A book on diabetes will remain grossly incomplete if we do not talk about stress. I suspect that many readers would be surprised that I did not even mention stress in the nine chapters so far. Some association between stress and health has been recognized for a long time by science, but till date, our understanding is rather naïve. I feel that the failure to understand stress and its effects clearly is owing to the lack of applying rigorous evolutionary logic to this field.

The history of stress-related health problems is interesting and has seen a number of ups and downs. From the beginning of the twentieth century, it has been speculated that diabetes and many other so-called “lifestyle” diseases are caused or aggravated by stress. These stress-related disorders included diabetes, hypertension, hypercholesterolemia, cardiovascular disease, gastric acidity, peptic ulcers, etc. After the World War I there was some setback to this argument due to evident reduction in some of the stress-related disorders, particularly diabetes and hypertension, during the Great War. It appeared that the stress of war not only failed to increase the incidence or severity but in fact reduced it dramatically [1–3]. Nobody doubted that there was tremendous stress on people during the Great War. But since there was no increase in stress-related disorders during the global stress period, there was a strong counterargument against stress having anything to do with these disorders. The antidiabetic effects of the stress of war were demonstrated once again during World War II. The story repeated itself in all subsequent long-drawn wars such as the Serbian or Sarajevan wars [3]. The war data are quite robust and consistent across wars. Clinical records, epidemiological surveys, and follow-up of known cases have consistently shown the ameliorating effects of war. Therefore it is unlikely to be a reporting or methodological bias. A common interpretation of the wartime suppression of diabetes and hypertension is that there is inevitable change in the diet during wartime. Unfortunately the diet hypothesis is more of hearsay and has not been pursued and tested with scientific rigor. Even if we get quantitative data on wartime diets, showing that there was a change in diet is not sufficient to conclude that the difference was caused by change in diet. A good test of the hypothesis would be to show that peacetime dietary interventions identical to the wartime diet produce the same effect. But this has never been done. Therefore all that we know is that war suppressed diabetes. Whether it was due to diet, stress of war, or some other effect is yet to be demonstrated.

While no one really knows why wartime stress decreased diabetes, in the last few decades, there has been an increasing belief in stress as the origin of an entire cluster of disorders, although some diseases such as peptic ulcers that were once thought to be stress induced are now excluded. The stress paradigm however is faced with a number of problems. The problems begin with the definition and measurement of stress itself.

Stress has been defined in a number of different ways, part of the difference being contextual. The word stress is used in animal, plant, and microbial physiology too. We would restrict ourselves to the use of the word in the field of medicine. Throughout the field of medicine, the definition and the methods (or the lack of them) of measuring stress have been elusive. Currently a disorganized set of concepts are being included under the name stress. Before we try to understand how and whether stress affects diabetes, we need to have a reasonable working definition of stress. A common and apparently reasonable definition is that of a challenge to one or more of the homeostatic mechanisms of the body or a departure from the normal homeostatic state in response to an environmental or behavioral stimulus. So, any environmental, social, or even an imaginary challenge that stretches the homeostatic mechanisms can be called a stress [4]. The definition sounds logical and should be acceptable as a working definition.

However there are a series of problems associated with this apparently simple definition. The homeostatic mechanisms of the body have substantial plasticity, and they appear to learn to accommodate a variety of challenges if faced repeatedly. Taking a simple example of a physical stress, if you run for a couple of kilometers, the mechanisms managing energy supply and energy balance are stretched rather too much, and you feel completely exhausted. However if you run 5 km every day, the mechanisms get adapted, and the systems adjust themselves such that you can easily run 2 km now without experiencing exhaustion. Something similar happens to psychological pressures too.

The corollary of it is that if you do not experience any challenge for sufficiently long time, the homeostatic mechanisms are slowly eroded. For example, a person who does not exercise for decades will be unable to run even 20 m. In this case, the homeostatic mechanisms appear to have deteriorated owing to chronic absence of challenge. Deterioration of a homeostatic system should also be considered a challenge to the homeostatic system. A challenge to the homeostatic system is stress by definition. This means that chronic lack of stress is a stress by itself. If lack of stress is a stress, then the definition of stress becomes meaningless. A definition of a concept needs to specify what it is and what it is not. If it is what it is not, the definition of definition is not satisfied, and the concept itself becomes meaningless.

This appears to reflect in the model stressors used by experimenters in different experiments too. For example, if rats are kept together, they indulge into fights and threat displays through which they establish and maintain a social hierarchy. This has been used as a model of “social stress” [5–8]. Other groups of experimenters addressing different questions have kept rats isolated in separate enclosures which they call “isolation stress” [9–12]. If being together is a stress and being isolated is also a stress, then one wonders whether there is a nonstress condition. Since various contradicting phenomena are dumped under a generic name of stress, there is contradiction in the effects of stress as well. Stress is immunosuppressive in certain context and enhancer of immunity in other [13–19]. Stress impairs learning and memory in some studies and enhances it in a few others [20–26]. Stress is proinflammatory in certain contexts [27, 28] and anti-inflammatory in others [29, 30]. The usefulness of stress as a generic term is therefore doubtful.

Another major problem lies in the circularity of stress-related arguments. If anything that challenges the homeostatic state is called stress, then stress changes the homeostatic state becomes a meaningless statement. Since inducing a physiological change is included in the definition of stress, the statement that stress affects physiology becomes a noninformative tautology. It effectively means that anything that changes the homeostatic state changes the homeostatic state. If you think that this statement is meaningful, then stress research is certainly meaningful. But if you feel this statement is meaningless, the logic behind entire stress research needs to be reexamined. A circularity problem of this type is not unique to stress. This philosophical problem is faced by many fundamental principles of science too. For example, the law of conservation of momentum states that the momentum of an object does not change unless an external force is applied. But force is measured by change in momentum, which means that there is no change in momentum unless the momentum changes.

Most researchers ignore this philosophical problem and indulge into the endocrinological nitty-gritty of the so-called stress-induced effects. For the time being we will follow this line of researchers and look at stress-related disorders without worrying about the definition of stress. All kinds of stresses are said to trigger a single pathway known as the HPA axis which stands for the hypothalamic–pituitary–adrenal axis [31–34]. This refers to the sequence in which a series of hormones are activated. The chain of events begins in the hypothalamus which receives nervous connections from many parts of the brain. After receiving certain kinds of signals which are believed to be “stress” signals, corticotropin-releasing factor (CRF) is produced by the paraventricular nucleus of the hypothalamus. CRF is carried to the anterior pituitary gland where it stimulates the production of adrenocorticotropic hormone (ACTH). ACTH release from the pituitary gland reaches the adrenocortical cells where it stimulates the production of glucocorticoids. This response takes a few minutes at the end of which raised levels of glucocorticoids can be detected in plasma. The rise can be as high as 20-fold of the basal levels. ACTH not only stimulates hormone secretion, it also induces hypertrophy and proliferation of the adrenocortical cells.

The end product of HPA activity is increased levels of glucocorticoids. In humans cortisol is the common form of the hormone, whereas in many animal models including rats, it is corticosterone. Either of them has a variety of actions on metabolism. What has been less appreciated is that they also have important effects on behavior. The metabolic effects consist of increased glucose production by liver utilizing amino acids that are mainly diverted from muscle. On the other hand the rate of muscle glucose utilization is reduced. Cortisol effectively depletes protein content of many tissues including muscle. It is also an effective immunosuppressor. The behavioral effects are less well known but nevertheless experimentally clearly demonstrated. Cortisol reduces risk-taking behaviors, increases submissive responses, and at the same time enhances learning, particularly learning in the context of coping strategies towards an environmental challenge [35]. The action of cortisol on hippocampus during sleep is important for memory consolidation [25, 36]. Cortisol enhances hippocampus-dependent declarative memory formation but suppresses amygdala-dependent emotional memory formation. The natural cortisol rise during late sleep may thus protect from overshooting emotional memory formation, a mechanism that could be potentially important for a diplomat who needs to take cool-headed decisions [25]. Although cortisol has many specific adaptive roles in fine manipulations of behavior, research on behavioral functions of cortisol appears to be largely overshadowed by the label of “stress hormone.”

Stress researchers are so much obsessed by the association of stress and HPA axis that knowingly or unknowingly most of them define stress as anything that triggers the HPA axis. Some authors even overtly define stress by the HPA effect. Here lies the circularity of the argument in a more detailed form. The belief that all kinds of stresses stimulate the HPA axis is augmented by defining stress this way. Let us take an example. It can be argued that an alpha male in a primate group is under stress because with every potential competitor, its alpha position is at stake. A low-ranking male, on the other hand, has little to lose from losing another combat, and therefore, it can easily give a submissive display and avoid a conflict. Therefore alpha males must be under greater stress and uncertainty than low-ranking males. However, stress researchers will not agree with this argument because alpha males generally have lower corticosteroid levels [4]. This makes the researchers believe that low-ranking males are at a greater stress. Whenever alpha males were found to have higher corticosteroid levels, it was immediately concluded that alpha males are under greater stress [37]. This exemplifies how people use HPA response to define stress. If stress is defined this way, stress induces HPA activity is a meaningless statement. We will be able to test the hypothesis that stress activates HPA only if we have any independent quantitative measurement of stress. If stress itself is measured by corticosteroid levels, then stress increases corticosteroid levels cannot be the inference. And if stress is equated to corticosteroid levels, then there is no need to use the word stress. We can simply treat the phenomenon as raised corticosteroid levels. This would be a more objective working approach that does not involve any ill-defined terms such as stress.

The example of romantic love is even more entertaining. A number of hormonal changes are associated with the first phase of falling in love. One of the observations is that the corticosteroid levels are raised [38, 39]. Based on this observation some researchers have concluded that love is a stress. They do not appear to feel the need for any other justification as to why they want to call love as stress. A rise in cortisol is a sufficient indication of stress for them. This aptly demonstrates how HPA axis is used as a criterion to identify something as stress. This biased definition of stress has given rise to the highly flawed perception that all types of stressors trigger the HPA axis. This becomes a tautological argument once again saying anything that triggers the HPA axis, triggers the HPA axis.

The idea of all types of stresses triggering a single pathway appears very strange from the standpoint of a behavioral ecologist. Behavioral ecologists study the diversity of behavioral strategies displayed in response to a diversity of environmental and social challenges and therefore naturally expect an equally wide diversity in the physiological response of the body that accompanies a behavioral response. The notion that all types of challenges trigger the same stereotyped response therefore appears weird. How and why could have evolution favored such a stereotyped dumb response at the physiological level when there is a wide array of flexible and smart responses at the level of behavior? The answer to the question lies in the circularity of definition and the selective choice of stress models. Since knowingly or unknowingly HPA axis is used in defining stress and choosing model stressors, there is a false impression that all kinds of stresses trigger HPA axis. A number of stressors could be triggering different pathways, but stress researchers are generally reluctant to call them stressors because they do not see rising corticosteroids.

The evidence is on the contrary. As mentioned earlier wartime stress appears to have different effects than peacetime stress. Some stressors are reported to evoke hypocortisolemia rather than hypercortisolemia [40–42]. Particularly interesting is the fact that stress and depression caused by war exposure is associated with low rather than high cortisol levels, and in war veterans, cortisol levels are negatively associated with exposure to actual combat and violence [43, 44]. Evidence also points that the same stressor can elicit different hormonal responses in different individuals. For example, to the same stressor of a conspecific competitor, chimpanzees responded by increasing testosterone whereas bonobos responded by increasing cortisol [45]. This is in accordance with the known differences in the behavior of the two species. Chimpanzees are more aggressive and risk takers than bonobos [46], and they responded by testosterone, whereas bonobos, the more social cousins of chimpanzees, responded by cortisol. Even within a species, in response to a potential competitor, high-ranking individuals give a testosterone response, whereas low-ranking individuals give a corticosteroid response [47, 48]. Apart from social ranking differences in individuals’ behavior and copying styles also affect stress responses [49]. It is possible therefore that in reality, different challenges evoke different responses and also the same challenge evokes different responses in different personalities, but it is the thinking trap of researchers that makes them believe that HPA is the only stress response pathway.

In fact some alternative responses are quite well known, and perhaps there are a number of others not so well known. Perhaps more important than HPA axis, there is another well-known stress response which is often referred to as a flight-or-fight response [50] with a quasi-­evolutionary explanation. This is a response of the autonomous nervous system. In the field of medicine it is assumed that a typical natural cause of stress is say suddenly detecting a predator. The natural reaction to a predator is either to run away as fast as possible or to fight back with all wits and vigor. This is called a flight-or-fight response. This response is mediated by activating the sympathetic nervous system which mobilizes a lot of energy in a short time by undertaking lipolysis, glycogenolysis, and liver gluconeogenesis resulting into release of more glucose and fatty acids that supply fuel to the fight-or-flight-related muscle activities. An old belief associated with this response is that although the stress and anxiety typical of modern urban life are of a very different nature, it triggers the same pathway as a predator response. So there is mobilization of fuel sources in the same way. Fatty acids and glucose levels in the plasma go up, but unlike classical predator situation, there is not any intense muscle activity to burn of the excess glucose and fatty acids. This eventually leads to the bad effects of stress including hyperglycemia, hyperlipidemia, hypercholesterolemia, atherosclerosis, and cardiovascular disease. The logic appears temptingly simple and convincing.

The only problem with the fight-or-flight hypothesis is that it is too naïve to be true. The first naïve assumption is that natural stressful situations are only predator or predator-like situations, and fight-or-flight response is the only adaptive response. This belief suits well for an urban biologist who does not know the complexities of life in the wilderness. The life of any social animal, primates in particular, is faced with a wide diversity of environmental and social challenges, and there is an equally wide diversity of behavioral responses to these challenges. Describing them would demand a separate book in itself, so I would avoid the temptation here. The only point relevant here is that fight or flight is not the only adaptive response to all types of challenges in the wilderness. This is recognized by many researchers, and alternatives are also suggested [51–53]. I would like to add two more “f”s to the response categories. They are fox and freeze in addition to fight or flight. Fox refers to the fox in Aesop’s fables. It is equivalent to the diplomat strategy that we introduced earlier consisting of coping with a challenge by social manipulation rather than physical strength. Fight and flight together constitute our soldier strategy. Whereas freeze is a response of being helpless and seeking neither a soldier nor a diplomat ­solution. It would be somewhat counterintuitive, but in real life, a freeze response can give substantial returns in a number of situations. In fact, some species are specifically adapted to the freeze response. Quite well known to naturalists, ­physiologists appear to have largely ignored the importance of a nonsoldier–nondiplomat freeze response. A variety of species of quails inhabit grasslands, scrub, or forest floor. They are mainly ground-dwelling birds with relatively poor flight muscles. Having limited flight capacity, they need alternative strategies for escaping ground predators such as wild cats and jackals and areal predators such as falcons and hawks. They rely on a superb camouflage aided by a completely immobile crouched and frozen posture on suspecting predator presence. With the sharp vision of falcons and eagles, a combination of superb camouflage and complete freezing is essential, and no other active defense is likely to work.

It is not only for species that specialize in freeze response. To relate some brief anecdotes, I have experienced attacks by two dangerous species of animals in the wild, elephant and tiger. In both the cases standing firm on the ground facing the animal is the most successful strategy, while an attempt to run away increases the risk, and fighting back is simply out of question. One of my close friends and a well-known wildlife photographer tells me that when attacked by a lone tusker, he wanted to run away, but fear gripped him in such a way that he just could not move. He froze helplessly facing the tusker, and the tusker, after a threatening rush towards him, suddenly turned away from a distance of about 10 m and vanished into the thickets. My friend still could not move for the next few minutes. I am sure this is an evolved response which has a survival value. When fight, flight, or fox are unlikely to work, there is one more class of responses and there must be an elaborate neuronal and physiological set of mechanisms mediating it. But stress biologists are so obsessed by the fight-or-flight idea that they have a mental block towards other possible responses.

The picture with four “f”s instead of two is still very naïve and unable to fully capture the context-dependent diversity of responses. Nevertheless it is better than the classical presumption that the body gives a single stereotyped response to all kinds of stresses. The main point that needs to be driven is that different types of responses exist in nature, and each response has different physiological requirements. Therefore we expect that different neuroendocrine pathways must be used in different types of responses. The classical belief of a single pathway covering all stresses is not logical.

Let us then look into the details of the physiological requirements of the four “f”s (Table 10.1). The fight-or-flight response is a soldier response and needs fast action, muscle strength, and rapid mobilization of energy. It is also more risk prone and should prepare the body in anticipation of injuries. This is achieved by the sympathetic response in full bloom. The sympathetic response is the fastest of all, particularly in comparison with the HPA response which takes several minutes for execution. The sympathetic response begins in fractions of a second and is characterized by very quick mobilization of energy by increased glycogenolysis, lipolysis, and gluconeogenesis. This is accompanied by increased muscle efficiency and muscle utilization of glucose. A battery of other responses elicited by sympathetic stimulation is also supportive to muscle action or its anticipated consequences. There is increased arterial pressure, and blood flow is managed in such a way that there is increased blood flow to active muscles concurrent with decreased flow to organs such as the gastrointestinal tract that can be temporarily ignored. In addition, a number of changes occur in response to anticipation of injuries such as increase in the rate of blood coagulation, spurt of EGF and NGF release in the saliva [54–57], and suppression of pain responses. Testosterone and sympathetic response interact in a synergistic way, and testosterone is needed for many of the above pathways. For example, the expression of β adrenergic receptors on adipocytes which receive the sympathetic response is partially testosterone dependent [58]. If the receptors are depleted owing to testosterone deficiency, sympathetic stimulation of lipolysis is ineffective. Similarly testosterone facilitates synthesis of but not release of EGF. EGF synthesized under the influence of testosterone is released on sympathetic stimulation [59–61]. Thus the effects of the two are synergistic and interdependent and together are supportive of an explosive muscle action such as in fight or flight. However, sympathetic activity in the absence of testosterone might have a different set of effects. There would be glucose production but no lipolysis. As a result amino acids instead of fatty acids will be used for gluconeogenesis. Testosterone facilitates muscle protein synthesis. Testosterone deficiency and increased gluconeogenesis would mean more protein degradation and conversion to glucose. Most of this protein comes from muscle, so sympathetic response in the absence of aggression will have diametrically opposite consequences. It will lead to weakening of muscle and building of fat. Aggression is therefore a necessary part of the fight-or-flight response.

The fox or diplomat response need not be as rapid as the fight–flight response. It is substantially slower and thoughtful involving more of cognitive brain and less of muscle strength and rapid action. This is when the HPA axis is most appropriate. A comparison of the similarities and differences between the sympathetic versus HPA mechanisms demonstrates their differential use. As opposed to the sympathetic response, the HPA response is slower. Rapid explosive action is sustained for a short time, whereas diplomat activities are more long drawn. In accordance with this, adrenalin or epinephrine, a marker of sympathetic response, has a very short half-life, whereas half-life of cortisol is substantially greater. One of the practical consequences of this difference is that for an experimenter, it is easier to study the corticosteroid response. Studying the epinephrine response is more difficult since it gives little time to collect a sample and preserve/process the sample before the levels change. This is likely to be another practical reason why stress researchers have focused so much on HPA [4]. Convenience of work and availability of methods have often decided the direction of research throughout the history of science. Both sympathetic and HPA responses mobilize energy and increase plasma glucose by liver glucose production, but cortisol reduces muscle glucose uptake implying that this increased glucose is not meant for muscle. We can suspect then that it must have evolved to supply the brain instead.

EGF and NGF are secreted mainly by salivary glands. Both are involved in wound healing, but NGF is also important in brain function. Cortisol stimulates salivary NGF selectively and not EGF [62]. This also implies that cortisol response is in support of brain and not physical activity. In addition the behavioral and cognitive effects of corticosteroids including risk aversion and facilitated learning [35, 63] indicate that HPA response is specifically an adaptation for diplomat behavior. Success of diplomat behavior depends largely on social support. Therefore cortisol would be an appropriate response when seeking social support. If and when good social support is achieved, there should ideally be a feedback regulation of cortisol response since the target for the response is already achieved. This is apparently done by oxytocin. Oxytocin is stimulated by friendly social contact such as hugging, and it negatively regulates cortisol production [64]. All the facts together can be interpreted to mean that cortisol is not a marker of “stress” in general, but it is a marker of diplomat or fox response to an environmental or social challenge. In a wide variety of vertebrate species, where physical strength and aggression appear to determine social dominance, subordinate individuals need to suppress aggression, and they have higher levels of cortisol [4, 65–78]. Long-term cortisol treatment inhibited aggressive behavior in fish [79, 80]. ACTH, another player in the HPA, also has aggression suppression action [81]. Cortisol is negatively associated with aggression in humans in a number of studies, but there is some inconsistency in these data presumably due to different definitions and measures used for aggression [82–86]. This interpretation is further backed by the finding that in low-cortisol individuals, testosterone is positively correlated with dominance, but in high-cortisol individuals, testosterone is either not related or negatively related to dominance [87, 88]. It is clear that testosterone would help achieve dominance if physical strength and aggression are a determinant of dominance. But if dominance is decided by nonphysical means, testosterone would be irrelevant and perhaps counterproductive. This is the state of diplomat behavior marked by high cortisol. Therefore in high-cortisol individuals, testosterone is not related or negatively related to dominance. This might be the most logical explanation of the interaction between testosterone and cortisol. The role of corticosteroid as a marker of diplomat behavior is also illustrated by the observation that the ability to aggressively attack another individual (displacement aggression) reduces shock-induced corticosteroid response in rats [4]. It might be surprising in this connection that corticosteroids are immunosuppressive since subordinate individuals are under risk of attack by dominant individuals. But interestingly there appears to be a built-in flexibility in the immunological effects of corticosteroids. In situations where the risk of injury is high, the immune system becomes resistant to the immunosuppressive action of corticosteroids [89–91]. It is possible therefore that in social situations where submissive displays can effectively inhibit aggression by dominant individuals, immunosuppressive effects are seen, whereas if submissive behavior does not prevent attacks, immunosuppressive action of corticosteroids is spared.

The freeze response shares some physiological requirements with fox response in terms of avoiding risk, suppressing aggression, and disinvesting from muscle. But the main difference is that here in addition to soldier functions, even the diplomat functions are switched off. Therefore even higher-order cognitive functions are not involved, and in addition to disinvestment from muscle, there would be a disinvestment from cognitive functions as well. Risk and aggression avoidance would need glucocorticoids and cholesterol, whereas disinvestment from muscle is mediated by insulin resistance. Since in the freeze response there is reduced demand on brain function, there should be disinvestment from brain energy supply too. This is particularly true for the freeze response since two main classes of brain functions are put off simultaneously. Complex nerve muscle coordination is not required, and many of the cognitive functions are also not required simultaneously. Therefore the activity and thereby energy requirement of the brain should come down substantially in comparison with both soldier and diplomats. Unlike muscle, disinvestment from brain is not brought about by insulin resistance since brain glucose uptake is insulin independent. The only way it can be brought down is by reducing the amount of glucose transporter glut-1 across the blood–brain barrier. This is a novel testable prediction of our model of multiple stress responses. There is little research input on glut-1 dynamics, but it is demonstrated in the rat model that in diabetes, glut-1 levels decrease substantially [92, 93]. It is also known that glut-1 levels are fine-tuned according to brain activity and energy demand [94]. But research so far is too scanty to be conclusive. Still we will explore the possible changes brought about by glut-1 dynamics in a later chapter.

Thus the different types of responses need different sets of physiological mechanisms. How frequently an animal gives these responses would slowly modulate the overall physiological and behavioral tendencies of it. What decides what type of response the body should give when faced with a challenge? It is partly decided extrinsically, i.e., by the nature of the challenge and partly by intrinsic factors that include the genetic constitution, developmental history, and all other factors that constitute the existing nature or personality of an individual. What happens when there is ambiguity in deciding which response is the most appropriate? For example, if a competitor is stronger, an HPA response is appropriate, and if it is weaker, then an aggressive response is more appropriate. But what if one is unable to judge the relative strength of the opponent? This is what happens when there is a conflict situation with a stranger or one of a comparable rank. In such a case as an interaction begins by giving threat displays, taking each other’s judgment and perhaps actual fight, sooner or later it becomes clear who is stronger. Until then it would be ideal to be prepared for both possible outcomes. This is what actually happens in nature. Facing an aggressive encounter with an uncertain outcome, both sympathetic response and HPA response are simultaneously launched. When the winner–loser status becomes clear, the winner rapidly retracts the HPA response and loser retains it. Winner has an elevated testosterone level, and loser lowers its testosterone. This has been demonstrated across a number of species [4, 95]. It is well known that many types of stressors trigger both catecholamine and cortisol response, but some are dominated by catecholamine, and others by cortisol [4] demonstrating the flexibility and fine-tuning of the responses.