFR increases (up to 5-fold) after feeding resulting in an increase in urine formation (up to 10-fold) depending on meal size. Overall, RPF and GFR are within the range of normal mammalian values but vary depending on feeding and fasting duration. With a few exceptions, marine mammals do not concentrate Na+ and Cl− to levels above that of seawater, suggesting that they do not regularly consume seawater, although incidental intake may occur during feeding (Ortiz 2001).

In contrast, when harbor seals are forcibly submerged, RPF, GFR, and urine production are greatly reduced or cease, and normal hepatic and gastrointestinal functions are probably interrupted (Irving et al. 1942; Kvietys and Granger 1982). Based on these studies, it appeared that renal and splanchnic organ function stopped during diving and resumed at the surface between dives. However, when it became apparent that blood flow to peripheral tissues and organs during most natural dives is sufficient to maintain aerobic metabolism, then the question arose: Are renal and splanchnic functions maintained? These organs are of particular interest because marine mammals feed while diving and must process water and nutrients during foraging sessions lasting many hours and consisting of many dives with only a few minutes between dives.

Unfortunately, there has been only one study of renal and hepatic function in seals making voluntary dives within their ADL (Davis et al. 1983) and no studies for cetaceans or sirenians. This is primarily due to technical challenges of making measurements of organ function in unrestrained animals. Once again, Weddell seals were the preferred species because the isolated ice-hole protocol enabled the infusion of tracers (inulin and indocyanine green) and blood sampling between voluntary dives, some of which appeared to involve feeding. Inulin is cleared exclusively by the kidneys and is used to measure GFR. Indocyanine green (ICG) is cleared by the liver and is dependent on blood flow. These tracers were injected through a catheter while subadult seals (mean mass 168 kg) were resting at the surface (Fig. 14). Blood samples were then taken after each dive or at 5- to 10-min intervals at the surface. The mean half time for inulin clearance (27.3 ± 13.0 min) and ICG clearance (18.3 ± 7.3 min) under resting conditions at the surface did not change significantly during dives that were within the ADL, although there was variability. The mean GFR was 3.6 ml min−1 kg−1 which is ca. 3-fold greater than measured previously in post-absorptive harbor seals and 2.8-fold greater than terrestrial mammals of the same mass (Hiatt and Hiatt 1942; Schmidt-Nielsen et al. 1959; Edwards 1975), but this may have resulted from the seals having recently fed. The results indicate that resting GFR and hepatic blood flow oscillate around a mean level as heart rate increases during ventilation and decreases 40–50 % during apneustic breath holding at the surface. During voluntary dives that do not exceed the ADL, heart rate decreases 40–50 % similar to apneustic breath holds. Hence, mean heart rate during dives that do not exceed the ADL is comparable to that during apneustic breathing at the surface, and renal GFR or hepatic blood flow are similar to resting levels under postabsorptive and postprandial conditions. During dives that exceed the ADL, renal and hepatic function decrease dramatically as would be expected with a profound dive response that more closely resembles forced submergence (Guppy et al. 1986). Nevertheless, it appears that renal function can be reestablished quickly after severe and prolonged ischemia in seal kidneys demonstrating a tolerance to hypoxia not present in terrestrial mammals such as dogs (Halasz et al. 1974).

Fig. 14
figure 14

Isolated ice-hole protocol used to study renal and hepatic function in Weddell seals making voluntary dives beneath the short-fast ice in Antarctica. a Weddell seal with a catheter in the intervertebral extradural vein returning to the ice hole from a dive, b Antarctic silverfish (Pleuragramma antarcticum) brought to the surface by a foraging seal, c taking a blood sample to measure inulin and indocyanine green concentrations between dives, d increasingly lipemic plasma from a seal during the course of a 5 h foraging bout indicating digestion and assimilation of prey, most likely silverfish

Full size image

The functions of the kidneys and liver are blood flow dependent, and they have a large reserve capacity in performing their functions; that is, they filter or process more blood than is necessary to serve their homeostatic functions in healthy, young animals. As a result, blood flow to these organs can be reduced by 80 % for extended periods without deleterious effects (Nishizaki et al. 2001; Levey and Coresh 2012). Assuming proportionate reductions in heart rate, cardiac output and regional blood flow (Johansen and Aakhus 1963; Elsner et al. 1964; Jobsis et al. 2001), a 75 % reduction from resting levels would result in a heart rate similar to the minimum (ca. 11 min−1; Fig. 4) observed by Davis and Williams (2012) for Weddell seals during aerobic dives. In their model, Davis and Kanatous (1999) showed that blood flow to the kidneys and splanchnic organs could be reduced by 80 % relative to resting levels, and convective oxygen transport would still be adequate to support the metabolism of these organs during aerobic dives. Hence, it appears that the metabolic and functional performance of the kidneys and liver can tolerate significant reductions in blood flow and maintain homeostasis, but the minimum flow is about 20 % of resting levels at the surface. This wide range in blood flow for normal renal and hepatic function enables seals to preferentially adjust cardiac output and convective oxygen transport to meet the metabolic needs of active muscles while maintaining the aerobic function of other organs.

While measuring inulin and ICG turnover, Davis et al. (1983) observed that the plasma became lipemic within an hour after the seals began making deep dives associated with foraging [probably on Antarctic silverfish (Pleuragramma antarcticum), Davis et al. 2013] (Fig. 14). The appearance of chylomicrons in the plasma indicated that the gastrointestinal tract and pancreas were functional during the 5–6 h foraging bout involving multiple dives. There was no indication that digestion was postponed while the seals were actively feeding. In addition, plasma glucose concentration was normal and stable throughout the foraging bout (Castellini et al. 1988). Taken together, these results indicate that: (1) renal and hepatic function continue during most aerobic dives, (2) physiological homeostasis is maintained and (3) digestion and assimilation occur even while the seal is feeding at depth.

The spleen of seals (e.g., hooded, harp, northern elephant and Weddell seals) is anatomically similar to that in other mammals but larger (2 ± 4 % of body mass) and capable of temporarily sequestering red blood cells (splenic hematocrit >90 %) that are expelled (~80 % of splenic volume) into the hepatic sinus by contraction of the splenic capsular smooth muscle during α-adrenoceptor activation associated with the dive response (Cabanac et al. 1997, 1999 and references therein). Terrestrial mammals show a similar splenic contraction during exercise. During forced submergence in northern elephant seals, contraction of the spleen increases the volume of the hepatic sinus, and the caval sphincter in the diaphragm reduces flow to the right ventricle to prevent overloading at a time when diving bradycardia is most profound (Thornton et al. 2001). During voluntary dives with submerged swimming, the dive response is less profound (Fig. 4) which probably results in a dilated caval sphincter that allows blood from the spleen and hepatic sinus to enter the central circulation more rapidly. Assuming full oxygenation, these additional red blood cells could increase the ADL (ca. 18 min) of Weddell seals by 25 % (4–5 min depending on diving metabolic rate) and increase convective oxygen transport to tissues and organs (Davis and Kanatous 1999). Cabanac et al. (1997) estimated a more modest increase of 9 % or less (1.8 and 1.3 min, respectively) in the ADL for hooded seals (17 min) and harp seals (10 min) (see also Folkow and Blix 1999; Folkow et al. 2004). However, these calculations depend on the difference in hematocrit between resting and diving and on the estimated ADL that varies with the level of exertion. When diving stops, the spleen dilates, and the hematocrit returns to predive, resting levels in ca. 20 min (Hurford et al. 1996; Thornton et al. 2001) thereby lowering blood viscosity. However, re-sequestration is often longer than the inter-dive interval of 1–3 min required to re-oxygenate blood and muscle, so hematocrit remains elevated throughout dive bouts that may last many hours (Castellini et al. 1988). The spleen in small cetaceans is relatively small and probably does not serve a blood storage role (Cowan and Smith 1999).

Oxidative stress

Repeated cycles of diving ischemia and surface reperfusion of tissues can potentially increase oxidant production and oxidative stress (Elsner et al. 1998). However, seal tissues do not possess higher levels of oxidative damage than terrestrial mammal tissues, and available data suggest that diving mammals have the ability to cope with increased oxidant production without experiencing tissue damage (Zenteno-Savín et al. 2002; Vazquez-Medina et al. 2012). Although the control and regulation of the adaptive responses to oxidative stress in seals is not completely understood, they possess higher concentrations and activities of non-enzymatic (e.g., glutathione) and enzymatic (e.g., glutathione S-transferase) antioxidants, respectively, than terrestrial birds and mammals.

Fuel homeostasis

Irving et al. (1935) were the first to measure the respiratory quotient (RQ; ratio of carbon dioxide produced by tissue metabolism to oxygen consumed) of harbor seals under resting, normoxic conditions. Although they did not take precautions concerning fasting, the mean RQ was ca. 0.70 indicating a reliance on fat for energy metabolism. Similarly, a low (ca. 0.74) postabsorptive RQ was found for seals by other investigators (Scholander 1940; Hart and Irving 1959; Kooyman 1973; Davis et al. 1991a) indicating that 87 % of the seal’s energy is derived from fat catabolism and 13 % from carbohydrate (see Davis et al. 1993 for calculations). During steady state swimming, the postabsorptive RQ decreases to 0.72 (Davis et al. 1991a), and the seal derives 95 % of its energy from fat and 5 % from carbohydrate. This is not surprising because fat is the primary fuel for aerobic metabolism in many carnivores and laboratory animals conditioned on a low-carbohydrate diet (Roberts et al. 1943; Blazquez et al. 1971; Kettelhut et al. 1980). Studies have shown that carbohydrate-deficient diets produce metabolic changes that conserve carbohydrate and minimize its oxidation (Randle et al. 1963; Suzuki and Fuwa 1970; Eisenstein et al. 1974). Furthermore, the primary energy reserve for pinnipeds and cetaceans is subcutaneous blubber, which enables many species to fast for prolonged periods as part of their natural history. Hence, the transition from fed to fasting does not alter the intermediary metabolism of fat and carbohydrate (Kirby and Ortiz 1994; Ortiz et al. 2003; Champagne et al. 2005, 2006, 2012; Houser et al. 2007), although mild hyperglycemia, insulin resistance and impaired pancreatic sensitivity to glucose develop in elephant seal pups toward the end (6–8 weeks) of their prolonged, post-weaning fast (Viscarra et al. 2011).

Unfortunately, there have been few quantitative studies of the intermediary metabolism of marine mammals during forced submersion and diving. Davis (1983) quantified the turnover and oxidation rates of lactate and glucose in post-absorptive harbor seals under resting conditions and after forced submergence using the continuous infusion of radioactively labeled metabolites. Under resting conditions, the plasma concentrations and turnover rates for lactate (0.44 mM and 27 µmol min−1 kg−1, respectively) and glucose (5.7 mM and 14 µmol min−1 kg−1, respectively) are similar to other mammals, although Guppy et al. (1986) observed lower turnover rates for resting Weddell seals using a single bolus injection technique. However, the percent oxidation for lactate (21 %) and glucose (3 %) are much lower than other mammals (based primarily humans and omnivorous laboratory animals) indicating that carbohydrate carbon is conserved through recycling and little is used in oxidative metabolism. Keith and Ortiz (1989) observed a similarly low percentage of glucose oxidation in fasting elephant seal pups. Instead, most of the glucose that is metabolized to lactate is resynthesized to glucose through pyruvate recycling (Davis 1983; Champagne et al. 2012). Even after forced submersion in which the post-dive lactate concentration increases 10-fold, only 27 % is oxidized and the remainder recycled back to glucose. There is very little change in post-dive glucose oxidation (ca. 3–5 %), although glucose concentration increases initially by 1.5-fold.

Although fat has been known to support most of a seal’s normoxic, postabsorptive metabolism for the past 70 years based on RQ, the possibility that fat continued to be the major metabolic fuel during diving was not seriously considered until the late 1970s (Kooyman et al. 1980). The earlier belief that carbohydrate was the major, if not only, source of energy during diving resulted from: (1) a significant anaerobic contribution to energy metabolism during forced submergence that was interpreted as representative of the free-ranging condition and (2) the results from early enzyme studies suggesting that the anaerobic capacity of marine mammal tissues was enhanced as an adaptation to diving (Scholander 1940; George et al. 1971; Hochachka 1981). Histochemical studies that showed that seal muscle has a high lipase activity (tributyrinase), even higher than fat-metabolizing avian pectoralis muscle, did not change this belief (George et al. 1971). Adaptive regulation of biochemical pathways was hypothesized that would enable energy metabolism in peripheral tissues to oscillate between anaerobic carbohydrate metabolism during diving and aerobic fat catabolism at the surface (Hochachka 1976). However, these hypotheses were abandoned by the mid-1980s as it became apparent that most free-ranging dives are supported by aerobic metabolism. The allosteric and hormonal consequences of a fat based metabolism prevent tissues from switching to carbohydrate even under conditions of hyperglycemia and hyperinsulinemia, although pinnipeds generally have low plasma insulin levels, a low insulin:glucagon ratio, and a glucose tolerance profile similar to that of obese and diabetic humans (Kirby and Ortiz 1994; Champagne et al. 2005, 2006). During forced submergence and long voluntary dives in which the dive response is maximum, the switch to anaerobic metabolism, especially in peripheral tissues, may occur soon after submergence. Under these circumstances, glycogen stores may be rapidly used as the sole fuel for anaerobic glycolysis, although the slow, post-dive uptake of glucose and replenishment of these glycogen stores may preclude consecutive dives that rely on anaerobic metabolism. However, studies of free-ranging seals indicate that most dives are within the ADL and that tissues are sufficiently perfused to remain aerobic (Kooyman et al. 1980; Davis and Williams 2012). Under these conditions, fat catabolism continues to be the primary source of energy.

This conclusion is supported by the results from Davis et al. (1991a) who studied the intermediary metabolism of glucose, lactate, and free-fatty acids (palmitic acid) in harbor seals during voluntary submerged swimming in a water flume at 35 and 50 % of maximum aerobic capacity (\( \dot{V}{\text{O}}_{2\hbox{max} } \)). The plasma lactate concentration (0.93 mM) and percent oxidation (26 %) showed little change from resting, pre-exercise levels, although turnover rate increased 2.1-fold (from 26.2 to 55.0 µmol min−1 kg−1) at 50 % \( \dot{V}{\text{O}}_{2\hbox{max} } \). In total, anaerobic metabolism and the oxidation of lactate contributed less than 10 % to ATP production during submaximal submerged swimming up to 50 % \( \dot{V}{\text{O}}_{2\hbox{max} } \) (Table 2). The plasma glucose concentration (8.0 mM), turnover (22.5 µmol min−1 kg−1) and percent oxidation (6 %) showed little change from resting, pre-exercise levels and contributed less than 2 % to ATP production during submaximal submerged swimming (Table 2). In contrast, there was a 25–37 % increase in the plasma free fatty acid (FFA) concentration (0.73–0.91 mM) and turnover (7.5–10.3 µmol min−1 kg−1), and the percentage oxidized increased from 33 % (67 % is re-esterified into triglycerides) under resting conditions to 85 % at 50 % \( \dot{V}{\text{O}}_{2\hbox{max} } \). As a result, 20–22 % of ATP production during submerged swimming at low and moderate levels of exertion comes from plasma FFA (Table 2). Comparable turnover rates for glucose and FFA have been observed in exercising grey seals (Castellini et al. 1985), fasting elephant seal pups (Castellini et al. 1987; Champagne et al. 2012), and during fasting with lactation in adult northern elephant seals (Champagne et al. 2006), although none of these studies quantified oxidation rates.

Table 2 Rates of ATP production from aerobic and anaerobic metabolism and from the oxidation of glucose, lactate, FFA and VLDL-TGFA
Full size table

The energetic contribution of plasma FFA oxidation alone (20–22 %) is less than estimates of the proportion (87–95 % depending on activity level) of total energy from fat catabolism based on the RQ. The oxidation of plasma FFA fails to account for 69–73 % of the energy derived from fat catabolism. Additional sources of fatty acids may come from very low-density lipoprotein triglyceride fatty acids (VLDL-TGFA) or chlyomicrons (Horowitz and Klein 2000; Stellingwerff et al. 2007). Total plasma triglycerides for marine mammals are similar to values for other mammals including humans (i.e., 45–155 mg 100 ml−1) (Davis et al. 1991b; Tift et al. 2011). Davis et al. (1993) estimated the turnover rate of VLDL-TGFA using isotopically labeled glycerol as a precursor for labeled triglycerides (TG). The concentration of plasma TGFA (0.54 mM) and VLDL-TGFA (0.36 mM), the percentage of plasma TG in VLDL (63 %), and the postabsorptive VLDL-TGFA turnover rate under resting conditions (4.8 µmol min−1 kg−1) were within the normal mammalian range. Assuming the same percentage of VLDL-TGFA is oxidized as the plasma FFA, then the VLDL-TGFA oxidation rate for harbor seals would be 1.6 µmol min−1 kg−1 at rest and 4 µmol min−1 kg−1 at 50 % \( \dot{V}{\text{O}}_{2\hbox{max} } \). This means that the percentage contribution of VLDL-TGFA oxidation to total ATP production would be about 13 % (Table 2). However, additional research is needed to measure VLDL-TGFA oxidation at rest and during submerged swimming and diving. In addition, most studies of fuel homeostasis have been conducted under postabsorptive conditions to avoid non-steady state changes in blood metabolites associated with digestion and absorption. However, marine mammals must dive to forage, and they digest food during foraging bouts that may last many hours and include multiple dives. After the first hour of foraging, Davis et al. (1983) observed that the plasma becomes very lipemic due to an increase in chylomicrons in Weddell seals (Fig. 14d). During this time, chylomicron-TGFA may be the primary source of lipid for energy metabolism. If we assume that the proportion of chylomicron-TGFA catabolized is proportional to the plasma concentration, then a fourfold increase during digestion and absorption could account for over 50 % of ATP production.

Overall, current estimates of the oxidation of glucose, lactate, FFA and VLDL-TGFA can account for about 40 % of total ATP production of which three-quarters comes from the oxidation of plasma FFA and VLDL-TGFA. Therefore, endogenous lipids such as intramuscular triacylglycerols (i.e., lipid droplets, see “Tissue level adaptations for maintaining aerobic metabolism in skeletal muscle”) must be an important source of lipid for energy metabolism and, based on lipid droplet size and density (Fig. 13), could provide enough energy to support several hours or more of muscle metabolism at routine levels of exertion (Fig. 5) during aerobic dives (Davis and Kanatous 1999; Kanatous et al. 1999). However, this contribution to energy metabolism has not been quantified in captive or free-ranging marine mammals.

There have been a few studies of intermediary metabolism of amino acids in marine mammals. Most research has focused on fasting northern elephant seals that appear to conserve protein while relying primarily on fat metabolism for energy. During prolonged fasting, the oxidation of protein provides less than 4 % of the energy metabolism of elephant seal pups (Pernia et al. 1980; Adams and Costa 1993; Houser and Costa 2001), but the percentage is higher (8–11 %) in fasting, lactating female elephant seals (Crocker et al. 1998) and fasting harp seals (ca. 20 %; Worthy and Lavigne 1987). Even when fasting elephant seal pups are placed in captivity and fed a high fat and protein diet, they use the dietary protein for the formation of lean tissue (i.e., growth) and continue to rely on fat for energy metabolism (Condit and Ortiz 1987). This suggests a metabolic poise that conserves both protein and carbohydrate while relying primarily on fat for energy metabolism under a variety of conditions, but especially during fasting in adults and growing pups.

There have been no quantitative measurements of the intermediary metabolism of dietary amino acids in postprandial, carnivorous marine mammals during normal foraging. Depending on the proximate composition of the prey (e.g., fat vs. lean herring), the amount of energy from fat can vary from 50 to 75 % while that from protein can vary from 25 to 50 %. Hence, the catabolism of amino acids during digestion and absorption can contribute significantly to the energy metabolism of an adult animal on a diet of low fat (ca. 10 % fresh weight) prey. Some of the amino acids will be used for body maintenance and growth, but some may enter the Krebs cycle to produce ATP or the gluconeogenic pathway to produce glucose. The balance between these pathways will probably depend on species, percentage body fat, season, reproductive status, and age. This aspect of fuel homeostasis remains a fertile field for additional research.

To summarize this section, carnivorous marine mammals rely primarily on fat for ATP synthesis under resting, normoxic conditions and during dives within their ADL. This metabolic poise, which is little affected by fasting, reflects the proximate composition of their diet and the substantial fat reserves in their subcutaneous blubber. The absence of significant carbohydrate in their diet results in the conservation of glucose, most likely for obligate glucose metabolizing tissues such as the central nervous system and red blood cells. In addition, it results in a blunted insulin response that is similar to that observed in obese and diabetic humans and omnivorous laboratory animals, but without the associated pathologies. The energetic importance of amino acids remains poorly understood, but they could contribute significantly to postprandial energy metabolism depending on the proximate composition of the diet. Nevertheless, carnivorous marine mammals represent extreme examples of adaptation to a high fat (dietary marine oil) and low-carbohydrate diet with a heavy reliance on aerobic lipid metabolism under normoxic and the hypoxic (but not anoxic) conditions associated with breath-hold diving.

Morphological and behavioral adaptations to reduce the energetic cost of diving

Perhaps the most obvious adaptation in marine mammals for an aquatic lifestyle is their streamlined body shape and limbs that are modified into flippers (pinnipeds) or flukes at the end of the spine (cetaceans and sirenians). The convergent evolution of a fusiform body shape has occurred in many aquatic vertebrates from fish to mammals, both extant and extinct (e.g., ichthyosaurs), in response to the greater density (820-fold) and viscosity (55-fold) of water compared to air (Vogel 1994). These two properties of water greatly enhance body drag [the resistive force to movement that increases exponentially (~power of 2) with speed] and the cost of locomotion, especially in non-hydrodynamic shapes (a bluff body such as a sphere). As a result, there is a significant energetic advantage to body streamlining. In general, the optimum streamlined body has a fineness ratio (ratio of body length/maximum body diameter) of approximately 4.5 (Fish 1996). Body streamlining reduces pressure drag by delaying flow separation and turbulence. Body shapes that have a ratio smaller than 4.5 experience greater pressure drag while those body shapes with a larger ratio experience greater friction drag (Vogel 1994). The pressure drag on a streamlined body is small (~20 %) relative to the friction drag. Marine mammals mimic the ideal hydrodynamic shape with their elongate bodies and internalized (or absent) limbs which reduce turbulent flow and minimizes the energetic cost of locomotion. Body contouring into a fusiform shape is enhanced by subcutaneous blubber which also provides thermal insulation and serves as an energy reserve.

To overcome drag during aquatic locomotion, marine mammals have evolved flippers and flukes with a high (3–9) aspect ratio (span2/area) to improve thrust production and propulsive efficiency (Fish 2000). With their streamlined cross-sections, flippers, dorsal fins, and flukes also function as control surfaces that provide maneuverability and stability (Weber et al. 2013). Since all marine mammals evolved from quadrupedal land animals, morphological changes for aquatic locomotion can be traced in the fossil record and include enlargement of the hands and feet with webbing, elongate dorsoventrally flattened tail, foreshortening or loss of limbs, expansion of the tail at the tip (i.e., flukes) and high aspect propulsive surfaces (Fish 1996). These morphological changes were associated with changes in locomotory mode from quadrupedal paddling to pelvic oscillation (cetaceans and sirenians), caudal oscillation (seals) and pectoral oscillation with performance changes from drag-based propulsion (paddling) to lift- based that increased efficiency, power, and speed (Fish 1996, 2000). Together, body streamlining and efficient, lift-based propulsive appendages reduce the cost of swimming and optimize energy use in aquatically adapted mammals (Fish 1996, 2000; Williams 1999).

The morphological adaptations that optimize aquatic locomotion are accompanied by behavioral changes that further reduce energetic costs. Surface swimming augments drag ca. 5-fold due to the generation of a wave. The additional wave drag can be eliminated by swimming three body diameters below the surface. Hence, most marine mammals minimize their time at the surface and travel submerged. In addition, many marine mammals take advantage of buoyancy changes during diving to reduce energetic cost. During descent, compression of the lungs results in the transition from positive buoyance at the surface to negative buoyancy which enables the animal to glide to depth without stroking thereby reducing the energetic cost depending on the type of dive (Skrovan et al. 1999; Williams et al. 2000; Davis et al. 2001; Davis and Weihs 2007). Furthermore, they generate hydrodynamic lift which enables them to use negative buoyancy to move horizontally as well as vertically without stroking during descent. To minimize the depth of transition to negative buoyancy (and to minimize exposure to high partial pressures of nitrogen in the lungs), deep diving marine mammals exhale prior to diving to reduce initial lung volume. In contrast, sea otters inhale prior to diving and, along with air trapped in their fur, remain positively buoyant throughout dives to routine depths (ca. 30 m or less) which increases the energetic cost. Sirenians, which are relatively shallow divers, have dense bones which enable them to remain neutrally buoyant at depth.

However, the most significant behavioral adaptation for marine mammals is to dive within (or close to) their ADL. By doing so, they minimize their surface time because it only takes a few minutes to recharge blood and muscle oxygen stores resulting from postdive hyperventilation and tachycardia. In contrast, dives that exceed the ADL and disrupt physiological homeostasis can take over an hour for full recovery. For example, in a hypothetical sequence of six, typical dives by a Weddell seal, each dive would last 15 min with 3 min for recovery between dives. The total time spent underwater would be 90 min (6 × 15 min) and the time at the surface would be 15 min (5 × 3 min) for a total duration of 105 min. As a result, the seal spends 86 % of its time submerged and only 14 % at the surface. In contrast, if the seal were to make an anaerobic dive of 45 min, it would take 60 min to fully recover (metabolize the post-dive blood lactate; Kooyman et al. 1980). The combined duration of the dive and recovery period also would be 105 min, but only 43 % of the time was spent underwater and 57 % at the surface. Hence, diving aerobically is twice as efficient in terms of submergence time, which increases opportunities capture prey. That does not mean that marine mammals always dive within their ADL, but the energetic return would need to justify a prolonged recovery at the surface for the animal to remain in positive energy balance.

Multi-level integration of adaptations for maximizing aerobic dive duration

The evolution of adaptations in marine mammals for maximizing aerobic dive duration has been driven by the enhanced ability to capture prey and to remain in positive energy balance which enables them to grow, maintain health, avoid predators, and reproduce. These adaptations are apparent in marine mammal species that feed in different aquatic habitats (geographically and vertically) and on different prey. Obviously, healthy animals in the wild have succeeded in obtaining sufficient food, but understanding how they do so is more difficult. One of the greatest challenges for understanding the diving and foraging behavior of free-ranging marine mammals is observing and tracking them for extended periods at depth. Early studies relied on a small number of spatial and temporal variables (e.g., dive depth and duration, rate of descent and ascent, time at depth, swimming speed, time-depth profile) obtained from a variety of animal-borne instruments to categorize dives based on behaviors that were inferred but not observed (see Schreer et al. 2001 for a review). A few investigators (Harcourt et al. 2008; Hindell et al. 2002; Wartzok et al. 1992) used acoustic tracking techniques to examine the three-dimensional movements of Weddell seals but did not distinguish types of dives or assign behavioral functions. Acoustic tracking and dive classification based on three-dimensional movements have been performed for free-ranging ringed seals (Phoca hispida) (Simpkins et al. 2001a, b, c). However, these studies were limited by the inability to associate diving characteristics with direct observations of animal behavior or prey capture. Ultimately, the only way to understand the significance of biological adaptation is to study the animal in its natural environment, but this is difficult with diving animals because in most cases we cannot observe them once they submerge. However, advances in the miniaturization of video and digital imaging technology now enable researchers to use animal-borne instruments that record the in situ behavior of diving vertebrates and their prey and, in some cases, data on three-dimensional movements, locomotor performance and the physical environment (e.g., temperature and light intensity) (Figs. 15, 16, 17). These instruments are providing new insights into the behavioral ecology of marine mammals and how physiology constrains behavior at sea (e.g., Bowen et al. 2002; Davis et al. 1999, 2001, 2003, 2013; Fuiman et al. 2007; Heithaus et al. 2001; Hooker et al. 2002; Madden et al. 2008; Marshall 1998; Mitani et al. 2004; Parrish et al. 2002; Ponganis et al. 2000; Sato et al. 2003; Watanabe et al. 2003).

Fig. 15
figure 15

a Artist’s rendition of the video and data recorder (VDR) on a Weddell seal capturing prey, b head mounted VDR on a Weddell seal at the surface, c a miniature VDR showing the lowlight sensitive camera with six near-infrared light emitting diodes (LEDs) as a light source and sensors for recording the animals’ three-dimensional movements, locomotory performance and environment. Figure a courtesy of W. Monteleone

Full size image
Fig. 16
figure 16

Images obtained from the video and data recorder of prey captured by Weddell seals: a, b Antarctic silverfish (Pleuragramma antarcticum), c icefish (Pagetopsis macropterus), d bald notothen (Pagothenia borchgrevinki), e Trematomous sp. and f a decapod crustacean (possibly Notocrangon antarcticus). From Davis et al. 2013

Full size image

For example, Davis et al. (2013) deployed animal-borne video and data recorders (VDRs) on free-ranging Weddell seals diving beneath the shore-fast ice in McMurdo Sound, Antarctica during the austral spring. Mid-water foraging dives had a mean maximum dive depth of 356 m, mean duration 16.8 min, and covered a total distance 1,470 m (Fig. 17). Video recorded prey capture and three-dimensional dive analysis is enabling us to understand the details of their foraging behavior and hunting strategies which, unlike terrestrial carnivores, have been hidden from us. We can now describe predatory behaviors such as searching, pursuit, attack, and handling. We now know that Weddell seals feed primarily on Antarctic silverfish in McMurdo Sound during the austral spring (Fig. 14a, b) and capture an average of 15.7 fish with a mean net energy gain of 4,123 kJ per dive (Appendix). The efficiency (percent energy consumed to energy expended) for these dives is 500 %. This remarkable ability to acquire food at depth is the culmination of millions of years of evolution resulting in the multi-level integration of adaptations for maximizing aerobic dive duration which, to varying degrees, occurs in all marine mammals. These adaptations are the long, evolutionary legacy of an aquatic lifestyle that directly affects the fitness of marine mammal species for different diving abilities and environments.

Fig. 17
figure 17

The three-dimensional profile of a typical mid-water foraging dive by a Weddell seal showing different foraging and non-foraging states and captures of Antarctic silverfish (black circles). Inset shows details of the search, pursuit, capture, and handling of a silverfish. The path is color coded for flipper stroke frequency. Figure courtesy of K. Madden

Full size image

Conclusions

Marine mammals exhibit a variety of adaptations that maximize aerobic dive duration that include:

  1. 1.

    A dive response during aerobic dives enables the efficient use of blood and muscle oxygen stores but is exercise modulated to maximize the ADL at different levels of exertion.

  2. 2.

    Blood volume and concentrations of blood hemoglobin and muscle myoglobin are elevated and serve as a significant oxygen store that increases aerobic dive duration.

  3. 3.

    Myoglobin is not homogeneously distributed in the locomotory muscles and is highest in areas that produce greater force and consume more oxygen during aerobic swimming.

  4. 4.

    Muscle fiber types are primarily fast and slow twitch oxidative with elevated mitochondrial volume densities and enhanced oxidative enzyme activities that are highest in areas that produce more force generation.

  5. 5.

    Most of the muscle mitochondria are interfibriller and homogeneously distributed. This reduces the diffusion distance between mitochondria and helps maintain aerobic metabolism under hypoxic conditions.

  6. 6.

    Mitochondrial volume densities and oxidative enzyme activities are also elevated in certain organs such as liver, kidneys, and stomach.

  7. 7.

    Hepatic and renal function along with digestion and assimilation continue during aerobic dives to maintain physiological homeostasis.

  8. 8.

    Most ATP production comes from aerobic fat metabolism. Glucose is conserved for tissues such as red blood cells and the central nervous system.

  9. 9.

    Marine mammals minimize the energetic cost of swimming and diving through body streamlining, efficient, lift-based propulsive appendages, and cost-efficient modes of locomotion that reduce drag and take advantage of changes in buoyancy with depth.

  10. 10.

    Most dives are within the animal’s ADL.

The result of these adaptations is increased breath-hold duration and enhanced foraging ability that maximizes energy intake and minimizes energy output while making aerobic dives to depth.