Autonomic Thermoregulation

Thad E. Wilson; Kristen Metzler-Wilson

doi:10.1093/acrefore/9780190264086.013.15

Summary

Thermoregulation is a key physiologic homeostatic process and is subdivided into autonomic, behavioral, and adaptive divisions. Autonomic thermoregulation is a neural process related to the sympathetic and parasympathetic nervous systems. Autonomic thermoregulation is controlled at the subcortical level to alter physiologic processes of heat production and loss to maintain internal temperature. Mammalian, including human, autonomic responses to acute heat or cold stresses are dependent on environmental conditions and species genotype and phenotype, but many similarities exist. Responses to an acute heat stress begin with the sensation of heat, leading to central processing of the information and sympathetic responses via end organs, which can include sweat glands, vasculature, and airway and cardiac tissues. Responses to an acute cold stress begin with the sensation of cold, which leads to central processing of the information and sympathetic responses via end organs, which can include skeletal and piloerector muscles, brown adipose tissue, vasculature, and cardiac tissue. These autonomic responses allow homeostasis of internal temperature to be maintained across a wide range of external temperatures for most mammals, including humans. At times, uncompensable thermal challenges occur that can be maintained for only limited periods of time before leading to pathophysiologic states of hyperthermia or hypothermia.

Introduction

Although J. N. Langley first substituted the term autonomic for visceral in 1898 (E. M. Tansey, 1999), the value of the term thermoregulation has been recognized for even longer as one of the most fundamental biologic concepts and examples of the homeostatic process in physiology. The opposite of thermoregulation is thermoconformation, which indicates internal temperature conforming to the environmental temperature. These definitions are often too simplistic, implying that if an organism does anything to modify heat production/loss or the local environment, it is not conforming (Bligh, 1979). Thermal biologists use several more descriptive terms for various body temperature patterns (Table 1). Nonetheless, thermoregulation is functionally subdivided into autonomic, behavioral, and adaptive divisions. Autonomic and behavioral divisions are neural processes—with autonomic being related to the sympathetic and parasympathetic nervous systems and behavioral being related to cortical memory and decisions, as well as aversions. Examples of cortical decisions based on perceptions of thermal comfort and discomfort include ambulation to alternate locations, fluid ingestion, shelter/clothing changes, and body positioning/orientation to maximize heat gain or loss. Autonomic thermoregulation is controlled at the subcortical level to alter the physiologic processes of heat production and loss to maintain a relatively stable internal temperature. Adaptive thermoregulation is often related to an endocrine or structural alteration, in which changes develop over time that affect or modulate heat production or loss. Examples of adaptive thermoregulatory changes to chronic cold stress in some mammalian species include increased thyroid hormones to increase metabolism and therefore heat production, and increased adipose tissue deposition or hair growth to increase skin-surface insulation to prevent heat loss. The longer-term phenotypic or even genetic changes associated with adaptation of a species are not the focus of this article; rather, the focus is specifically on acute autonomic changes attempting to keep behavioral and adaptive/acclimatizing factors consistent, when possible. This is especially difficult because individuals and strains within a species can have different thermoregulatory strategies (Gordon, 1993). Finally, the article addresses mammalian autonomic thermoregulation, with an emphasis on humans.

Table 1. Summary of Terms Related to Patterns of Body Temperature

Term	Definition
Bradymetabolic	The low level of basal metabolism of reptiles and other nonavian and nonmammalian animals relative to those of birds and mammals of the same body mass and at the same tissue temperature. Synonym: cold-blooded. Antonyms: tachymetabolic, warm-blooded.
Cold-blooded	The thermal state of an animal in which core temperature remains close to ambient temperature when subjected to a low ambient temperature. Synonym: bradymetabolic (preferred term), poikilothermic. Antonyms: tachymetabolic, warm-blooded.
Ectothermic	The pattern of temperature regulation of animals in which body temperature depends mainly on the behaviorally controlled exchange of heat with the environment. Autonomic thermoeffectors may be temporarily important in certain species. Antonym: endothermic.
Endothermic	The pattern of thermoregulation in which the body temperature depends on a high (tachymetabolic) and controlled rate of heat production. Behavioral responses are also often used by endotherms. Antonym: ectothermic.
Eurythermal	The tolerance by organisms of a wide range of environmental temperatures, or the accommodation to substantial changes in the thermal environment. Antonym: stenothermal.
Heterothermy	The pattern of temperature regulation in a tachymetabolic species in which the variation in core temperature exceeds the variation that defines homeothermy.
Homeothermy	The pattern of temperature regulation in a tachymetabolic species in which the cyclic variation in core temperature is maintained within arbitrarily defined limits despite much larger variations in ambient temperature (i.e., homeotherms regulate their body temperature within a narrow range).
Poikilothermy	Large variability of body temperature as a function of ambient temperature in organisms without effective autonomic temperature regulation. As a rule, bradymetabolism implies poikilothermy, with only temporary exceptions in certain species. Tachymetabolism excludes poikilothermy, except in pathological conditions (impairment of temperature regulation), but permits heterothermy or torpor to occur in a number of species. Antonym: homeothermy.
Stenothermal	Descriptive of organisms that occur naturally in a narrow range of environmental temperatures and that, singly or collectively, are intolerant of, or accommodate ineffectually to, wide changes in their thermal environment. Antonym: eurythermal.
Tachymetabolic	The high level of basal metabolism of birds and mammals relative to those of reptiles and other nonavian and nonmammalian animals of the same body mass and at the same tissue temperatures. Synonym: warm-blooded. Antonyms: bradymetabolic, cold-blooded.
Warm-blooded	The thermal state of an animal that maintains its core temperature considerably higher than that of the environment when subjected to a low ambient temperature. Synonym: tachymetabolic (preferred term). Antonyms: bradymetabolic, cold-blooded.

Note: Definitions adapted from International Union of Physiological Societies—Thermal Commission (2001).

Systemic Thermal Challenges and Regulation

Passive and active heat stress imparts energy to cells, tissues, and organ systems of the body and, if not dissipated, can result in damage and potentially organismal death. Hyperthermic temperatures (> 40°C and especially > 42°C) can impair DNA and protein synthesis, can unfold thermolabile proteins, and can be cytotoxic (Lepock, 2003; Roti, 2008). In addition, certain tissues and organs (e.g., the central nervous system) are especially vulnerable to functional alterations during hyperthermia (W. P. Cheshire, 2016; Davis, Wilson, White, & Frohman, 2010; J. R. S. Hales, Hubbard, & Gaffin, 1996). In response to heat stress, the autonomic nervous system coordinates a proportional multiple-efferent pathway response to variables affecting systemic heat loss and metabolism, which acts to prevent large gains in thermal energy (i.e., heat storage). The preventive effect primarily occurs by decreasing tissue insulation to facilitate dry heat exchange (i.e., convection, conduction, and radiation) and by engaging eccrine sweat glands and other means to facilitate wet heat exchange (i.e., evaporation). Heat dissipation from internal tissues requires heat transfer via blood, which is shunted to the skin surface by the cardiovascular system, so that heat is transferred to the external environment.

Passive or active cold stresses can be an equally dangerous temperature challenge, involving heat removal or reduced heat production from cells, tissues, and organ systems of the body. Hypothermic temperatures can decrease enzymatic reaction rates, thus depressing cellular function; can alter hemostasis toward clotting; and can cause cell lysis if water crystals form (Hamlet, 1988; Pozos, Iaizzo, Danzl, & Mills, 1996). In response to cold temperatures, the autonomic nervous system coordinates proportional multiple-efferent pathway responses to variables affecting systemic heat gain and loss, which act to prevent large losses of thermal energy (i.e., negative heat storage). Through the control and regulation of blood flow, the autonomic nervous system can adjust tissue insulation and thus can partially adjust heat loss through dry heat exchange. Although the physiologic response to cold stress primarily involves mitigating heat losses via increasing surface-tissue insulation, sometimes increasing heat production by shivering or nonshivering thermogenesis is an additional strategy. Disorders of the central and peripheral nervous system can severely alter the body’s ability to mount a response to a cold challenge and thus can lead to hypothermia (W. P. Cheshire, 2016).

When describing a thermal challenge, it is important to note that autonomic responses have finite capacities—i.e., some challenges are too extreme or the rate of change is too rapid. Thermal challenges exceeding thermoregulatory capacity are considered “uncompensable,” while those that are within capacity are considered “compensable.” Compensable thermal challenges, after an initial adjustment period, result in a thermal steady state, where heat loss and heat gain are homeostatically matched. Unbalanced thermal states lead to heat gain or heat loss, which is referred to as positive or negative body heat storage. Uncompensable thermal challenges occur in nature but can be maintained for only limited periods of time prior to the development of a heat illness or pathophysiology—such as hyperthermia or hypothermia. When determining the nature of the thermal challenge (compensable or uncompensable), it is imperative to measure an internal temperature index that is representative of the organism’s internal state and has a rapid response time; in humans, temperature measures that fit these parameters are blood and esophageal temperature, as opposed to rectal, tympanic, or forehead skin temperature (W. P. Cheshire, 2016; Taylor, Tipton, & Kenny, 2014).

Homeostatic regulation of body temperature has classically been described in control system language borrowed from engineering. In this approach, four primary components describe a regulated variable: the variable itself, sensors, integrator/controller, and effectors. A few additional concepts must be added: negative feedback, set-point, and error signal. This approach works very well in engineering; for example, in a heating/ventilation/air-conditioning system, where temperature is the regulated variable and thermocouples in a wall unit (thermostat) are sensors. The temperature measured by the wall unit is the sample, which is used to regulate the whole building. Thus, a single room’s temperature is not directly regulated, rather the only regulated temperature is that at the wall unit. This becomes problematic only when the effector (such as a blower) does not adequately or appropriately circulate air. Similar problems occur in vivo if blood is not adequately or appropriately circulated. The integration/control steps are often combined, but they can be divided into: collecting and integrating sensory inputs, comparing the inputs to a set-point, and controlling the efferent responses. The set-point is the desired temperature, and the error signal is the difference between the sensor input and the set-point. To extend the example, the effectors for heating in response to a cold challenge may be a heat pump and a furnace. The heat pump exchanges heat with the external environment; if the building temperature is cold enough, or the rate of change is too great for the heat pump, the furnace engages to generate heat for the building. This is similar to the body’s use of the following effectors: skin for heat exchange, and skeletal muscle and brown adipose tissue (BAT) for heat generation. The effectors stop their activity when sensors provide negative feedback to the integrator/controller and it reaches a temperature that does not produce an error signal. The system may act either in an “on/off” manner or proportionally to the challenge.

A fundamental question is whether mammals, including humans, actually utilize a classically defined control system. It is possible to use many of the control-system terms to describe autonomic thermoregulation: the regulated variable is temperature, the sensors are the preoptic area of the hypothalamus and skin and visceral thermoreceptors, the integrator/controller is the preoptic area and other areas of the hypothalamus, and the effectors for cold stress are blood vessels, arrector pili muscles, BAT, and skeletal muscle. The principle of negative feedback holds up well in vivo, but the set-point has eluded precise identification. A set-point helps to describe modifications in the system, such as a fever, which allows the body to switch to a regulated, but elevated, temperature. Determination of where such a set-point is precisely located, how it is consulted, and even more fundamentally, how the precise set-point value is established has been elusive. What appears to fit the data best is a notion of a “balance point” (Romanovsky, 2007). In this model, multiple pathways act independently (McAllen, Tanaka, Ootsuka, & McKinley, 2010); integration (or balancing) of the independent pathways, each of which might have slightly different thresholds of activation and inactivation, allows for a relatively well-maintained internal temperature (Fealey, 2013; E. A. Tansey & Johnson, 2015). The balance-point framework aids in explaining the physiology but still allows for negative feedback and for factors like fever, biological rhythms, and acclimatization, which are known to affect the control and regulation of internal temperature.

Acute Systemic Responses to Heat Stress

The precise autonomic response to heat stress is highly dependent on the type of stress (passive or active), duration, and whether the stress is compensable or not. Active heat stress typically involves heat generation via an increase in metabolism mediated by ambulation, work, and/or exercise—although there are some pathophysiologic mechanisms of active heat stress, such as thyroid storm or malignant hyperthermia. Voluntary skeletal muscle contractions associated with work or exercise also alter the cardiovascular, respiratory, and endocrine systems independent of heat stress. Therefore, this article focuses on passive heat stress. Passive heat stress studies in mammalian species can be performed in both field and laboratory conditions, each having methodological advantages and disadvantages. Field studies are more generalizable to the natural environment but often suffer from the inability to control variables that may affect mechanisms of action, while the advantages and disadvantages are flipped for laboratory-based studies. Mammalian laboratory studies often utilize an environmental chamber with telemetry; alternatively, animals must be restrained or anesthetized and placed under a heating blanket or radiant heat source. Restraint causes additional stress to the animal, and anesthesia may blunt certain autonomic reflexes (Stocker & Muntzel, 2013). Human participants can also be studied in environmental chambers, as well as by hot water immersion and whole-body heating with a water-perfusion suit. The latter provides a well-controlled and reproducible passive heat stress, without inducing other issues, such as the increases in hydrostatic pressure associated with water immersion (Wilson, Klabunde, & Monahan, 2014). In humans, passive heat stress does not substantially alter metabolic rate, while in some mammals, heat stress can create a state of hypoactivity (Folk, Riedesel, & Thrift, 1998). It is difficult to assess if the hypoactivity is an autonomic response or a means of behavioral thermoregulation.

Initial autonomic responses to a passive heat stress include increasing overall skin vascular conductance or conductance to a specific skin/epithelial region to offload heat via dry heat exchange (Cramer & Jay, 2016; Johnson, Brengelmann, Hales, Vanhoutte, & Wenger, 1986; Kellogg, 2006). Offloading heat via dry heat exchange is limited by ambient temperature; when ambient temperature is above skin temperature, then the thermal gradient shifts to onloading heat. A benefit of increasing cutaneous vascular conductance is that it can take advantage of evaporation via hypotonic sweat secretions, saliva spread on the skin surface, or serous gland secretions in epithelial airway surfaces (Hainsworth & Stricker, 1968; Robertshaw, 2006; Sato, Kang, Saga, & Sato, 1989). To increase skin perfusion, heart rate and cardiac output increase, and blood flow is redistributed away from the renal and splanchnic vascular beds (Crandall & Wilson, 2015; Wilson, 2016). To enact these autonomic thermoregulatory responses, a number of neural processes need to occur: thermosensation, afferent signaling, integration plus coordination, and efferent responses, including postganglionic sympathetic nerve activity. The neural control and regulation then activate or inhibit biological end organs to elicit a homeostatic response to the heat stress (Table 2).

Sensation/Central Processing

Open in new tab

Figure 1. Functional neuroanatomical model for the fundamental thermoregulation pathways. Consult the text for description. Abbreviations: dorsal root ganglia (DRG), dorsal horn (DH) of the spinal cord, lateral parabrachial nucleus (LPB), preoptic area (POA) of the hypothalamus, dorsomedial hypothalamus (DMH), rostral raphe pallidus (rRPa), median preoptic area (MnPO), warm-sensitive (W-S) neurons, medial preoptic area (MPA), brown adipose tissue (BAT), prostaglandin E₂ (PGE₂), cutaneous vasoconstriction (CVC), sympathetic preganglionic neurons (SPNs), intermediolateral nucleus (IML), and ventral horn (VH) of the spinal cord.

Used with permission from Morrison (2016).

Temperature information regarding warmth arises primarily from the preoptic area of the hypothalamus and from skin and visceral afferent warm-sensitive thermoreceptors (Benarroch, 2007; Fealey, 2013; Morrison, 2016; Nakamura, 2011), either via thermosensitive transient receptor potential vanilloid type-3 and 4 (TRPV3 and TRPV4) channels (Caterina, 2007; E. A. Tansey & Johnson, 2015) or other means. Cutaneous warm-sensitive afferent terminals are located in the base of the epidermis and throughout the dermal layer and are unmyelinated branches of thinly myelinated class C fibers (Ivanov, 1990; Pierau, 1996). There are other nerves in the skin that have thermal stimulation properties, but the C fibers appear to be the most important for thermoregulation. The warm-sensitive afferents can be ordered as follows (Figure 1): First-order neurons project to the dorsal horn of the spinal cord. Second-order neurons travel up the lateral funiculus of the spinal cord as part of the anterolateral system. These neurons, which carry information from warm-sensitive receptors, synapse at the dorsal subnucleus of the lateral parabrachial nucleus in the midbrain/pons junction of the brainstem. Third-order neurons in this pathway travel from the lateral parabrachial nucleus to synapse with warm-sensitive neurons (thermal integrators) in the preoptic area of the hypothalamus, via interneurons (Morrison, 2016; Nakamura, 2011).

The preoptic area acts as a response controller; it regulates temperature in that its excitatory and inhibitory outputs to neurons in the dorsomedial hypothalamus are balanced (Figure 1; Morrison, 2016). If hot temperatures and signals predominate, heat-defense pathways and responses (e.g., skin vasodilation and sweating) result (Nakamura & Morrison, 2010). Warm-sensitive neurons inhibit, and temperature-insensitive preoptic area neurons activate, the effector neurons for the heat-defense responses (Boulant, 1996). The pathways for heat responses are less clear, in part due to the fact that typical model systems, such as rodents or most other mammals, do not use eccrine sweating as a heat-loss mechanism. To address this issue, a series of brain imaging studies using functional magnetic resonance imaging during whole-body heating identified that, in humans, the following areas are activated: preoptic area of the hypothalamus, anterior and posterior cingulate cortex, insula, premotor cortex, thalamus, lentiform nuclei, cerebellum, and multiple brainstem nuclei (Farrell, Trevaks, & McAllen, 2014; Farrell, Trevaks, Taylor, & McAllen, 2013, 2015). Future studies will ascertain which of the areas are in the heat-defense pathway and which are concurrently activated.

Sympathetic Nerve Activity (SNA)

Heat stress can be considered hyperadrenergic (Rowell, 1990). This heightened sympathetic state is needed to coordinate the multiple primary and supportive functions of thermoregulation. In humans, passive heat stress with a water-perfused suit increases both sudomotor and related vasomotor components of skin SNA (Figure 2; Cui et al., 2006; Wilson, Cui, & Crandall, 2001, 2005) and vasoconstrictor components of muscle SNA (Cui, Wilson, & Crandall, 2002, 2004; Wilson & Ray, 2004). In other species, such as in passively heated anesthetized rats (Kregel, Kenney, Massett, Morgan, & Lewis, 1997), these responses can be observed as increases in lumbar SNA and as decreases in tail SNA (Blessing, McAllen, & McKinley, 2016). Increases in renal SNA also occur during heat stress in the rat (Kenney, Barney, Hirai, & Gisolfi, 1995; Kenney, Claassen, Bishop, & Fels, 1998; Kenney & Fels, 2002, 2003; Kenney, Musch, & Weiss, 2001; Kenney, Pickar, Weiss, Saindon, & Fels, 2000). Similar to lumbar SNA, renal SNA responses to radiant heat are abated by ganglionic blockade agents, such as trimethaphan (Kenney et al., 1995), indicating the accuracy of the postganglionic SNA measurements. Heat stress can also increase blood levels of catecholamines (Robertshaw, 1977), which provides an additional adrenergic stimulus to end organs.

Open in new tab

Figure 2. Total skin sympathetic nerve activity (SNA) across mean body temperature (0.9 × internal temperature + 0.1·mean skin temperature). Mean body temperature was manipulated via skin-surface cooling and whole-body heating. The open square is the mean body temperature in normothermic baseline conditions. Values are means ± SE (N = 7).

Used with permission from Cui et al. (2006).

Besides directly recording SNA from various nerves during heat stress, an alternate experimental approach to understand the increase in activity is to electrically simulate the nerve to mimic the sympathetic responses observed during heat stress. In humans, directly stimulating the common fibular nerve results in changes in electrodermal activity (an index of sweating) in the area of innervation (dorsal aspect of the foot; Ray & Wilson, 2004). Stimulation of the vidian nerve mimics heat-stress-induced nasal serous secretions in panting species, such as the dog (Wells & Widdicombe, 1986). Renal nerve stimulation decreases renal blood flow in a frequency-dependent manner (DiBona & Kopp, 1997); a decrease in renal blood flow is a common observation during heat stress in both rats and humans (Wilson, 2016). Splanchnic nerve stimulation decreases both vascular conductance and vascular volume in dogs (Karim & Hainsworth, 1976; Noble, Drinkhill, Myers, & Hainsworth, 1997); a decrease in splanchnic blood flow is a common observation during heat stress in rats and humans (Massett, Lewis, Bates, & Kregel, 1998; Rowell, Brengelmann, Blackmon, Twiss, & Kusumi, 1968). Hence, the data provide evidence that the increase in postganglionic SNA is indicative of the effect being proposed for the end organ during passive heat stress.

End Organs

End organs are the targets of the autonomic nervous system; during heat stress in humans, the most important end organs in facilitating heat loss are eccrine sweat glands and blood vessels of the skin and respiratory epithelia. Certain species may target or localize sweating/serous secretion and vasodilation rather than eliciting global responses, either for efficiency or for other reasons. A few of these specialized mammalian structure−function relationships, such as the carotid rete, will be highlighted. The blood vessels of other vascular beds and the heart are heat-exchange-supportive end organs that ensure that the skin has adequate interstitial fluid and blood flow.

Sweat Glands

Evaporation of sweat is one of the most effective cooling mechanisms available for thermoregulation. It has been estimated that one liter of sweat can liberate 580 kcal during evaporation (Wenger, 1972). The evaporation requirement is one of the primary determinants of whether a heat stress is compensable or not (Cramer & Jay, 2016). However, heat stress is not the only stress that can induce sweating; mental stress, arousal, or anxiety can induce focal sweating (Quinton, 1983; Wilke, Martin, Terstegen, & Biel, 2007). To engage sweating in response to thermal or nonthermal stimuli, postsynaptic SNA activates neurotransmitters and neuromodulators from nerve terminals. Human eccrine sweat glands express a number of postsynaptic adrenergic and cholinergic receptors that initiate and sustain sweating.

The human sweating response to heat stress is primarily cholinergic, as evidenced by observations that anticholinergics and botulinum toxin block sweating (Blackburn, Sammons, & Wilson, 2012; Cheshire & Fealey, 2008). Cholinergic sweating utilizes muscarinic type-3 receptors (Torres, Zollman, & Low, 1991); these G-protein coupled receptors mediate K⁺, Cl⁻, Na⁺, and Ca²⁺ influx across the basolateral membrane as well as Ca²⁺ release from the sarcoplasmic reticulum in the secretory coil (Metzler-Wilson et al., 2014; Sato & Sato, 1981). The initial ionic effects either cause cell shrinkage or activate second messengers that activate apical NCCK1 and basolateral Cl⁻ channels (Wilson & Metzler-Wilson, 2015). Sustained sweating appears to also utilize store-operated channels to replenish Ca²⁺ in the endoplasmic reticulum (Concepcion et al., 2016). The passage of Cl⁻ across the basolateral membrane establishes a transepithelial gradient that allows paracellular Na⁺ transport across tight junctions. The osmotic gradient that results from the Na⁺ and Cl⁻ transport then allows water movement through aquaporin-5 channels (Song, Sonawane, & Verkman, 2002).

Besides the cholinergic pathway, ~ 50% of cultured sweat gland cells are activated by β‎-adrenergic agonists and β‎-adrenergic pathway stimulants (Reddy & Bell, 1996). The effect of β‎-adrenergic agonists also translates to in vivo human preparations (Shamsuddin, Reddy, & Quinton, 2008). Circulating catecholamines have also been implicated in inducing sweating in horses and many species of cattle (McEwan Jenkinson, Elder, & Bovell, 2006; Robertshaw, 1975). However, endocrine-mediated sweating does not appear to play a primary role in in vivo human sweating. β‎-Adrenergic sweating involves stimulation of the cAMP pathway and the activation of Cl⁻ channels and paracellular Na⁺ transport in the secretory coil.

Precursor fluid, regardless if stimulated via cholinergic or adrenergic means, then travels up through the ductal region of the sweat gland to the skin surface. Sweat gland ducts consist of a two-cell structure containing ENaC and CFTR channels (Jones & Quinton, 1989). The two channels interact to reabsorb Na⁺ and Cl⁻ (Reddy, Light, & Quinton, 1999; Reddy & Quinton, 2003, 2005). β‎-Adrenergic receptors appear to regulate the ion reabsorption (Reddy & Stutts, 2013). Thus, by the time the fluid reaches the skin surface, it is hypotonic compared to plasma.

Airway Epithelia—Thermal Tachypnea

Heat-stress-induced alveolar hyperventilation decreases partial pressure of arterial carbon dioxide (PaCO₂) and end-tidal CO₂ (Brothers, Ganio, Hubing, Hastings, & Crandall, 2011; Cabanac & White, 1995; Fujii et al., 2008; Petersen & Vejby-Christensen, 1977; Tsuji, Honda, Fujii, Kondo, & Nishiyasu, 2012; White, 2006). The increase in ventilation increases air movement across the respiratory epithelia to facilitate evaporation. Panting species activate special serous nasal glands to increase fluid volume secreted onto the airway epithelial surface (Blatt, Taylor, & Habal, 1972). Panting can occur by a few different respiratory flow patterns: for example, in through the nose and out through the mouth in the dog (Schmidt-Nielsen, Bretz, & Taylor, 1970). Some respiratory flow patterns allow for high airflow but partially attenuate the increased alveolar ventilation and thus make respiratory alkalemia less of an issue. To take advantage of evaporation, specialized blood vessels (e.g., the carotid rete) within the nasal region of panting species increase conductive heat transfer to the nasal region for heat transfer. The carotid rete is under the control of the sympathetic nervous system (Maloney, Fuller, Mitchell, & Mitchell, 2002). In sheep, the total amount of blood flow has been observed in the nasal-buccal area; it approximately doubles during mild heat stress and then can be reduced during severe stress (J. R. S. Hales, 1973). There are also increases in tongue and nasal epithelial blood flow that occur for similar reasons as for the increased flow in the carotid rete.

Combined, these mechanisms have the potential to cause selective brain cooling, which keeps the brain at a temperature lower than the body. This is important because nervous tissue is especially sensitive to hot temperatures (J. R. S. Hales et al., 1996). Because of measurement difficulties, direct evidence for selective brain cooling in humans is minimal.

Heart

In humans, cardiac output increases in response to whole-body heating (Crandall & Wilson, 2015; Wilson & Crandall, 2011). In contrast, other mammals (e.g., rodents, sheep, and baboons) do not appreciably increase cardiac output during passive heat stress. The reason for this large comparative cardiovascular difference is that humans perfuse their entire skin rather than focally perfusing small thermoregulatory regions, such as the tail, ears, or carotid rete. This whole-skin perfusion could be seen as being less efficient, but it does have a greater overall capacity for heat exchange because of the larger area available for heat transfer.

To provide for the increase in cardiac output, heart rate increases significantly. The effect of temperature on heart rate has been widely documented across a wide range of temperatures (Crandall & Wilson, 2015). Using anticholinergics and β‎-adrenergic antagonism to functionally remove parasympathetic and sympathetic influences, it has been identified that there is an 8.4 ± 0.8 bpm increase per 1.0°C increase in internal temperature in the baboon (Gorman & Proppe, 1984). Of this increase, ~ 40% is due to direct increases in cardiac temperature and ~ 60% is due to stress-induced changes in autonomic nervous system activity, with ~ 25% due to sympathetic activation and ~ 75% due to parasympathetic withdrawal (Gorman & Proppe, 1984). The latter autonomic nervous system changes during heat stress facilitate focal and whole-body skin perfusion, and thus thermoregulation.

Open in new tab

Figure 3. Representative pressure waveforms during cool stress induced by skin-surface cooling. A, representative tracings of pulmonary artery pressure (PAP) and central venous pressure (CVP) during normothermia and skin-surface cooling. B, representative tracing of pulmonary capillary wedge pressure (PCWP) during the same thermal conditions. Variables in A are time-sequenced with each other but not to PCWP in B.

Used with permission from Wilson, Tollund, et al. (2007).

Open in new tab

Figure 4. Schematic of the effect of thermal stress on Frank-Starling relations. Panel A represents normothermia with the labeled point being the operating point. Panel B represents associated changes in the curves related to heat stress and cold stress as well as the location of the supine operating points on the curves. Panel C highlights slope changes in the curves where a similar decrease in pulmonary capillary wedge pressure would cause a relatively large change in stroke volume during heat stress and a relatively small change during cold stress.

Used with permission from Wilson and Crandall (2011).

Stroke volume does not appreciably change during whole-body heating in humans (Wilson & Crandall, 2011). This is somewhat surprising given that there are significant decreases in the pressure returning to the heart, especially when combined with orthostatic stress (Wilson, Tollund, et al., 2007). Thus, it can be surmised that other cardiac or vascular factors, such as changes in preload and/or afterload, are also important in heat stress in humans. Indexes of preload, such as central venous and pulmonary capillary wedge pressures (Figure 3) and left ventricular end-diastolic volume, decrease during passive heat stress in humans (Brothers, Keller, Wingo, Ganio, & Crandall, 2011; Crandall et al., 2008; M. D. Nelson et al., 2010, 2011; Stohr et al., 2011; Wilson et al., 2009). In addition, direct measures of central blood volume via gamma scans of technetium (Tc 99) radiolabeled erythrocytes identify clear decreases in central blood volume during passive heat stress (Crandall et al., 2008, 2012). The mechanisms for this decrease in cardiac preload during heat stress are multifactorial but are most likely related to a sweating-induced decrease of plasma and interstitial fluid volumes and/or translocation of blood volume to the cutaneous vasculature (Crandall et al., 2012; Deschamps & Magder, 1990). The change in vascular volume occurs despite a volume reduction from the splanchnic, renal, and central reservoirs via venoconstriction (Crandall et al., 2008; J. R. Hales, 1973; J. R. Hales, Rowell, & King, 1979). The vascular volume changes occur in most species but are especially pronounced in humans. Similar to indices of cardiac preload, indices of cardiac afterload (i.e., left ventricular end-systolic wall stress, systemic vascular resistance estimate, and arterial vessel stiffness) also decrease during heat stress (Crandall & Wilson, 2015). A decrease in cardiac afterload in humans is likely mediated by the large decrease in cutaneous vascular resistance and may allow for the aortic valve to open at an earlier left ventricular developed pressure. Thus, stroke volume is increased at any given filling pressure in humans. The final factor affecting stroke volume is cardiac contractility (inotropy). Indices of cardiac inotropy, such as increases in ejection fraction and left ventricular twist rates, isovolumic acceleration of the septal and lateral mitral annulus, and a leftward and upward shift in the Frank-Starling relation, increase during whole-body heating in humans (Brothers, Bhella, et al., 2009; Bundgaard-Nielsen, Wilson, Seifert, Secher, & Crandall, 2010; Crandall et al., 2008; M. D. Nelson et al., 2010, 2011; Stohr et al., 2011). This is likely due to cardiac sympathetic activity increases or adrenal medulla release of epinephrine, rather than a heat-related change in β‎-adrenergic receptor sensitivity or responsiveness (Klabunde, LePorte, & Wilson, 2013). The increase in cardiac contractility likely facilitates cardiac emptying during instances where filling pressure is challenged, such as during heat stress.

Combined, the effects of cardiac preload, cardiac afterload, and inotropy in stroke volume regulation can be observed in the alteration of Frank-Starling relations during systemic heat stress, where: decreased cardiac preload moves the operating point leftward down the same curve, and decreased afterload and increased inotropy cause a leftward and upward shift in the Frank-Starling relation to a new sister curve (Figure 4). Thus, stroke volume can be maintained because there is a greater stroke volume at any given filling pressure within the heat-stress curve.

Vasculature

As discussed, blood flow changes during heat stress are in part related to a mammal’s unique thermoregulatory strategies (e.g., panting, saliva spreading, and ear or tail vasodilation), which can have variable effects on blood pressure and volume changes. For example, with environmental heating to 37°C to 40°C, rats exhibit increased arterial blood pressure and no change in blood volume (Horowitz & Samueloff, 1988; Kregel, Wall, & Gisolfi, 1988), while humans maintain or decrease arterial blood pressure and decrease central blood volume (Crandall & Wilson, 2015). In humans, the increase in cardiac output is balanced by a decrease in systemic vascular resistance, which maintains arterial blood pressure in steady state until reaching severe heat stress, at which point arterial blood pressure is reduced. During heat stress, the decrease in systemic vascular resistance is primarily mediated by increases in cutaneous vascular conductance, but it is offset from decreasing too much by decreased vascular conductance of many other vascular beds (splenic, renal, muscular, and cerebral).

The skin is the primary source for the increase in vascular conductance, and diversion of blood flow to the skin−environment interface occurs for heat dissipation purposes. In humans, the total amount of blood flow that can be diverted to the skin has been calculated to be around 8 L/min (Johnson et al., 1986). The cutaneous vasodilation occurs as a twofold process in hairy, or nonglabrous, skin: vasoconstrictor tone is decreased and then removed, and active vasodilation follows. The decrease and removal of vasoconstrictor tone is due to the decrease of basal neuronal norepinephrine release into the synaptic cleft and subsequent binding to postsynaptic vascular smooth muscle α‎-adrenergic receptors. Active vasodilation is a more complex event that is still mechanistically unclear in that it requires a sympathetic cholinergic nerve but not its primary neurotransmitter, acetylcholine (Kellogg et al., 1995). It is currently unknown if the sympathetic cholinergic nerve that controls skin blood vessels is the same one that innervates the sweat gland (e.g., sudomotor and vasomotor), or if it is unique to the cutaneous vasculature (i.e., vasomotor only; Smith & Johnson, 2016). Two candidate co-transmitters for vasodilation during heat stress, each with some supporting data, are vasoactive intestinal peptide and pituitary adenylate cyclase activating peptide (Johnson, Minson, & Kellogg, 2014). Histamine released from mast cells and local sensory nerve activation can account for a portion of active vasodilation during heat stress (Wong & Hollowed, 2017). Also, nitric oxide appears to be permissive, meaning that it is required for full expression of vasodilation but does not induce full expression in active vasodilation. Furthermore, endothelial-derived hyperpolarizing factor and cyclooxygenase products appear to contribute to the overall vasodilatory response (Wong & Hollowed, 2017). It is likely that some of these mechanistic pathways are redundant or are unique to specific perturbations and circumstances. Glabrous, or nonhairy, skin does not contain an active vasodilator system in humans but only utilizes withdrawal of vasoconstriction (Johnson, Pergola, Liao, Kellogg, & Crandall, 1995). Rat tail skin also increases blood flow solely by the withdrawal of vasoconstriction (Blessing et al., 2016). Glabrous skin of humans and certain skin of other mammals, such as sheep, contain arteriovenous anastomoses. Because the arteriovenous shunt vessels have a larger luminal diameter than capillaries, higher flows can be maintained when vascular smooth muscle is relaxed. In addition to increases in cutaneous vascular conductance, there are also increases in cutaneous venous volume and flow (Deschamps & Magder, 1990, 1994). The increase in cutaneous venous volume likely allows for greater thermal exchange at the skin−environment interface.

The splanchnic vascular bed (comprised of the gastrointestinal tract, spleen, pancreas, and liver) and renal vascular bed receive approximately 40% to 50% of cardiac output in basal preprandial conditions. Splanchnic blood flow decreases during whole-body heating in humans (Rowell, Blackmon, Martin, Mazzarella, & Bruce, 1965; Rowell et al., 1968; Rowell, Detry, Profant, & Wyss, 1971). Splanchnic vascular resistance increases with increasing internal temperature, to between 20% and 60% from baseline. Whole-body heating also increases renal vascular resistance, which reduces renal blood flow 15% to 30% from baseline values (Wilson, 2016). Moreover, removal of the adrenal medulla attenuates the increases in renal and splanchnic vascular resistance, indicating combined neural-hormonal control and regulation of these vessels in the rat (Kregel & Gisolfi, 1989). The splanchnic and renal vasculatures provide a blood flow reserve that can be drawn upon during sympathetic events, such as heat stress.

Resting skeletal muscle receives approximately 15% to 20% of cardiac output, although this can increase dramatically during and after skeletal muscle contractions (Andersen & Saltin, 1985; Richardson et al., 1993; Saltin, Radegran, Koskolou, & Roach, 1998). Positron emission tomography suggests that direct, but not indirect, heating results in small increases in muscle blood flow during heat stress (Heinonen et al., 2011). The mechanism by which heating has the capability to slightly alter muscle blood flow may be related to a direct effect of heat on the muscle vasculature (Bagher & Segal, 2011; Cooke, Shepherd, & Vanhoutte, 1984) or heat-induced release of a local vasodilator (Harris, Blackstone, Ju, Venema, & Venema, 2003; Pearson et al., 2011). The vasodilatory changes compete with the increases in muscle SNA, causing the vasoconstriction described earlier (Cui et al., 2002, 2004, 2011; Keller, Cui, Davis, Low, & Crandall, 2006; Low, Keller, Wingo, Brothers, & Crandall, 2011; Niimi et al., 1997). Thus, the overall result is likely the small changes in blood flow.

Passive heat stress decreases cerebral perfusion in humans, as indexed via cerebral artery velocities using transcranial Doppler (Bain, Nybo, & Ainslie, 2015), although the precise magnitude of the reduction in cerebral perfusion to whole-body heating varies among subjects (Figure 3). The mechanisms behind the decreases are likely multifactorial but could be related to carbon dioxide, sympathetic innervation, and cerebral autoregulation. The cerebral vasculature is very sensitive to changes in PaCO₂, resulting in an ~ 2% to 4% reduction in cerebral perfusion for each 1 mm Hg reduction in PaCO₂ (Ringelstein, Van Eyck, & Mertens, 1992), a relationship that is unchanged by whole-body heating (Low et al., 2008). Thus, it is possible that thermal tachypnea causes a reduction in PaCO₂ and cerebral perfusion. Others suggested that increases in cerebral sympathetic stimulation during heat stress may also contribute (Brothers, Wingo, Hubing, & Crandall, 2009). Functionally, the cerebral vasculature is in part under sympathetic control (Ide et al., 2000; E. Nelson & Rennels, 1970; Ogoh, Brothers, Eubank, & Raven, 2008; Purkayastha, Saxena, Eubank, Hoxha, & Raven, 2013; Umeyama et al., 1995; Zhang et al., 2002), and thus it is reasonable to conclude that a component of the reduction in cerebral perfusion during heat stress is sympathetically mediated. Although the idea is not universally accepted (S. J. Lucas et al., 2010), cerebral blood flow is generally thought to be maintained over a range of perfusion pressures via intrinsic changes in the cerebral vasculature (Heistad & Kontos, 1983; Paulson, Strandgaard, & Edvinsson, 1990). It is possible that changes in the sympathetic activity or cerebral vascular responses could alter cerebral autoregulation; currently, there is some evidence for this hypothesis, but the evidence is equivocal in humans (Brothers, Wingo, Hubing, Del Coso, & Crandall, 2009; Low et al., 2009; Wilson, Cui, Zhang, & Crandall, 2006).

Table 2. Summary of Comparable Systemic and Local Responses to Acute Passive Heat Stress Before the Development of Hyperthermia in Three Commonly Studied Species of Mammals

Variable/Purpose		Human Response	Sheep (Ovis aries) Response	Rat (Rattus rattus) Response
Primary mechanism of heat loss	Evaporative heat loss	Eccrine sweating	Thermal tachypnea (panting)	Saliva spreading
			Apocrine sweating^*
	Dry heat loss	Increased skin blood flow	Increased nasobuccal and carotid rete blood flow	Increased tail blood flow
				Increased ear blood flow
Cardiovascular support of heat loss	Cardiac output	Large increase	No change	No change
	Systemic vascular resistance	Large decrease	Small decrease	Increase
	Arterial blood pressure	No change or small decrease	Small decrease	Increase

Note: ^*This mechanism’s importance is increased when sheep are shorn but is more minor with a full fleece.

Acute Systemic Responses to Cold Stress

The precise response to a cold stress is highly dependent on the type, duration, severity, and pain involvement during the exposure (Castellani, Sawka, DeGroot, & Young, 2010; Frank et al., 1997; Giesbrecht, 2000; Leblanc, 1975; Stocks, Taylor, Tipton, & Greenleaf, 2004; Toner & McArdle, 1996). Similar to studies of heat stress, mammalian studies of cold stress can be performed both in the field and in the laboratory. The studies often utilize an environmental chamber, with measurements being recorded on a data logger or transmitted via telemetry to a recording station; alternatively, animals must be restrained and/or anesthetized to be instrumented and placed on or near a cooling source. Human participants can also be studied in environmental chambers, as well as by cold water immersion and skin-surface cooling with a water-perfusion suit. The latter provides a well-controlled and reproducible passive cold stress, without inducing other issues, such as the increases in hydrostatic pressure associated with water immersion (Wilson et al., 2014). Additional advantages of this research method are that it is nonpainful, because cold pain enacts its own set of autonomic responses, and the suit can clamp skin temperatures just above or just below a shivering threshold.

Initial autonomic responses to a cold stress include decreasing cutaneous vascular conductance to increase tissue insulation and thus reduce dry heat exchange (Cramer & Jay, 2016; Johnson et al., 1986; Kellogg, 2006). Other methods to increase tissue insulation include piloerection and reduction of blood flow to nearby muscle (Herrington, 1951; Rennie, 1988). If cold stress is severe and/or prolonged, then increased shivering or nonshivering thermogenesis is required to maintain internal temperature. Nonshivering thermogenesis often involves specialized mitochondrial proteins or increased cellular pump activities. Many of these cold-induced responses require cardiovascular adjustments to maintain functionality during cold stress. To enact the autonomic thermoregulatory responses, many neural processes must occur: thermosensation, afferent signaling, integration plus coordination, and efferent responses, including postganglionic SNA. The neural control and regulation then activate or inhibit biological end organs to elicit a homeostatic response to the cold stress (Table 3).