Adrenal Responses to Stress

Abstract

Based on concepts proposed by Langley, Cannon, and Selye, adrenal responses to stress occur in a syndrome that reflects activation of the sympathoadrenal system and hypothalamic–pituitary–adrenocortical (HPA) axis; and a “stress syndrome” maintains homeostasis in emergencies such as “fight or flight” situations, but if the stress response is excessive or prolonged then any of a variety of clinical disorders can arise. The idea of a unitary sympathoadrenal system does not account for evidence that different stressors elicit different patterns of autonomic responses, with exposure to some stressors differentially affecting sympathetic noradrenergic and adrenomedullary hormonal activities. Instead, adrenomedullary responses to stressors are more closely tied to adrenocortical than to sympathetic noradrenergic responses. Distress involves concurrent activation of the HPA and adrenomedullary neuroendocrine systems.

Cannon, Homeostasis, and the Sympathoadrenal System

Walter B. Cannon coined the term “homeostasis” to describe the maintenance within acceptable ranges of physiological variables such as blood glucose and core temperature (Cannon 1929a, 1939). Cannon extended this concept to include psychosocial threats to homeostasis. In the early 1900s, he described for the first time the acute changes in adrenal gland secretion associated with what he called “fight or flight” responses.

Cannon taught that noxious environmental or internal stimuli threatening homeostasis (e.g., exposure to cold, hypotensive hemorrhage, traumatic pain, insulin-induced hypoglycemia, emotionally distressing antagonistic encounters) concurrently activate the adrenal gland to release epinephrine (EPI, synonymous with adrenaline) and the sympathetic nervous system. He considered the two effectors to function as a unit, which came to be termed the “sympathoadrenal,” “sympathico-adrenal,” of “sympathoadrenomedullary” system. In emergencies, sympathoadrenal activation would tend to restore homeostasis. Indeed, in 1939, Cannon formally and erroneously proposed that the chemical messenger of the sympathetic nervous system was identical to EPI (Cannon and Lissak 1939). Cannon emphasized that disparate threats to homeostasis incite the same sympathoadrenal response (Cannon 1929b, 1939). The notion of a unitary sympathoadrenal system continues in medical thinking (Cryer 1980; Shah et al. 1984; Sofuoglu et al. 2001; Kvetnansky et al. 1995).

Selye, Stress, and the General Adaptation Syndrome

Hans Selye, who popularized stress as a medical scientific idea, defined stress as (or a state resulting in) “the nonspecific response of the body to any demand upon it” (Selye 1974). By non-specific he meant a shared element regardless of the nature of the demand or stressor. In this regard, his view agreed with that of Cannon.

Selye proposed three universal stages of coping with a stressor—the “General Adaptation Syndrome”—an initial “alarm reaction,” analogous to Cannon’s “fight or flight” response, a stage of adaptation, with resistance to the stressor, and eventually a stage of exhaustion and organismic death. In Selye’s early experiments, after injection of any of a variety of tissue extracts or of formalin into rats, the animals developed a pathological triad of enlargement of the adrenal glands, atrophy of lymphoid tissue in the thymus, spleen, and lymph nodes, and bleeding gastrointestinal ulcers. It was later demonstrated that these changes are associated with and to at least some extent result from activation of the HPA axis. Steroids released into the circulation from the adrenal cortex are required for resistance but are also responsible for pathological changes. Selye’s concept that prolonged stress can produce disease is now widely accepted.

Selye acknowledged that responses to stressors have a specific component that tends to counter effects of the stressor; however, after removal of specific responses from consideration, a non-specific stress syndrome would remain. Chrousos and Gold (1992) modified Selye’s doctrine of non-specificity by proposing that above a threshold intensity, any stressor elicits the stress syndrome.

More than a half century elapsed before Selye’s doctrine of non-specificity underwent experimental testing, which failed to confirm it (Pacak et al. 1998). Even so, modern lay and even scientific literature continues to accept the notions of a unitary stress response, a central neural stress system, and a stress “syndrome.” For instance, a Google search yielded about 4,000,000 hits for “the stress response,” 714,000 hits for “the stress system,” and 202,000 for “the stress syndrome.”

Just as Cannon emphasized activation of the sympathoadrenal system to maintain homeostasis, so did Selye and his students emphasize activation of the HPA axis in the General Adaptation Syndrome.

Systems Concepts of Stress, Allostasis, and Allostatic Load

Modern concepts view stress as a sensed threat to homeostasis (McEwen and Stellar 1993; Goldstein and McEwen 2002), in which the response has a degree of specificity depending on the particular challenge to homeostasis and the organism’s perceptions of the stressor and ability to cope with it (Goldstein 2001).

Concepts of Scientific Integrative Medicine

Scientific integrative medicine is not a treatment method or discipline but a way of thinking that applies systems concepts to conceptualize physiology and pathophysiology. The scientific integrative medicine approach provides a framework for understanding highly complex and dynamic challenges to our integrity as organisms and in turn for developing novel treatments based on this complexity and dynamism (Goldstein 2006). It also leads to systems definitions of stress and allostatic load, as discussed below.

Homeostats

Central to scientific integrative medicine in general and to systems concepts of stress specifically is that the body possesses homeostatic comparators, called “homeostats” (Goldstein 1995). Each homeostat compares information with a set point for responding, determined by a regulator (Fig. 1). A sufficiently large sensed discrepancy between afferent information about the level of the monitored variable and the set point for responding (error signal) elicits altered activities of effectors, the actions of which on the monitored variable decrease the discrepancy.

Fig. 1.

A physiological homeostatic system. As the level of the monitored variable changes, afferent information is compared with a set point or other algorithm for responding, and the sensed discrepancy leads to altered activities of effectors. Note the odd number of (−) signs, indicating a negative feedback loop. In response to a continuous perturbation, the level of the monitored variable reaches an apparent steady state

Homeostatic systems operate according to a few principles that despite their simplicity can explain complex physiological phenomena and might help resolve controversial issues in the area of stress and disease.

Negative Feedback Regulation

Physiological homeostatic systems entail negative feedback regulation of monitored variables such as core temperature, blood pressure, serum osmolality, and glucose levels. Disruption of a negative feedback loop, by preventing afferent information from reaching the brain, inability to process the information and regulate effector functions correctly, or dysfunction or loss of effectors, increases fluctuations in levels of monitored variables (Sharabi et al. 2003).

Positive feedback loops are inherently unstable, and conversion from a negative to a positive neurocirculatory feedback loop presages rapid decompensation. For instance, one can understand transitions from heat stress to heat shock and from compensated to decompensated heart failure in terms of positive feedback loops (Goldstein 2006).

Multiple Effectors

Multiple effectors regulate levels of most monitored variables of the body (Fig. 2). Having available multiple effectors extends the range of control, allows at least some regulation of the monitored variable if a particular effector fails (compensatory activation), and enables elaboration of specific, adaptive effector patterns.

Fig. 2.

Compensatory activation. One advantage of multiple effectors is compensatory activation of alternative effectors if one effector fails, enabling control of the monitored variable. For instance, thyroidectomy augments sympathetic nervous system (SNS) responses to cold exposure

Examples of compensatory activation in physiology include augmentation of sympathoneural responsiveness by adrenalectomy, hypophysectomy, or thyroidectomy (Udelsman et al. 1987; Goldstein et al. 1993; Fukuhara et al. 1996). Compensatory activation of other vasoactive systems after destruction of the sympathetic noradrenergic system helps to explain why many workers, including Cannon, erroneously concluded that the sympathetic nervous system acts only as an “emergency” system (Gauthier et al. 1972; Julien et al. 1990).

Patterning of neuroendocrine, physiological, and behavioral effectors increases the likelihood of adaptiveness to the particular challenge to homeostasis, providing another basis for natural selection to favor the evolution of systems with multiple effectors.

Effector Sharing

Different homeostatic systems can interact by sharing effectors (Fig. 3). This can help understand clinical associations that might otherwise be unexpected. For instance, sharing of the adrenomedullary hormonal system by the barostat and glucostat can explain hyperglycemia in gastrointestinal hemorrhage, and sharing of the vasopressin system by the barostat and osmostat can explain hyponatremia in heart failure.

Fig. 3.

Effector sharing. Sharing of an effector by multiple homeostats can explain unpredicted consequences and syndromic features of disease processes. For instance, in heart failure, decreased aortic filling increases levels of vasopressin (AVP), which, as the anti-diuretic hormone, promotes retention of free water, explaining hyponatremia attending heart failure

Primitive Specificity

Beginning soon after adequately sensitive assay methods of plasma levels of norepinephrine (NE) and EPI became available, evidence rapidly accumulated for different noradrenergic versus adrenergic responses in different situations (Cryer 1980; Robertson et al. 1979; Young and Landsberg 1979; Young et al. 1984). These findings did not fit with the notion of a unitary sympathoadrenal system.

In the sheltered confines of a laboratory, with controlled temperature and ad libitum water, nutrients, and calories, mammals do not seem to require an intact sympathetic nervous system (Cannon 1931). It is by now clear that even under resting conditions pulse-synchronous bursts of skeletal muscle sympathetic nerve activity are detectable, and NE continuously enters the venous drainage of most organs. Activities of daily life, such as standing up (Lake et al. 1976), digesting a meal (Patel et al. 2002), speaking in public (Gerra et al. 2001), and walking—i.e., not only emergencies—are associated with rapid adjustments in sympathetic nervous system outflows.

Adrenomedullary hormonal system activity, and thereby EPI levels, respond to global or metabolic threats, such as hypoglycemia, hemorrhagic hypotension, exercise beyond an anaerobic threshold, asphyxiation, emotional distress, and shock. Evidence also has accumulated for an association between NE and active escape, avoidance, or attack, and an association between EPI and passive, immobile fear. Thus, in contrast with Selye’s doctrine of non-specificity, according to the systems theory of stress activities of effector systems are coordinated in relatively specific neuroendocrine patterns.

Studies of humans exposed to cold or with mild core hypothermia have provided support for the notion of primitive specificity of neuroendocrine stress responses. Cold exposure increases plasma NE levels, with little if any increases in plasma EPI or ACTH levels, consistent with sympathetic neuronal activation and relatively less adrenomedullary and adrenocortical activation. Mild core hypothermia also increases antecubital venous levels of NE but not EPI (Frank et al. 2002). Both NE and EPI levels in arterial plasma increase in this setting, but with larger NE responses (Goldstein and Frank 2001). When these homeostatic mechanisms are overwhelmed and core temperature falls, then increased adrenomedullary secretion results in high circulating EPI levels, increasing generation of calories (Staten et al. 1987) and eliciting cutaneous vasoconstriction, which decreases evaporative heat loss. The adrenomedullary activation is associated with distress, which motivates escape and avoidance learning.

Allostasis

Homeostasis implies set goal values for monitored variables; however, ranges of acceptable values are decidedly inconstant. Levels of physiological activity required to re-establish or maintain homeostasis differ, depending on continually changing conditions in which the organism finds itself—e.g., running versus standing versus lying down. “Allostasis,” a term used by Sterling and Eyer in 1988 (McEwen 1998), refers to levels of activity required for the individual to “maintain stability through change”—i.e., to adapt (McEwen 1998, 2000; Schulkin et al. 1998). Regulation around an altered apparent steady state is the essence of allostasis. Adaptations involving allostasis to cope with real, simulated, or imagined challenges are determined by genetic, developmental, and experiential factors.

Homeostat resetting redefines the conditions required to maintain homeostasis. This would be analogous to a different thermostatic setting in the winter compared to the summer. A neuroendocrine example is the hyperglycemia of exercise. Even in anticipation of the need for metabolic fuel, by activation of “central command,” the blood glucose level increases to a new steady-state value. Classically conditioned hyperglycemia provides another example of allostasis (Siegel 1972).

Resetting alters activities of multiple effector systems required to maintain allostasis, at least for short durations. During stress, short-term changes in homeostatic settings generally enhance the long-term well-being and survival of the organism. When superimposed on a substrate of pathology, however, homeostatic resetting can cause harm. For instance, in the setting of ischemic heart disease, global or patterned increases in sympathetic outflows from homeostat resetting would increase cardiac work, and the resulting imbalance between oxygen supply and demand could precipitate angina pectoris, myocardial infarction, or sudden death.

Allostatic Load

While they may be effective over a short interval, allostatic alterations may have cumulative long-term adverse effects. For instance, chronic elevation of blood pressure to ensure adequate blood flow to the brain might eventually lead to heart or kidney failure. “Allostatic load” (McEwen and Stellar 1993) refers to effects of prolonged continuous or intermittent activation of effectors involved in allostasis. In the analogy of the home temperature control system, allostatic load would increase if a window or door were left open. In this situation, one or more effectors might be activated frequently or even continuously. Long-term allostatic load—the wear and tear cost of adaptation—provides a conceptual basis for studying long-term health consequences of stress (Fig. 4).

Fig. 4.

Systems definitions of stress and allostatic load. In stress, the organism senses a discrepancy between afferent information about a monitored variable and a set point and other instructions for responding, altering activities of effectors to decrease the discrepancy. Allostatic load reflects wear and tear, which, if sustained and substantial enough, decreases effector efficiency, further activating the effector and accelerating wear and tear. Allostatic load can therefore eventuate in a destabilizing and pathologic positive feedback loop

Chronic effector system activation might alter the efficiency of the homeostatic system itself. For instance, chronic sympathetic nervous stimulation of the cardiovascular system could promote cardiovascular hypertrophy, “splinting” arterial baroreceptors in stiff blood vessel walls, in turn contributing to systolic hypertension and the risk of heart failure, kidney failure, and stroke. Moreover, an inappropriately large adrenomedullary response to a stressor might exaggerate the experience of distress (Schachter and Singer 1962). Exaggerated distress responses might increase the risk of worsening an independent pathologic process, such as in panic-induced angina pectoris (Wilkinson et al. 1998; Mansour et al. 1998).

Recently, panels of specific indices have been proposed as biomarkers of allostatic load (Juster et al. 2010). Because of the primitive specificity of stress response patterns, one might question the universal applicability of such panels. On the other hand, in chronic distress, a set of indices might reflect consequences of persistent activation of the adrenal cortex and medulla.

Systems Definitions of Stress and Distress

Even a simple homeostatic reflex reflects stress when a perceived discrepancy between a set point for a monitored variable and information about the actual level of that variable elicits compensatory responses that decrease the discrepancy.

Thus, one way of looking at stress is as a condition where expectations, whether genetically programmed, established by prior learning, or deduced from circumstances, do not match current or anticipated perceptions of the internal or external environment, and this discrepancy between what is observed or sensed and what is expected or programmed elicits patterned, compensatory responses (Fig. 4).

Distress is aversive to the organism, as evidenced by motivation for learning to escape or avoid the stressor. Selye characterized distress as unpleasant or harmful (Selye 1974), without separating these two very different characteristics. The systems theory of stress does not assume an equivalence of noxiousness (i.e., negatively reinforcing properties) with production of pathological changes; that is, the theory does not assume that distress causes disease.

Selye’s theory emphasized the non-specificity of the stress response, whereas according to the systems theory, the experience of distress depends on the character, intensity, and meaning of the stressor as perceived by the organism and on the organism’s perceived ability to cope with it. This is in line with Lazarus’s views on emotional stress and psychological coping (Somerfield and McCrae 2000). For an organism to experience distress therefore seems to require consciousness, to interpret the situation in terms of the ability to cope. This notion can help explain the finding that sedation with the benzodiazepine, alprazolam, attenuates the ACTH and EPI responses to glucoprivation (Breier et al. 1992).

Distress responses, as all stress responses, have a “purpose,” mitigating effects of a stressor in some way. This applies not only to neuroendocrine aspects of those responses (such as the glucose counter-regulatory actions of pituitary-adrenocortical and adrenomedullary stimulation during insulin-induced hypoglycemia) but also to psychological aspects (such as conditioned aversive and instrumental avoidance learning).

Distress responses evolved and continue to be expressed even in higher organisms, including humans who actually are only rarely exposed to truly “fight-or-flight” agonistic encounters, partly because of the importance of those responses in instinctive communication. Selye’s theory did not consider the communication function of distress.

An Adrenal Distress System

A fourth characteristic of distress is adrenal activation (Goldstein 1995, 2001).

A sufficient number of studies have been published in which adrenomedullary, adrenocortical, and sympathetic noradrenergic activities have been monitored simultaneously to enable a meta-analysis of the literature, to examine inter-relationships among alterations in activities of these key effector systems upon exposure to different stressors and reassess long-standing concepts about these inter-relationships.

A recent study based on computer searches of PubMed (Goldstein and Kopin 2008) retrieved publications describing original data about plasma EPI, ACTH, and NE levels measured before and during or after exposure to stressors. Magnitudes of responses were categorized according to the following criteria. If there was no significant change in plasma levels of the dependent variable, a score of 0 was assigned. If there were a statistically significant increase but less than a doubling of the pre-stress baseline level, a score of 1 was assigned. If there were at least a doubling of the baseline value, up 3 times the baseline value, a score or 2 was assigned. If there were a large increase, from 3 up to 10 times the baseline value, a score of 3 was assigned. If there was a massive increase to ≥10 times the baseline value, a score of 4 was assigned. A total of 15 different stressors were identified for which the available literature satisfied the above criteria.

Mean EPI responses were strongly positively correlated with mean ACTH responses (Fig. 5) and less strongly with NE responses. Plasma EPI responses were larger than expected for NE responses during hypoglycemia and smaller than expected for NE responses during cold exposure without hypothermia, orthostasis, and active escape/avoidance. Plasma NE responses were larger than expected for ACTH responses during cold exposure without hypothermia and severe/exhausting exercise and smaller than expected for ACTH responses during hypoglycemia.

Fig. 5.

Mean values across 15 different stressors for plasma levels of a epinephrine (EPI) and corticotrophin (ACTH); b EPI and norepinephrine (NE); and c NE and ACTH. Dashed lines indicate lines of best fit

The results of this meta-analysis therefore support a close association between adrenomedullary and HPA responses across a variety of stressors, an association that is stronger than that between adrenomedullary and sympathetic noradrenergic responses. The findings favor the concept of an adrenal distress system and support the notion of “primitive specificity,” according to which stress responses occur in relatively specific neuroendocrine patterns. By promoting homeostasis, such patterning would have provided clear advantages in natural selection and therefore evolved. In contrast, Cannon’s and Selye’s theories, based on stereotyped responses regardless of the stressor, do not account adequately for outliers in the scatter plots relating EPI to NE and NE to ACTH responses. For instance, plasma EPI and ACTH responses to hypoglycemia are much larger than expected for NE responses, and plasma NE responses to cold exposure without hypothermia are much larger than expected for EPI or ACTH responses.

Thus there seems to be at least as good a justification for the concept of coordinated adrenocortical–adrenomedullary responses as for coordinated adrenomedullary–sympathoneural responses in stress. This coordination may involve corticotropin releasing hormone driving adrenocortical and adrenomedullary outflows (Yoshida-Hiroi et al. 2002) as well as interactions between adrenocortical and adrenomedullary chromaffin cells (Zuckerman-Levin et al. 2001; Ehrhart-Bornstein and Bornstein 2008).