Abstract
The ability to maintain health, or recover to a healthy state after disease, is an active process involving distinct adaptation mechanisms coordinating interactions between all physiological systems of an organism. Studies over the past several decades have assumed the mechanisms of health and disease are essentially interchangeable, focusing on the elucidation of the mechanisms of disease pathogenesis to enhance health, treat disease and increase healthspan. Here, I propose that the evolved mechanisms of health are distinct from disease pathogenesis mechanisms and suggest that we develop an understanding of the biology of physiological health. In this Perspective, I provide a definition of, a conceptual framework for, and proposed mechanisms of physiological health to complement our understanding of disease and its treatment.
Introduction
The past fifty years of biological research has taken a disease-centric approach to understanding the mechanistic underpinnings of health. Indeed, it is easier to conceptualize a disease state than a healthy state of an individual. We know that disease is caused by the presence of an insult that can be genetic, environmental, or caused by a decline in normal physiological function, such as during aging. Health on the other hand, seems to be more metaphysical, a so-called state of “well-being”. This makes it problematic for describing health mechanistically. We know that an individual is healthy when there is an absence of disease, and this has led to a view of health as a passive state. In that view, disease occurs when there is a disruption in the system due to the presence of an insult, and therefore the mechanisms that drive disease are related to the ones that promote health: simply removing an insult or antagonizing a disease pathogenesis pathway is sufficient to promote health. We know that this is not necessarily true. One of the triumphs of biomedical research is that we can describe in exquisite detail the disease path taken by an individual when it is sick and the disease pathogenesis pathways that drive the disease course. We also have an excellent understanding of risk factors and mechanisms that allow avoidance or eradication of an insult. Thus, our primary methods for treating diseases are to block pathogenic responses or remove the insult, rather than inducing pathways that work to maintain health. However, the goal of modern medicine is to restore a patient to a healthy state, and a goal for biologists is to define methods to extend the healthspan of an individual. As we embark on a new decade of biomedical research, I propose that rather than continuing to ask “how should we treat disease?” we might instead ask, “how can we promote and maintain health?” These are not the same thing.
Before exploring the possible directions we can pursue to address this question it is important to provide a brief explanation of the viewpoint I take in this Perspective. Health is not a passive state. Health is an active process that enables an organism to adapt to fluctuations in its intrinsic and extrinsic environments to maintain health or recover to a healthy state after disease occurs. This has important implications for our thinking about health mechanistically – that we have evolved adaptive mechanisms that promote the healthy state of an individual and that these health mechanisms are generally distinct from those that drive disease. We are comfortable with this idea in the context of homeostasis, where the plasticity of an organism’s physiology enables it to adapt to changes that occur due to normal physiological processes or the normal fluctuations that occur in an organism’s environment. We are also comfortable with this idea in the context of defense mechanisms which protect an individual from hostile environments and enable the individual to antagonize the primary causes of disease. However, we have also evolved adaptive mechanisms that enable an individual to withstand the presence of an insult rather than antagonize it. As an example, much of immunology and infectious disease research has been based on the assumption that to remain healthy when challenged with an infection, all that is necessary is to mount an immune response to eliminate it (Raberg et al., 2007; Schneider and Ayres, 2008). For fifty years, we have focused our efforts in controlling infection on developing vaccines and antibiotics and elucidating the mechanisms of the immune response. A challenge to this way of thinking arises when a pathogen infects and replicates inside a host without causing disease. With the discovery of the cooperative defense system that enables a host to adapt to the presence of a pathogen with disease tolerance and anti-virulence defensive health mechanisms, we are beginning to understand this phenomenon (Sanchez et al., 2018; Schneider and Ayres, 2008). We appreciate that we have specialized health mechanisms that enable an individual to withstand a hostile environment such as drought or heat, but the past decade of research with infectious disease has demonstrated that we have barely tapped into defining the full spectrum of these defensive health mechanisms for hostile conditions (Jamieson et al., 2013; Luan et al., 2019; Rao et al., 2017; Sanchez et al., 2018; Schieber et al., 2015; Wang et al., 2016). Furthermore, why individuals vary in their ability to homeostatically adapt to normal fluctuating conditions and how an individual can endure a new internal state when homeostasis cannot be maintained remains largely unknown. By contrast, we can describe in detail how disease occurs under these conditions in which homeostasis is disrupted.
The goal of this perspective is to encourage a shift in how we conceptualize health. I provide a conceptual framework for investigators from diverse disciplines within biomedical research for thinking about health in the context of their field and disease interests, propose the mechanisms of health and provide a discussion on how to elucidate health mechanisms in our experimental systems. I focus my discussion largely at the physiological level and in mammals. While I discuss a diverse array of health conditions, I largely borrow examples from the host-microbe/infectious disease field to help constrain my discussion and allow for comparisons and conceptualization, all of which can apply to non-infectious diseases. Although I focus on physiological health, the framework, concepts and approaches I describe here are applicable to mental health as well. Only when we understand health mechanistically will we truly be able to treat disease.
What is physiological health?
Conceptualizing health in a simple two-dimensional health space
I anticipate that by this point, the reader has said to themselves, “I do study health”. However, scientists largely study disease, irrespective of the context, and elucidating the mechanisms of disease pathogenesis does not equate to elucidating the mechanisms of health. Imagine a simple two-dimensional health trajectory of an individual over time. The y-axis represents the continuum of health, with maximal health at the top and minimal health (disease) at the bottom, and the x-axis represents time (Figure 1A). At any time, the individual can shift along the y-axis representing changes in the health state of the individual. For example, if the individual is exposed to hostile environments such as an infection, it can become sick, shifting the trajectory towards disease. For an individual on the path to death, this trajectory will continue to decline until termination. For individuals that will ultimately recover and return to health (resilience), the trajectory will exhibit an inflection point at some time post insult-exposure and move the trajectory along the y-axis back to a healthy state. For other individuals, the health trajectory does not move along the y-axis after insult-exposure, indicating the capacity of that individual to maintain health in the presence of the insult (maintenance). What are the mechanisms that control the position of the individual in this health space?
Figure 1. The continuum of health.
(A) The health of an organism exists on a continuum (y-axis) and can shift along it at any time under homeostatic conditions (vigor). When challenged with a hostile environment, such as a pathogen, the organism can maintain health in response to the insult (maintenance), become sick but then recover (resilience) or become sick and die. (B-C). The LD50 infection approached used by Sanchez et al. to define mechanisms of health with stochastic models. (B) LD50 death curve of C. rodentium infected mice. (C) The health trajectory of survivors and those that succumbed to the LD50 dose of C. rodentium. The health trajectory can be used to define individuals that fall into each outcome group prior to death that can be subjected to systems analyses to elucidate health mechanisms and methods of intervention to promote health. (B-C) Data adapted from (Sanchez et al., 2018).
The traditional definition of health, “the absence of disease or injury”, gives the impression that removing the source of the disease or injury necessarily restores health, resulting in a focus on disease pathogenesis mechanisms that shift the health trajectory to disease and the mechanisms that eradicate the insult. However, we intuitively realize that health is an active process that can decline, for example with age, and that physiological processes independent of the mechanisms of disease pathogenesis can affect the maintenance or resilience capacity of an individual. The distinction between these two ways of thinking is highlighted by considering a simple host-pathogen system. It is rare for a single type of microbe to cause disease in 100% of the hosts that it infects. Instead, pathogen virulence, the ability of a pathogen to cause disease, exists on a continuum (the y-axis), and whether a pathogen will cause disease in large part is dependent on how the host responds to the pathogen. A failure to cause disease in 100% of hosts that are infected with a microbe is almost always ascribed to the relative resistance to the infection and not the ability of some hosts to remain healthy by withstanding the presence of the pathogen. Consider the fascinating phenomenon of lethal dose 50 (LD50), which describes the dose of the pathogen that will kill 50% of a genetically identical host population that is exposed to identical environmental conditions (Figure 1B). Most would assume that at such a dose, infection would drive illness to different extents in different individuals (due, perhaps to random processes), such that all individuals would become sick, but only half would die. Surprisingly, though, this is often not the case (Sanchez et al., 2018).
We followed the health trajectories of genetically identical mice that were challenged with the LD50 dose of the diarrheal pathogen, Citrobacter rodentium (Sanchez et al., 2018) (Figure 1B and C). We made the critical observation that while the mice received the same dose of pathogen and had equivalent levels of the pathogen infecting their organs, there was a bifurcation in the health trajectories of the two outcome groups. Those that survived the LD50 dose remained healthy by a number of criteria over the course of the infection, while only those that subsequently died became sick, as one would predict, prior to death (Sanchez et al., 2018) (Figure 1B and C). What happens mechanistically to dictate whether a host will develop illness and eventually succumb to the challenge or stay healthy in the presence of the insult? While our work revealed that the two outcome groups of the LD50 differed in the maintenance of health, one can imagine for other LD50 systems the health trajectory can reveal differences in resilience. What happens mechanistically to dictate whether a host will recover to original health after developing illness or will succumb to the challenge in the presence of the insult? We understand a great deal about the disease pathogenesis mechanisms that shift the health trajectories of sick, dying individuals. However, we have relatively little understanding about the mechanisms that maintain the trajectories at maximal health in the presence of the insult, or those that control the resilience of an individual, and we have no reason to assume that the maintenance capacity or the resilience path of an organism involves the same mechanisms that drive disease.
Health mechanisms minimize the fitness costs of diseases
Natural selection functions to promote reproductive fitness. A common misconception is that health and fitness are interchangeable, which is not the case. Indeed, for many diseases, the very traits that were selected to promote fitness are the ones responsible for disease due to incompatibilities with novel environments, physiological trade-offs, and antagonistic pleiotropy, where one gene controls for multiple phenotypes that have opposing effects on organismal fitness (Corbett et al., 2018; Paaby and Rockman, 2013). For example, the transition of humans from hunter-gatherers to the post-industrial revolution and farming represents a mismatch of our physiologies with novel environments. As hunter-gatherers, humans underwent prolonged periods of fasting and daily exercise. In the modern world, humans are now more sedentary and undergo ad libitum feeding leading to obesity and metabolic syndrome (Pontzer et al., 2012). Therefore, natural selection does not work to protect from disease unless the conditions for natural selection have been met. There must be: 1) a disease that acts as a selective pressure by reducing reproductive success in a portion of a population; 2) variations in reproductive success within diseased individuals in the population; 3) variation in a trait that is heritable and that provides a fitness advantage when the selective pressure of the disease is present. There can then be selection of adaptive mechanisms that minimize the fitness costs of disease by promoting health that fall into two classes described below: defensive health mechanisms and homeostatic health mechanisms.
Defensive health promotes maintenance and resilience
Defensive health mechanisms are specialized health mechanisms that evolved to promote maintenance or resilience of the health trajectory when challenged with extrinsic threats or hostile environments that drove the evolution of our physiologies, including food scarcity, predators, infections, heat/cold stress, hypoxia/hyperoxia, dehydration and toxins (Figure 1A, 2A). Defensive health mechanisms are inducible mechanisms that enables the organism to antagonize or withstand the insult. Antagonizing the insult is mediated by avoidance and resistance strategies that enable the organism to avoid or eliminate the insult. Withstanding the insult is mediated by disease tolerance and neutralization strategies that alleviate the fitness costs by promoting maintenance of health or resilience in the presence of the insult (Figure 2A). (Hart, 2011; Raberg et al., 2009; Schneider and Ayres, 2008).
Figure 2. Physiological health strategies.
(A) In response to a hostile threat, an organism can theoretically defend itself by avoiding the threat, eradicating the threat with resistance mechanisms or withstand the presence of the threat with disease tolerance and neutralization. The context and cost of each defense will dictate which strategy an organism will employ. (B) Homeostatic control mechanisms promote vigor under normal conditions. When these fail they have the potential to cause disease due to the inability to maintain variables within the set-point. Homeostatic tolerance promotes an apparent vigor to this new internal environment. Red arrows indicate negative affect on tissue/organ/physiology.
Let’s consider a host-pathogen system to conceptualize the distinction between these defensive health strategies. Avoidance mechanisms are innate and learned behavioral mechanisms that operate to prevent a host from becoming infected with a pathogen. These behaviors include grooming and other hygienic practices, reduced food consumption, physically moving away from the threat and social avoidance to limit interactions with potentially infected individuals (Figure 2A) (Hart, 2011). Olfactory, visual, gustatory and even auditory (individuals informing others to stay away because they are sick) sensory cues trigger avoidance behaviors. In birds, males invest in secondary characteristics to provide visual cues to females about their infection status – brighter plumage negatively correlates with infection status. Females are less likely to mate with those with duller feathers and this serves as an avoidance mechanism (Trivers, 2002). In humans, many avoidance behaviors are suggested to be driven by the emotion of disgust triggered by these sensory cues (Curtis et al., 2011). Resistance strategies operate to eliminate the pathogen once it has infected the host (Figure 2A). These mechanisms are largely encoded by the immune system, which kills pathogens, and are also mediated by physical and chemical barriers including the intestinal epithelium and the acidic environment in the upper gastrointestinal tract. Damage during an infection is caused by pathogenic signals induced by the pathogen and the host response to the pathogen that act on the host at all levels, resulting in pathology and a decline in health. A host can withstand the presence of the pathogen and alleviate the fitness costs of the infection by limiting physiological damage and promoting health (Figure 2A) (Raberg et al., 2009; Schneider and Ayres, 2008). This can be done by neutralizing the pathogenic signals induced by the pathogen or the host response to the pathogen or changing the susceptibility of the host to these signals. Anti-virulence mechanisms are a neutralization strategy employed by changes in host physiology that limit pathogenic signals during infection without affecting the pathogen’s ability to infect or replicate in the host (Rao et al., 2017; Sanchez et al., 2018). By contrast, disease tolerance mechanisms limit damage during the infection by minimizing tissue susceptibility to damage cues, maintain physiological function and promote repair (Medzhitov et al., 2012; Schneider and Ayres, 2008). The critical distinction is the way in which these strategies affect the health trajectory with respect to the extrinsic insult. Avoidance and resistance mechanisms operate to promote maintenance or resilience by avoiding or eradicating the extrinsic insult. Disease tolerance and anti-virulence (neutralization) operate to promote health by withstanding the presence of the insult. For example, considering the C. rodentium LD50 experiment discussed above, an examination of the relationship between host health and pathogen fitness over the course of the infection indicates that the maintenance of health in the survivors of the LD50 can be explained by enhanced disease tolerance or anti-virulence mechanisms rather than avoidance or pathogen eradication due to enhanced resistance, because their health trajectory remained constant along the y-axis while accommodating comparable levels of the pathogen as to those found in the sick, dying group (Figure 1B and C) (Sanchez et al., 2018).
Which defense strategy an organism will employ will be largely dependent on its cost (Steppuhn A., 2008). For an infection, resistance responses will kill the infection, but the collateral damage caused by the resistance response can compromise the host. This contributes to symptoms of the infection and helps explain why resistance mechanisms can negatively shift the health trajectories of individuals, and why health of a host cannot be sustained solely by killing pathogens. Disease tolerance alleviates those costs by maintaining physiological function and health during the infection (Raberg et al., 2009; Schneider and Ayres, 2008). For some infections, there is an apparent cooperation between the host and the pathogen, mediated by disease tolerance and anti-virulence mechanisms (Rao et al., 2017; Sanchez et al., 2018; Schieber et al., 2015). This can lead to the evolution of attenuated pathogens, resulting in the persistent carriage of avirulent strains (Sanchez et al., 2018). In these contexts, it appears that cooperation with the pathogen and colonization with these new strains provides less of a cost to host fitness than mounting a resistance response to kill the pathogenic form of the infection. It is also possible that these new strains may confer a fitness benefit to the host. Indeed, cooperation with microbes via disease tolerance and neutralization mechanisms have been proposed to drive the evolution of beneficial relationships between hosts and microbes (Ayres, 2016).
Hostile environmental factors drive selection for different defensive health strategies, functionally classified as avoidance, resistance, disease tolerance, and neutralization (Figure 2A). In addition to cost, which defensive strategy will be selected by a hostile threat is dependent on the context of the threat. Pathogens are a unique extrinsic insult in that they can drive the selection for each of these strategies but other hostile environments will drive the selection for a subset of these strategies. For example, most organisms can avoid a hypoxic environment by moving to one with greater oxygenation, and therefore hypoxia drove the selection of such avoidance behaviors. An ability to withstand hypoxia, therefore, would provide an opportunity to exploit such a niche. Indeed, naked mole rats (NMRs), whose normal habitat is hypoxic subterranean burrows, employ defensive strategies other than avoidance. Instead, they employ tolerance defenses to maintain health and limit the fitness consequences of hypoxic conditions. NMRs use fructose as a substrate to fuel anaerobic glycolysis and in doing so limit tissue damage under hypoxic conditions (Park et al., 2017).
Anti-virulence defenses may appear to only make sense for infections with professional pathogens, however the resident microbiota can become pathogenic when an organism is challenged with a non-infectious external threat. Anti-virulence mechanisms protect the host from the microbiota when this partnership becomes compromised. In the sickness-induced anorexic response, the host must ensure energy demands of the microbiota are satisfied under food scarce conditions (Ayres, 2013). To do this, the host employs a fucosylation mechanism to increase the amount of fucosylated species presented on the intestinal epithelial surface on which the intestinal microbiota can forage (Pickard et al., 2014). While anti-virulence applies to protection from microbe-driven pathology, the concept of anti-virulence is similar to other defensive health mechanisms that fall into the neutralization category and that protect from other external threats such as detoxification of noxious substances (Figure 2A).
Homeostatic health promotes vigor and apparent vigor
Health under normal, unchallenged conditions is called vigor and the ability to maintain vigor over time is dependent on evolved homeostatic control mechanisms (Figure 1 and 2B). Unlike defensive health mechanisms, homeostatic control mechanisms are constantly operating to control the plasticity of an organism, facilitating their ability to adapt to fluctuations in physiology and the environment to maintain a constant internal environment. These mechanisms participate in a dynamic equilibrium, requiring sensing of a continuously changing internal variable, integration of this information into a control center that provides output information to mediate an effector response that responds appropriately to these changes to stay within the set point for a particular variable. Homeostatic control mechanisms operate at all levels of the body – molecular, cellular, tissue, organ and physiological to: 1) meet supply and demand to ensure sufficient resources are available to cells and tissues; 2) perform quality control assessment and maintenance on each physiological system; 3) perform detoxification functions to deal with the removal of waste and byproducts of normal homeostatic processes that have the potential to cause pathology; and 4) perform repair functions (Figure 2B). Because no system exists in isolation, the health of an organism is dependent on how well all the systems of the body function together to execute homeostatic control mechanisms and maintain vigor.
A failure to induce defensive health mechanisms will result in disease (loss of maintenance or resilience) caused by a hostile environment. By contrast, a decline in vigor occurs when there is a loss of the constitutive functions of homeostatic control mechanisms, leading to dysfunction or failure to sense and respond to internal variables to return to the desired set-point (Figure 2B). This is reflected in the health trajectory of the organism by a shifting towards a diseased state, due to a deficiency of a resource, the gain of a byproduct or waste that can induce pathogenic responses, or a failure to maintain the quality or operations of a system. Diseases of aging, such as cancer and neurodegenerative diseases, are due, at least in part, to a decline in homeostatic control functions. Studies of aging in experimental animals reveals an interesting phenomenon related to what is observed with LD50 studies. Genetically identical mice housed in identical environments do not die at the same rate and they exhibit differences in the kinetics of health trajectory decline, indicating differences in physiological age (health) in otherwise identical animals of the same chronological age (Turturro et al., 1999). One explanation for this variation is that there is a difference in the rate of decline of function of homeostatic control mechanisms. A second explanation is that these differences represent variations in the ability of the individuals to adapt to the new internal environment. This would involve variations in physiological compensation mediated by homeostatic tolerance mechanisms (Figure 2B). These mechanisms are largely unknown, but are conceptually like disease tolerance mechanisms that mediate defensive health against hostile environments in that they facilitate the ability of systems in the body to adapt to a new internal environment. This apparent homeostasis, would maintain vigor, or rather result in an apparent vigor when homeostatic health mechanisms decline (Figure 2B). For example, cardiac hypertrophy occurs in response to the increased resistance that the heart experiences in individuals with hypertension to maintain cardiac output (Ertl et al., 1991). Many organs in the body exhibit compensatory growth to maintain health when there is damage, dysfunction, or even removal of another organ (Cerf, 2013; Taner et al., 2015). Homeostatic tolerance mechanisms also involve physiological remodeling and limiting tissue susceptibility to damage to maintain vigor (Del Coso et al., 2009; Metivier et al., 2000; Seifter and Chang, 2017). These mechanisms have the capacity to maintain the health trajectory of an organism over time. While a failure to compensate will result in a decline in health, prolonged physiological compensation can also ultimately lead to disease. The underlying mechanisms of homeostatic tolerance will be different for each tissue depending on their ability to adapt to the disrupted internal environment, to support vital organs and vital physiologies. Similarly, sustained changes in tissue susceptibility to disruptions in intrinsic variables can also eventually cause disease and shift the health trajectory towards disease.
How can we experimentally study health mechanisms?
Phenotypes to study
Health is the integration of all systems and their interactions within a body, making it a property of the whole organism that therefore should be studied at the whole organismal level. The same clinical parameters we use to study disease can be used to study health: overall morbidity, weight, anemia, body temperature, diarrhea, grimace, anorexia, blood pressure, heart rate, respiration, confusion, social withdrawal, grooming, activity, and circulating markers of organ function. What is important is that they are quantifiable, have a broad dynamic range, and are readily measurable (Ayres and Schneider, 2012). While biologists often use survival as a readout for health in their systems, longevity does not always equate to health. Manipulation of disease pathogenesis mechanisms can increase lifespan independently of health, yielding a longer-lived organism that is morbid (Pearson-Stuttard and Gregg, 2019; Shankar-Hari and Rubenfeld, 2016; Stollman et al., 2009). Once a health phenotype is defined at the organismal level, the mechanistic basis for this health phenotype can be elucidated at the physiological, tissue, cellular, genetic and molecular levels.
When interpreting clinical phenotypes, the context in which they occur must be considered. For example, unintentional weight loss would typically be considered a sign of disease. However, in the context of acute starvation, the increased lipolysis and decreased lipogenesis, which would clinically present as weight loss, is a disease tolerance mechanism to increase the availability of energy stores until food is available. Reduced activity is often a sign of lethargy and a decline in health, yet for some organisms this serves promotes disease tolerance when challenged with hostile conditions such as cold temperatures or hypoxic conditions with an extreme example being the induction of torpor or hibernation. For infections, clinical signs and symptoms such as fever, vomiting and coughing can be signs of the defense response. In some cases, these changes in health are the result of the costs of employing a defensive heath mechanism. Thus, examining multiple parameters of health and understanding the context can reveal information about how physiological systems cooperate to maintain health and the resulting trade-offs. The engagement of defensive health mechanisms can therefore be the cause of some signs of disease, but in some cases, represent physiological changes that are part of the defense response for promoting health.
Models for phenotypic variations in health
We already use models that perturb the internal and external environments of an organism to understand the underlying mechanisms of disease pathogenesis, avoidance, and resistance defenses. We can use these same models to elucidate the mechanisms of homeostatic control, homeostatic tolerance, disease tolerance and neutralization mechanisms. Models with perturbations or variations in intrinsic properties such as genetics or the resident microbiota are intrinsic models. These can be examined alone or together with other intrinsic perturbations under homeostatic environmental conditions to reveal how these differences in the intrinsic property influences homeostatic control mechanisms and vigor. They can also be coupled with hostile extrinsic challenges to determine how that intrinsic factor controls defensive health mechanisms. Apparent stochastic models such as LD50 infections provide an exciting opportunity to reveal the mechanistic underpinnings of physiological health mechanisms (Sanchez et al., 2018). Additional layers of complexity can be added to a system by coupling such stochastic models with other perturbations to reveal how modifiers of the stochasticity shift the lethal dose. For example, how does chronic stress or acute starvation shift the LD50 curve of an infected group of an animal, and what mechanisms drive this shift in the curve? This approach can also be applied to intrinsic perturbation models for which there is not 100% penetrance of a disease phenotype. For example, genetic mouse models that develop spontaneous colitis exhibit variation in the penetrance of the disease phenotype (Keubler et al., 2015). We attribute much of this variation to differences in environmental factors, but are there other differences that act as modifiers of the mutation and enable an organism to withstand the genetic insult? Similarly, with microbiota-driven diseases for which there is not 100% penetrance, what stochastic events dictate whether an organism will be able to withstand the microbiota and remain healthy? Animals with extreme physiologies, such as hibernating or migratory animals, or those that undergo suspended animation, provide an exciting source for phenotypic variations and discovery of health mechanisms with comparative physiological approaches.
For elucidation of disease tolerance and neutralization defensive health mechanisms that protect from hostile environments, it is critical that one uses models that mimic the environments that drove the evolution of the physiology of the subject organism. We often use models that represent diseases of the modern world, employing factors such as antibiotics and hygiene. However for diseases caused by novel environments and other evolutionarily “new” external factors, while they may incur a fitness cost, our evolution may not have “caught up,” and we likely do not have adaptive mechanisms to maintain health in the presence of these threats (Corbett et al., 2018). Hostile environmental factors that drove the evolution of defensive health mechanisms can be classified based on whether the insult can co-evolve with the organism. Temperature stress and hypoxia do not co-evolve with organisms. However infectious diseases, predators, and intra-species fighting represent co-evolutionary systems where reciprocal adaptations occur, revealing important principles about health mechanisms. With host-pathogen systems, disease tolerance and anti-virulence defenses protect the host while having a neutral to positive influence on pathogen fitness. This has important implications for understanding the mechanisms of health. For example, microbes can evolve strategies to promote disease tolerance and anti-virulence mechanisms in the host (Ayres, 2016). We appreciate that members of our microbiota have the capacity to induce health in the host. We have less appreciation that pathogens also can promote disease tolerance or anti-virulence health mechanisms despite the well accepted understanding that host-pathogen systems evolve towards reduced virulence. Microbes, both beneficial and pathogenic, represent a unique opportunity to define these health mechanisms and identify methods of intervention that have the potential to be applied to diseases beyond infections (Ayres, 2016; Rao et al., 2017; Schieber et al., 2015).
Distinguishing homoeostatic from defensive health mechanisms in experimental systems
An organism with reduced vigor may also have increased susceptibility when challenged with a hostile environment. The challenge is to determine if this increased susceptibility is due to dysfunction of a specific defensive health mechanism or if the reduced vigor renders the organism generally more susceptible to hostile environments. When one is studying a defensive health mechanism, how can one determine if they are studying avoidance, resistance, disease tolerance or neutralization? Below, I discuss three methods to identify groups that differ in homeostatic or defensive health due to a perturbation in an intrinsic or extrinsic factor such as genetics, microbiota or diet: tracking health over time (Sanchez et al., 2018); reaction norms (Raberg et al., 2009; Schlichting and Pigliucci, 1998) and tracing trajectories in a health space (Torres et al., 2016). These can be used to distinguish these possibilities in experimental systems. In each case, once two groups that differ in health are identified, the groups can be used in subsequent analyses to elucidate how those perturbations control health and for elucidation of homeostatic and defensive health mechanisms.
Tracking health over time
This approach examines how health of an organism changes over time due to an intrinsic or extrinsic perturbation under homeostatic conditions or when challenged with a hostile environment, and therefore has the potential to reveal differences in homeostatic control, homeostatic tolerance or defensive health. It involves an XY plot with the health continuum on the y-axis with maximal health at the top and minimal health at the bottom (as in Figure 1A). The x-axis represents time. Each point on the curve represents the quantification of a health parameter at that time point. The line generated represents the health trajectory of the organism. The shape of the curve together with the context (homeostatic or hostile conditions), indicates if the differences in health are due to vigor, apparent vigor, maintenance or resilience Two organism groups that have overlapping health curves would indicate no differences in health. For vigor, apparent vigor and maintenance, a heath trajectory shifted to the left with respect to another would have reduced health and a health trajectory shifted to the right would have greater health compared to the other group. For resilience, shifts right indicate less resilient and shifts left indicate more resilient. (Figure 3A, B). There are an infinite number of possible curve shapes but will fall into the classes described in Figure 3. The health parameter used to make these plots should be relevant to the context being studied such as overall morbidity, body temperature, and weight loss, and can be an individual parameter or an integration of all health parameters quantified – whichever is most appropriate for the system being studied.
Figure 3. Tracking health over time.
Hypothetical maintenance and resilience health curves. Shifts to the left and right indicate differences in (A) maintenance and (B) resilience respectively. (C) Examination of the health curve and insult levels or burdens differentiate health differences due to antagonism or the ability to withstand the insult. For an example curve of differences in maintenance due to differences in disease tolerance or neutralization see Figure 1C.
For groups exposed to a hostile environment, once a group has been identified to have a difference in health, the relationship between heath and insult levels/burden over time are examined to determine if the differences in defensive health are due to differences in avoidance, resistance, disease tolerance or neutralization. The context must also be considered. For hostile environments, when avoidance mechanisms are ruled out and resistance is not a possibility, and for insults that can be controlled by insult level (dose and exposure time) such as hypoxia, acute starvation, and temperature stress, measuring how health changes over time is sufficient to reveal the contributions of a variable to disease tolerance/neutralization mechanisms against that threat. For hostile environments where insult burden (amount of insult measured in the organism) can vary, (e.g., pathogens), burdens should be measured over time (not at only a single time point) to avoid misinterpretation of the data (Ayres, 2017) (Figure 3C). Differences in health curves without differences in burdens would indicate that the two groups differ in disease tolerance or neutralization (Figure 1C, 3C).
Groups that differ in health and burdens would indicate differences in resistance (or avoidance if a possibility with the experimental system) (Figure 3C). Sanchez et al. applied this approach to their LD50 Citrobacter model to define a novel anti-virulence mechanism (Figure 1B, C and (Sanchez et al., 2018). Stochastic models that vary in resilience can also be used (Figure 3C) but can provide a challenge because the less resilient group may succumb to the disease prior to the recovery of the healthy organisms. In this case, a predictive biomarker or longitudinal monitoring of readily accessible parameters before the bifurcation in the health trajectories occurs. This approach is limited in identifying differences in disease tolerance/neutralization between two groups if they differ in vigor. A special case where this is not true would be if the vigor defect results in enhanced disease tolerance/neutralization in response to a hostile insult. Then this would suggest there are trade-offs between vigor and the maintenance or resilience capacity of an organism for certain variables.
This same approach can be used to determine how intrinsic or extrinsic perturbations affect homeostatic control mechanisms. This is a simple approach as it only requires measuring relevant health parameters over time and does not require measuring insult levels or burdens (as these are homeostatic conditions rather than hostile). Manipulations that result in the enhancement of an aspect of health will likely come at a cost to some other aspect of health. Thus, multiple parameters of health and analysis at different stages of life of an organism will reveal those trade-offs. For the identification of homeostatic tolerance mechanisms, a similar strategy to that performed for disease tolerance/neutralization can be done, where the effects of health for an intrinsic or extrinsic factor is compared to the level of the perturbation that results because of the failure of homeostatic control mechanisms.
Reaction norms
This approach is useful to define differences in homeostatic control, homeostatic tolerance, disease tolerance and neutralization mechanisms. Evolutionary biologists use a conceptual and statistical framework based on the reaction norm concept to measure phenotypic plasticity of a single genotype in response to changes in environmental conditions (Schlichting and Pigliucci, 1998). It is a tool that maps all the phenotypes that a genotype can yield across a range of environmental conditions within an XY plot and thus represents the phenotypes of a genotype as a function of its environment. Reaction norms can be used to define groups that vary in defensive health when exposed to a range of extrinsic insults due to differences in an intrinsic property or when exposed to different extrinsic variables even when there are differences in vigor, which would be represented as health when not exposed to the insult (Raberg et al., 2007). On an XY plot, health could be represented by the y-axis. On the x-axis insult level (dose/duration of the insult) or insult burden (the level of insult detected in the organism) is plotted. Each point on the curve represents the health phenotype at that insult level or burden. The y-intercept represents vigor. The line that is generated represents the intrinsic or extrinsic variable for that group. The slope of the line indicates how health changes over the range of extrinsic insult. The shallower the slope, the more health is maintained over the range of extrinsic insult. The steeper the slope, the more health changes over the range of extrinsic insult. The benefit of using a reaction norm is that it tests for the interaction between an intrinsic or extrinsic variable and the extrinsic insult. Thus, even if two different groups differ in vigor, one can conclude if any differences in health are due to general defects in vigor or if a variable specifically mediates defensive health against an extrinsic threat by examining the slope of the curves (Raberg et al., 2009). Reaction norms also allows for the identification of differences in health in response to an extrinsic insult without assumptions about the shape of the relationship between health and insult levels for an organism with an intrinsic or extrinsic variable (Figure 4) (Raberg et al., 2009).
Figure 4. Reaction norms to define differences in health.
Hypothetical infection of host group green and pink. Vigor indicated as the y-intercept. (A) Host groups differ in vigor indicating differences in homeostatic control mechanisms and respond the same to infection indicating no differences in defensive health mechanisms. (B) Same vigor and green host has shallower slope indicating greater maintenance or resilience in response to infection due to enhanced disease tolerance or neutralization. (C) Green host has greater vigor indicating differences in homeostatic control mechanisms. The shallower slope of Green indicates it also has greater disease tolerance or neutralization mechanisms because it can maintain health over a range of pathogen better than pink. (D) Green host has greater health because it is more resistant (increased health and less pathogen burden) to the pathogen than pink host because the curve is shifted along the x-axis.
Reaction norms plot the summary of health for an individual at each insult level and some decisions about how health is quantified must be considered. Health can be represented by a single time point, the maximum or minimum health exhibited by an organism within a defined timespan, or the integral of the health curve over time. For representing insult, for extrinsic threats that can be controlled by dose and exposure time, the dose amount or exposure time is sufficient for representation on the XY plot. For extrinsic threats that can vary over time, such as pathogens, one can represent levels by the dose of the pathogen, or use maximal pathogen burden or the rate of pathogen growth or the integral of pathogen levels over the course of the infection. With reaction norms, it can be difficult to conclude if variations in defensive health are due to difference in maintenance of health or resilience because they plot the summary of the health curve. Ideally one would use tracking health over time in combination with reaction norms to define groups that vary in defensive health that can be used in subsequent analyses for the and elucidation of defensive health mechanisms. The distinction between reaction norms and tracking health over time is that reaction norms they allow for the identification of disease tolerance/neutralization differences between two groups even if vigor is different, plot the summary of the infection and requires testing multiple doses of insult.
Charting physiological paths in a health space
The reason an individual dies from a disease, regardless of the insult, is because they ultimately suffer from irreversible cessation of heartbeat and breathing (cardiopulmonary death) or irreversible cessation of brain functions (brain death). Despite this, the path an individual takes through a health space towards death is somewhat like a pinball machine. In certain cases, individuals challenged with the same insult can take different physiological paths that lead to the same cause of death. Two different individuals suffering from the same insult can also take different paths that cause a different cause of death. A simple way to think about this is to visualize the health continuum as a health space with organs, tissues and physiological processes positioned within this space (Figure 5A). The position of each within this space correlates with how severe the consequences would be to health if that system was damaged. Vital organs and vital process are towards the bottom at minimal health. All others are positioned further up along the health trajectory. Each organ, tissue, processes can be viewed as a node and movement down this space towards disease due to an intrinsic or extrinsic insult depends on: 1) damage to the node and 2) the severity of the consequences of the damage to that node for the organism. Individuals can enter into this space at different nodes and shift to other organ systems with similar consequences to health before moving down the path. The further down the space they begin their disease path, the more unhealthy they are. For example, with malaria, a patient can develop renal failure, a consequence of which is multi-organ failure such as heart failure. A patient can develop liver failure that progresses to kidney failure, or cerebral edema. Other patients can develop cerebral malaria (Figure 5B). However, despite the apparent complexities of disease courses, there are a finite number of physiological paths one can take that leads to death. Even seemingly disparate causes of diseases ultimately converge on similar physiological courses leading to the same pathology and cause of death. This implies that there is then a finite number of health paths that one can take for vigor, apparent vigor, maintenance and resilience in response to perturbations and insults. At each node, health mechanisms can prevent the movement of the individual further down the continuum and shift them back up towards maximal health by: 1) enabling the node to adapt to withstand the damage it experiences so that it’s function is not affected and 2) if function is affected, enabling the other systems to adapt through physiological compensation to protect from the consequences (Figure 5C).
Figure 5. Charting physiological paths in a health space.
(A-C)The health continuum can be visualized as a health space with hierarchies of organs/physiological systems that are ranked based on the consequences of their damage or dysfunction on health. Each system represents a node with defined health mechanisms to promote health of that system and shift the path back up towards health in the health space. (B) Hypothetical disease paths for malaria infection and (C) methods of intervention at disease nodes to shift the trajectory back towards health. (D-E) From Schneider and colleagues (Torres et al., 2016), (D) the hypothetical path a resilient individual takes through the health space when sick and (E) topological network map of the transcriptomes of circulating immune cells of surviving and dying mice infected with malaria. Mice that die (red) do not loop back to the original health state. Resilient mice (blue) do. Curves show where physiologically the two fate outcomes differ based on transcriptome.
A significant limitation of the models that track health over time and reaction norms is that they summarize the path an individual takes to disease and death by plotting a summary or integration of quantitative health measurements. Furthermore, for patients, longitudinal information regarding the disease course of an individual is typically not available. Clinicians only have data points that begin when the patient comes into to the clinic. What is important for mechanistic studies of health and ultimately translating the concepts presented here to practice is to: 1) determine the disease path of an individual when challenged with an intrinsic or extrinsic insult; 2) identify where the individual is on their disease path with readily accessible markers; 3) know the health path of an individual that remains healthy in response to the intrinsic or extrinsic insult; and 4) understand the mechanisms that can shift the disease path to the health path based on where they are on their disease course. For most diseases, we know the different disease paths and the mechanisms that shift an individual further along their disease course towards death. We now need to focus on defining the latter two items. Schneider and colleagues developed an elegant system to do just that (Louie et al., 2016; Schneider, 2011; Torres et al., 2016). They imagined the path a resilient individual takes through a health space. The map has XY coordinates and point 0,0 represents maximal health. As an individual gets sick, they move away from this point and as they recover, they loop back to return to this point (Figure 5D). Using the infection Plasmodium chaubaudi in mice, they found that by plotting quantitative measurements of multiple parameters of health and using topological data analysis, non-dimensional networks can be generated to visualize the underlying shape of the curve. Mice that recovered from the infection had distinct curves from mice that succumb to the infection (Figure 5E) that correlated with differences in specific health parameters. These curves are useful for identifying the current location of the organism in the health space and predict the future route of the disease course. Once the path of an individual is constructed and the path differences between healthy and sick individuals are established, one can introduce intrinsic and extrinsic perturbations into the system to determine mechanistically how they affect the different phases of the infected and use this tool to discover methods of interventions to shift the disease paths at different phases of the disease course. Because cross-sectional data can be used, this approach is the most amenable approach for applying to humans. Health spaces can also be used to define differences in vigor, apparent vigor and maintenance but the curves will look different – they will not shift into sick and then recover but rather will maintain health. Therefore, these curves will shift into different dimensions into this space compared to sick/dying individuals that will shift down into the disease area.
The mechanisms of physiological health
We have an excellent understanding of homeostatic control mechanisms, but less so of what contributes to the variations in vigor among individuals. We appreciate that intrinsic factors (including genetics, the resident microbiota, and lifestyle, such as activity and diet) will influence overall homeostasis and vigor. However, we generally lack a mechanistic understanding of how these factors influence homeostatic control mechanisms. For homeostatic tolerance, we know very little mechanistically about the ways in which the body can physiologically compensate to facilitate adaption to new internal environments when homeostasis control mechanisms fail. What causes variations in homeostatic tolerance mechanisms is unknown. While we also have an excellent understanding of the disease tolerance mechanisms that have evolved to mediate defense in some specific, hostile environments, the host-pathogen field over the past decade has revealed that we have barely tapped into the mechanisms that enable an organism to remain healthy when faced with hostile environments. These areas will be a focus in the upcoming decades of biological research, providing a better mechanistic understanding of physiological health for improved disease treatment. Below I provide a mechanistic framework to conceptualize how these mechanisms function and how they relate to homeostatic control mechanisms.
While homeostatic control, homeostatic tolerance, disease tolerance and neutralization health mechanisms are protective against distinct insults, there are important conceptual and mechanistic parallels between these health mechanisms. Homeostatic control mechanisms facilitate the plasticity of an organism that is needed to adapt to the internal and external fluctuations present under homeostatic conditions to maintain vigor. Homeostatic tolerance facilitates the physiological adaptation of an organism to a new internal state caused by homeostatic control mechanisms failing. Similarly, disease tolerance and neutralization facilitate the adapation of an organism to a new state caused by the presence of an extrinsic insult (Figure 2). Each facilitates physiological adaptations that promote health and prevent pathology and disease. The same physiologies that are vulnerable to dysfunctions in homeostatic control mechanisms are vulnerable to failure of homeostatic tolerance, disease tolerance, and neutralization, and they therefore mediate adaptation of the same systems to protect the same physiologies. For vigor, homeostatic control mechanisms prioritize the protection of vital organs and vital physiological processes (Figure 5A–C). The same systems are prioritized to protect from damage with homeostatic tolerance, disease tolerance and neutralization, and the capacity of each tissue to adapt for vigor will be similar for apparent vigor, maintenance and resilience. This conceptually means that homeostatic tolerance, disease tolerance, and neutralization mechanisms perform the same functions as homeostatic control mechanisms to meet the supply and demand of systems, to perform quality control and maintenance of systems, detoxify waste products or neutralize pathogenic signals, and perform repair functions. While the mechanisms are distinct, homeostatic tolerance, disease tolerance, and neutralization mechanisms likely evolved from homeostatic control mechanisms.
One important distinction is the way different health mechanisms are activated. Homeostatic control mechanisms are constantly operating for continuous surveillance of internal variables. By contrast, homeostatic tolerance, disease tolerance, and neutralization are inducible health mechanisms. This is, in part, due to the cost of these mechanisms. The inducibility reduces the cost by ensuring they are active only when the relevant insult is present. It is also a product of their selection and defines the specificity for protection against specific perturbations and insults. Homeostatic tolerance, disease tolerance, and neutralization can be induced by direct sensing of the perturbation or insult, sensing the loss of a signal, or sensing of a normal homeostatic variable that reaches levels outside the homeostatic set-point. There can also be combinatorial sensing of multiple signals, for example, direct sensing of the insult plus some physiological parameter that leads to induction of the health mechanism (Figure 2).
Maintaining vigor requires coordination of interactions between all physiological systems in the body (cardiovascular, digestive, immune, respiratory, renal etc). Central to this is the neuroendocrine system, with the hypothalamus coordinating these interactions to control the following physiological variables: growth and development, macro/micro-nutrient and vitamins, socialization, thermoregulation, energy balance, oxygenation, detoxification, acid-base balance, and osmoregulation (Figure 6). The control of each of these variables is dependent on homeostatic control mechanisms that operate at each level (molecular, cellular, tissue, organ, physiological), each contributing to homeostasis at the next level which ultimately translates to vigor at the organismal level. For example, at the molecular and cellular levels, some categories would be cellular stress responses, protein repair and mitochondrial turnover. At the tissue level maintaining proper cell number through generation and clearance as well as tissue repair mechanisms will contribute to vigor. Homeostatic tolerance, disease tolerance and neutralization mechanisms will also operate at each of these levels, with the neuroendocrine system and derivatives of homeostatic control mechanisms functioning to promote apparent vigor, maintenance and resilience of an organism by controlling the same physiological variables (Figure 6, 7). Here I provide a framework to begin to conceptualize the mechanisms of health based on the classes of physiological variables. While each category can be refined further, and while there is overlap in some systems and responses for these categories, they serve as an important framework for predicting the health mechanisms engaged and ultimately how their function promotes health of the organism. Due to space constraints, I will only discuss examples for a subset of these categories to establish the principles. I largely confine my discussion to examples that promote maintenance of health via disease tolerance, neutralization, and homeostatic tolerance mechanisms to focus the discussion and make it easier to draw parallels and distinctions.
Figure 6. Categories of physiological variables for health.
The hypothalamus coordinates interactions between all physiological systems in the body to control physiological variables including growth and development, macro/micro-nutrient and vitamins, socialization, thermoregulation, energy balance, oxygenation, detoxification, acid-base balance, and osmoregulation. The control of each of these variables is dependent on homeostatic control mechanisms that operate at each level (molecular, cellular, tissue, organ, physiological), each contributing to homeostasis at the next level which ultimately translates to vigor at the organismal level. Mechanisms have evolved to regulate each of these variables to promote disease tolerance, neutralization and homeostatic tolerance. For some, multiple systems will contribute to the regulation of a variable. For example, appetite regulation will also contribute to nutrient homeostasis and osmoregulation contributes to increased socialization.
Figure 7. Physiological mechanisms of health.
Representative disease tolerance, neutralization and homeostatic tolerance mechanisms that fall into the classes of physiological variables that are controlled to maintain health under homeostatic and hostile conditions. A) Derivatives of growth and development homeostatic control mechanisms promote health during challenge. A disease tolerance mechanism has evolved to maintain skeletal muscle mass during infections. When homeostatic control mechanisms decline, cardiac hypertrophy occurs to increase cardiac output and function. B) Derivatives of glucose homeostatic control mechanisms promote health during challenge. Acute insulin resistance induced by iron sequestration in fat causes allocation of glucose to intestine for gut microbes to forage on suppressing their virulence and thus acting as an anti-virulence mechanism. During chronic hyperglycemia and insulin resistance, beta cell compensation including increased insulin secretion and growth increases glucose uptake to maintain glucose levels. C) Derivatives of socialization homeostatic control mechanisms promote health during challenge. A disease tolerance mechanism to maintain health during hostile environments involves increasing socialization behavior in rats to increase the likelihood they will drink water at common water sources. In humans, maintenance of socialization may promote homeostatic tolerance when there is a decline in homeostatic control mechanisms leading to accumulation of amyloid β. D) Derivatives of thermoregulatory homeostatic control mechanisms promote health during challenge. During infection, a raise in the thermal set point promotes the febrile response that acts as an anti-virulence mechanism to limit pro-inflammatory signals that can cause pathology. During hibernation induced by nutrient scarcity, the thermal set point is reduced to facilitate hypothermia that promotes disease tolerance by protecting from oxidative and ischemic injury. E) Derivatives of energy balance homeostatic control mechanisms promote health during challenge. Bacterial inflammation induces sickness-induced anorexia to promote ketosis and disease tolerance by protecting the brain from oxidative stress. Salmonella has evolved an effector, SlrP, to manipulate the gut-brain axis and inhibit the anorexic response to dampens its virulence. F) Derivatives of oxygenation homeostatic control mechanisms promote health during challenge. Treatment with erythropoietin during Trypanosome infections promotes RBC release, preventing anemia and promoting disease tolerance. Dysfunctional erythropoiesis leads to physiological compensation including redistribution of blood flow and vascular resistance. G) Derivatives of detoxification homeostatic control mechanisms promote health during challenge. Detxification of HO-1 generated during malaria infection promotes anti-virulence defenses. CO, a byproduct of HO-1 detoxification acts on smooth muscle of the gut to increase motility and promote disease tolerance during sepsis.
Growth and development
While homeostasis functions to maintain a constant internal environment, an organism itself must change at different life stages during maturation and aging. The normal growth and development of an organism during these critical stages is a quantitative indicator of vigor (Rauber, 1990), and thus are intimately connected to homeostasis, and are largely controlled by the endocrine system and homeostatic control mechanisms. The growth hormone (GH)-insulin like growth factor −1 (IGF-1) axis is a critical regulator of somatic growth and development. Secretion of GH by the pituitary gland is induced by the hypothalamic peptides, growth hormone releasing hormone (GHRH) and somatostatin (SS), as well as the peptide ghrelin that is released by the stomach. GH induces the release of IGF-1 from tissues, which acts on tissues to promote the linear growth effects of GH. Liver IGF-1 is the predominant mediator of somatic growth, however, autocrine effects of IGF-1 on adipose tissue also has an important role in regulating circulating IGF-1 levels and somatic growth (Kloting et al., 2008).
Derivatives of this axis have evolved to promote disease tolerance to hostile environments and for homeostatic tolerance. During infections, skeletal muscle wasting is induced by proinflammatory cytokines acting directly on muscle and indirectly through induction of the sickness-induced anorexic response. The inflammatory response also causes changes in liver IGF-1 and a decline in circulating IGF-1 levels. In mouse infection models, microbial recognition by the innate immune system in adipose tissue induces IGF-1 production and release by fat deposits into circulation, sustaining the levels of IGF-1 during infection. This maintenance of circulating IGF-1 is necessary to protect from infection-induced cachexia and death, without affecting the inflammatory response or pathogen burdens, thus promoting disease tolerance (Schieber et al., 2015). The maintenance of skeletal muscle mass by IGF-1 is specific to the inflammatory state, because neutralization of IGF-1 action in uninfected, healthy animals does not affect muscle mass (Schieber et al., 2015). An inability to maintain vascular homeostasis leads to reduced blood flow to the heart that can cause myocardial infarction. In the heart, IGF-1 promotes cardiac hypertrophy and GH causes increased prevalence of the myosin V3 isoform found in cardiac hypertrophy (Castellano et al., 2009). Administration of IGF-1 to rats induces cardiac hypertrophy and in animals with evolving myocardial infarction, additional hypertrophy induced by IGF-1 shows improved cardiac function, leading to an apparent vigor, suggesting a role for homeostatic tolerance (Duerr et al., 1995). In both cases, the actions of IGF-1 sustain the health trajectory to promote maintenance or apparent vigor, respectively (Figure 6, 7a).
Macro/micro-nutrients and vitamins
The levels of macro- and micro-nutrients as well as vitamins are tightly regulated by homeostatic control mechanisms. Homeostatic tolerance, disease tolerance and neutralization mechanisms have evolved from many of these mechanisms to regulate nutrient and vitamin variables during challenge to promote an apparent vigor, maintenance and resilience (Figure 6, 7b). Here, I focus on glucose regulation. Glucose is the main energy source in most mammals, and its maintenance within the homeostatic range is essential for vigor. The regulation of blood glucose under homeostatic conditions is mediated by a complex interplay between glucose sensing mechanisms and effector systems at multiple levels of the body. When blood glucose levels are high, the beta cells of the pancreas release insulin into the bloodstream. This induces glycogenesis to convert glucose into glycogen and acts on multiple cells in the body to induce the transport of glucose via glucose transporters into cells, decreasing blood glucose. A failure in this homeostatic control mechanism can lead to chronic insulin resistance, where the body produces insulin but cells no longer respond appropriately, causing hyperglycemia that can lead to diabete.
Beta cells are dynamic and adapt to the physiological state of the organism (Cerf, 2013). This adaptability serves as an important homeostatic tolerance mechanism when sustained hyperglycemia occurs due to insulin resistance (IR), enabling beta cells to increase the amount of insulin they synthesize and secrete. Beta cell hypertrophy also occurs, resulting in increased synthesis and secretion of insulin during the IR state (Cerf, 2013). This beta cell compensation protects from progression to diabetes (Blandino-Rosano et al., 2012) sustaining the health trajectory, supporting the notion that this is a homeostatic tolerance mechanism mediating apparent vigor (Figure 7b).
Acute IR occurs in response to hostile insults, including acute starvation and infections, suggesting that acute IR (but not chronic) may be an adaptive response to hostile environments. One possible explanation is that acute IR may serve as a switch between survival and growth/development (Koshiyama, 2012). Another hypothesis is that acute IR may serve as a mechanism to increase glucose availability to prioritized physiological systems during insult. In agreement with this, in mice infected with C. rodentium, acute IR is induced by adipose tissue sequestration of iron during the infection. The systemic IR causes a change in intestinal physiology to reduce the amount of glucose absorbed from the gut, increasing the amount of luminal glucose in the intestine on which C. rodentium foraged. Glucose, as the carbon source for this pathogen, suppresses the induction of the virulence program without affecting the ability of the pathogen to colonize and replicate in the host, indicating that IR in the context of this infection serves as an anti-virulence mechanism by promoting maintenance of health and sustaining the healthy trajectory (Sanchez et al., 2018) (Figure 7b). Micronutrient regulation, including that of iron, is an additional homeostatic control mechanism. Therefore, derivatives of different homeostatic processes can synergize to enable a host to withstand a hostile threat. We appreciate that the brain, heart, and lung, are prioritized tissues, because damage to them can ultimately lead to death. These findings, along with the fucosylation response to starvation, discussed above (Pickard et al., 2014), support the idea that in hostile environments, “feeding” the resident microbiota may also be a priority to the host to maintain health via anti-virulence.
While the above examples involve IR, all aspects of glucose physiology are affected by hostile environments and derivatives of homeostatic control mechanisms that control additional aspects of glucose homeostasis are likely important for homeostatic tolerance, disease tolerance, and neutralization health mechanisms (Figure 6, 7b). My discussion focused on glucose regulation, however the same principles will apply to micronutrients and vitamins. For example, the insulin resistance in the C. rodentium example above was due to the sequestration of iron to white adipose tissue. In addition to maintaining levels of nutrients and vitamins, changes in their utilization, their allocation within the body and their ability to serve as signaling molecules will be important evolved mechanistic strategies to promote disease tolerance, neutralization and homeostatic tolerance in an organism (Figure 6).
Socialization
The effector response for some homeostatic control mechanisms includes a behavioral component to return the variable to the set-point (i.e. feeding, thermoregulation, osmoregulation). Most mammals are group-living, social animals and physiological health is controlled by social interactions (Umberson and Montez, 2010), suggesting that the regulation of frequency and quality of socialization behavior may serve as a homeostatic regulatory mechanism that operates on a different time-scale than other homeostatic regulatory mechanisms. This idea has been described as “social allostasis” that facilitates physiological adaptation to fluctuations in social and physical environments of an organism through modification of social behaviors (Schulkin, 2011). Homeostatic control mechanisms that regulate socialization are mediated in part by the hypothalamic-oxytocin system. Oxytocin neurons of the paraventricular nucleus of the hypothalamus project to regions of the brain to release oxytocin to modulate neuronal activity and social behavior (Young, 2009).
Derivatives of this axis have evolved to promote disease tolerance to hostile environments and for homeostatic tolerance (Figure 6, 7C). High temperature hostile environments can lead to dehydration and the hyperosmolar condition, i.e., hypernatremia involving elevated serum sodium levels due to decreased total body water content. Part of the defensive health response involves osmoreceptors in the hypothalamus that sense changes in osmolarity, leading to release of vasopressin from the pituitary to promote thirst, in addition to reabsorption of water in the kidney to promote disease tolerance. While driving an elevated thirst response, hypernatremia in rats also leads to greater social behavior when challenged with a novel intruder. This increased socialization in rats is associated with increased circulating levels of oxytocin, which is similar to vasopressin, and c-Fos expression in PVN oxytocin neurons, and is proposed to promote social behavior for animals to satisfy thirst at communal water sources (Krause et al., 2011). Thus, increased socialization in this example could be viewed as a defensive health mechanism to promote disease tolerance and maintenance of health trajectory (Figure 7C).
Social-empathetic behavior is induced in mammals when members of their group are sick or injured, encouraging the healthy individuals to increase socialization frequency and duration with the sick individual (Hart, 2011). This may be beneficial during infections, where lethargy is induced, and can reduce social interactions sought out by the sick individual. This idea is supported by the well-established phenomenon that social isolation negatively affects the maintenance and resilience capacity of an individual when challenged with an infection (Kappeler et al., 2015). Oxytocin is proposed to induce such social-empathetic behaviors (Geng et al., 2018). Derivatives of the hypothalamic-oxytocin homeostatic control mechanisms may be induced in healthy individuals when a group member is sick to mediate social interactions with the sick individual when they are not ambulatory. The idea that uninfected animals cooperate with infected animals during infection is supported by social immunity studies in social insects (Cremer, 2019).
Social interactions may be important for homeostatic tolerance mechanisms. During aging, the homeostatic control mechanisms that act at the cellular level and are necessary for clearance of protein accumulation in neurons declines. Accumulation of these aggregates causes neuronal death and is a driver of neurodegenerative diseases. In humans, individuals with similar levels of accumulation of the Alzheimer’s associated protein, amyloid-β, differed in their socialization. After three years, those that socialized maintained their mental health trajectory, while those who failed to socialize suffered mental health decline and a drop in their mental health trajectory (Biddle et al., 2019). While correlative, if socialization was indeed responsible for maintaining mental health, it would be acting as a homeostatic tolerance mechanism, mediating adaptation to the new intrinsic state caused by a decline in homeostatic control mechanisms (Figure 7C).
Thermoregulation
The core body temperature set-point and central control of body temperature in homothermic animals is mediated by the hypothalamus. Temperature receptors located in the hypothalamus sense changes in the core body temperature, resulting in the induction of physiological responses to balance heat produced and heat loss to maintain the body temperature within this desired set-point. These responses include vasodilation, vasoconstriction, and hormonal responses to change metabolism. In hostile environments of extreme temperature, infections and food scarcity, derivatives of the homeostatic control mechanisms that are important for thermoregulation are important for maintenance of health (Figure 6, 7D).
In extreme cold exposure, disease tolerance is mediated by the induction of shivering in muscles, as well as brown adipose tissue (BAT) thermogenesis and browning of white adipose tissue. In extreme heat conditions, sweating or increased respiration changes are induced in different mammals to promote disease tolerance. During infections, the febrile response is induced in homothermic animals upon recognition of pathogen-associated molecular patterns (PAMPs) on immune cells as well as multiple cell types in the brain, including endothelial cells, microglia, and neurons. Cytokines produced in response to recognition of PAMPs act via prostaglandins on the hypothalamus to change the thermoregulatory set-point, such that what was thermal neutral is now cold to the organism. The organism reaches a new thermal equilibrium by driving BAT thermogenesis in response to norepinephrine release. This increases the metabolic rate to induce shivering, vasoconstriction to reduce heat loss, as well as behavioral modifications, including seeking warmer environments and curling up to minimize heat loss (Schieber and Ayres, 2016). While fever can promote resistance responses to infections, it also promotes anti-virulence defenses by neutralizing the host inflammatory response to infection and preventing tissue damage via the induction of heat shock responses (Schieber and Ayres, 2016). Heat shock factor 1 (HSF-1) is activated during febrile responses in mammals. Mice deficient for HSF-1 are more susceptible to systemic infection with Listeria monocytogenes without a change in pathogen burdens. Instead, Hsf1−/− mice have elevated levels of TNFα and IFN-γ compared to infected wild type animals and neutralization of the proinflammatory cytokines is sufficient to protect from infection in HSF-1-deficient mice (Murapa et al., 2011). In a variety of sepsis models, Hsf1−/− mice exacerbate proinflammatory responses, and this is associated with impaired cardiac contraction and relaxation (Barber et al., 2014) (Figure 7D).
Hibernation represents another extreme example of thermoregulation where the temperature set-point is reduced, in some cases so low that the animal reaches ambient temperatures. In hibernating animals, metabolism and heart rate are reduced and BAT deposits are increased prior to entering hibernation to sustain body temperature during hibernation (Carey et al., 2003). The hibernating state can be thought of as a disease tolerance mechanism as it allows the animal to adapt to a hostile environment of nutrient scarcity and mediates tissue protective effects by lowering the metabolic demands of tissues, lowering the need for oxygen and nutrition and protecting against oxidative stress damage (Schieber and Ayres, 2016). Indeed, hibernation and hypothermic states protect from ischemic injury (Schieber and Ayres, 2016). In the L. monocytogenes-induced sepsis and fever, the increased temperature set-point promotes maintenance of the health trajectory, while in hibernation, a reduction in the temperature set-point promotes maintenance during conditions of food scarcity. The induction of the thermogenic responses and mechanisms of action are distinct for each hostile environment but originated from the homeostatic control mechanisms for thermoregulation (Figure 7D).
Energy balance
Homeostatic control of energy balance involves the regulation of the input energy through the regulation of feeding behavior, the regulation of energy flow or energy expenditure, as well as energy storage and allocation to systems in the body. Homeostatic control mechanisms regulate all aspects of energy balance in an organism, and homeostatic tolerance, disease tolerance, and neutralization mechanisms have evolved from many of them (Figure 6). Here, I focus on energy intake. Leptin is a peripheral signal that regulates body weight homeostasis in large part through its regulation of appetite. After a meal, leptin is secreted by the white adipose tissue and signals to the hypothalamus to suppress hunger by inhibiting neuropeptide Y (NPY) and agouti-related peptide (AgRP), which induce feeding, and by promoting the synthesis of α-melanocyte-stimulating hormone (α-MSH), which suppresses hunger. Ghrelin, synthesized in the stomach during periods of food deprivation, is a peripheral signal that promotes feeding behavior. Ghrelin acts on the hypothalamus to induce hunger by activating NPY/AgRP neurons that express ghrelin receptors (Sohn et al., 2013).
Derivatives of homeostatic controls mechanisms that control feeding promote disease tolerance. During infections, peripheral activation of the pro-inflammatory cytokine IL-1β, in response to innate immune sensing of microbial PAMPs, is sensed by the vagus nerve and leads to an anorexic signature in the hypothalamus that causes a reduction in food consumption by the host (Pavlov and Tracey, 2012). While sickness-induced anorexia can potentially serve as an avoidance mechanism during infections (for orally acquired pathogens) and resistance (Murray and Murray, 1979), it is also an adaptive host defensive health mechanism that can optimize host defense against infection by mediating disease tolerance (Ayres and Schneider, 2009; Wang et al., 2016). In mice, injection with lipopolysaccharide (LPS) mimics an inflammatory state caused by systemic bacterial infections and caused sickness-induced anorexia. The fasted state caused by anorexia increases the amount of ketosis and ketone bodies available for the brain to metabolize during bacterial inflammation. Brain utilization of ketone bodies rather than glucose as an energy source during bacterial inflammation promotes disease tolerance by protecting the brain from reactive oxygen-induced neuronal dysfunction, seizing, and death (Wang et al., 2016) (Figure 7E). However, in the context of an oral Salmonella infection, the anorexic response promotes virulence of the pathogen, increasing disease and death of the host at a cost to pathogen transmission to new hosts. Salmonella evolved an effector called SlrP, that antagonizes IL-1β maturation in the intestine, preventing IL1β-mediated signaling along the vagus nerve that induces the anorexic response. This promotes the fitness of the pathogen through the maintenance of host health (Rao et al., 2017). In the case of LPS, anorexia promotes the maintenance of health by promoting disease tolerance, but in the case of Salmonella, antagonizing the sickness-induced anorexia response serves as an anti-virulence mechanism to promote the maintenance of health. While simplified models such as LPS can be useful for understanding the host response to an infection, they are limiting if we want to understand how a host response may affect the health trajectory and outcome of an infection when co-adaptation is a possibility (Figure 7E).
Oxygenation
The maintenance of vigor requires the homeostatic control mechanisms that coordinate the complex interactions between multiple systems in the body including lungs, cardiovascular system, kidneys, and bone marrow to allow for adequate oxygen to be delivered to each system in the body. Derivatives of each aspect of these homeostatic control mechanisms have evolved to promote disease tolerance and homeostatic tolerance (Figure 6). For example, erythropoietin is a hormone that is secreted by the kidney to stimulate red blood cell production in the bone marrow. The kidney is constantly secreting low levels of this hormone to replenish the supply of red blood cells due to the normal turnover of aged and damaged RBCs. When challenged with hostile environments that have inadequate oxygen levels, the hypoxic conditions are sensed by receptors in the kidney, triggering the release of erythropoietin. This hormone then signals the bone marrow to induce erythropoiesis, increasing the number of red blood cells that are released into the circulation to increase O2 levels, increasing the supply of this resource to cells and tissues of the body. In addition to maintaining vigor in response to the homeostatic maintenance of the RBC population, this mechanism is actively engaged under hypoxic conditions and the elevated levels of erythropoietin are necessary to maintain the health trajectory of an organism under these hostile conditions. Infections with Trypanosoma parasites causes anemia and the compensatory induction of erythropoietin and erythropoietin receptor (Nairz et al., 2012). Administration of erythropoietin to T. congolense infected mice promotes the maintenance of health by protecting from anemia despite sustained parasitemia, suggesting that erythropoietin promotes disease tolerance during this infection to maintain oxygen supply and the maintenance of the health trajectory (Suzuki et al., 2006) (Figure 7F). In conditions under which this homeostatic control mechanism does not function properly, homeostatic tolerance mechanisms compensate for these deficits in meeting oxygenation demands to maintain an apparent vigor. In cases of anemia, where adequate tissue oxygenation cannot be met with the levels of red blood cells, physiological compensation occurs by increased blood flow to tissues, caused by decreased systemic vascular resistance, leading to increased stroke volume and cardiac output. There can also be a redistribution of cardiac output so that blood flow is redistributed to tissues with a higher extraction ratio such as the brain and the heart (Gillies et al., 1976; Metivier et al., 2000). There is also a change in the oxygen-hemoglobin affinity, such that the dissociation curve shifts to release oxygen into tissues with higher partial pressures (Figure 7F). Each of these mechanisms of physiological compensation sustain health and the health trajectory when oxygen homeostasis is disrupted due to the failure of the homeostatic tolerance mechanism and therefore promotes an apparent vigor (Figure 7F).
Detoxification
Byproducts of normal cellular and physiological processes in the body can be toxic. The detoxification processes that occur in the liver and subsequent excretion is essential for maintaining homeostasis and vigor. As an example, senescent red blood cells are removed via a process called erythrophagocytosis by macrophages mainly in the spleen, and by Kupffer cells of the liver. During this process, hemoglobin is broken down, generating free heme as a byproduct that can be toxic to cells and tissues. To deal with this byproduct, heme oxygenase-1 (HO-1), present in Kupffer cells and macrophages detoxifies heme, producing bilirubin that is then conjugated to albumin. In hepatocytes, bilirubin is conjugated to glucuronic acid by glucuronidation, and the soluble conjugate is excreted via bile (Fraser et al., 2011). Derivatives of this mechanism have evolved to promote health through neutralization and disease tolerance mechanisms when challenged with hostile environments (Figure 6). Plasmodium, the causative agent of malaria, replicates within red blood cells in its vertebrate host. When parasite densities increase, red blood cell lysis occurs, releasing hemoglobin into the circulation, that when exposed to reactive oxygen species becomes oxidized and releases free heme (Seixas et al., 2009). In mouse models of Plasmodium infection, free heme synergizes with TNFa, released from macrophages, and causes liver damage. Hmox1, the gene encoding HO-1, is induced in hepatocytes during Plasmodium infection. The HO-1 detoxification mechanism neutralizes the pathogenic effects of free heme and promotes survival without altering parasite burdens, indicating that this activity of HO-1 promotes the maintenance of health and the health trajectory via an anti-virulence mechanism (Seixas et al., 2009) (Figure 7G). Carbon monoxide (CO) is a byproduct of HO-1 mediated detoxification of free heme. Administration of CO alone promotes disease tolerance by limiting tissue susceptibility to damage in response to a variety of infection types and stressors (Hoetzel et al., 2007). In sepsis, intestinal dysmotility contributes to the pathological process. CO regulates the voltage gradient of the smooth muscle in the intestine regulating its contractility. CO administration maintained the smooth muscle potential gradient and motility that would promote disease tolerance and the health trajectory during sepsis (Hoetzel et al., 2007) (Figure 7G). This is one of many examples of detoxification as a potential mechanism for neutralization defenses. Whether other mechanisms of detoxification function in processes of disease tolerance in other contexts is unknown.
Conclusions
While we understand in exquisite detail how many diseases occur, health is a fundamental and active biological process that we have difficulty defining and describing mechanistically. This perspective described a new framework for the conceptualization of what it means to be healthy mechanistically and the application to experimental systems to elucidate health mechanisms. Achieving health relies on evolved mechanisms that enable an individual to remain healthy under homeostatic and hostile conditions. Homeostatic control mechanisms operate under normal conditions to maintain health under normal conditions. When there is a failure in homeostatic control mechanisms, homeostatic tolerance mechanisms facilitate the individual’s ability to adapt to this new internal state to yield an apparent vigor. When challenged with hostile conditions, defensive health mechanisms that enable the individual to antagonize the threat (avoidance and resistance) or withstand the threat (disease tolerance and neutralization) promote maintenance and resilience of the individual. While the mechanisms are distinct, homeostatic tolerance, disease tolerance and neutralization mechanisms are derivatives of and likely evolved from homeostatic control mechanisms to control the same physiological variables including growth and development, macro/micro-nutrient and vitamins, socialization, thermoregulation, energy management, oxygenation, detoxification, acid-base balance and osmoregulation. As I hope is apparent, with a shift in our perspective, over the upcoming decades we can elucidate the full spectrum of health mechanisms to complement our understanding of disease. In order to do this, we must understand the health path of an individual that remains healthy in response to the intrinsic or extrinsic insult; and understand the mechanisms that can shift the disease path to the health path based on where they are on their disease course. This knowledge coupled with our existing knowledge on the disease paths an individual can take when challenged with an intrinsic or extrinsic insult and our ability to define where an individual is on their disease path with readily accessible markers will enable us to better treat patients and override physiological decline to extend healthspan.
Acknowledgements
JSA is supported by the NOMIS Foundation, The Keck Foundation, an NIH Pioneer award and NIH grant R01 AI114929. Images for some figures include icons from BioRender.
Footnotes
Declaration of interests. JSA holds an adjunct faculty position in the Biological Division at UC San Diego. The author has no financial interests to declare.
References
- Ayres JS (2013). Inflammasome-microbiota interplay in host physiologies. Cell Host Microbe 14, 491–497. [DOI] [PubMed] [Google Scholar]
- Ayres JS (2016). Cooperative Microbial Tolerance Behaviors in Host-Microbiota Mutualism. Cell 165, 1323–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ayres JS (2017). Microbes Dress for Success: Tolerance or Resistance? Trends Microbiol 25, 1–3. [DOI] [PubMed] [Google Scholar]
- Ayres JS, and Schneider DS (2009). The role of anorexia in resistance and tolerance to infections in Drosophila. PLoS Biol 7, e1000150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ayres JS, and Schneider DS (2012). Tolerance of infections. Annu Rev Immunol 30, 271–294. [DOI] [PubMed] [Google Scholar]
- Barber RC, Maass DL, White DJ, Horton JW, Wolf SE, Minei JP, and Zang QS (2014). Deficiency in Heat Shock Factor 1 (HSF-1) Expression Exacerbates Sepsis-induced Inflammation and Cardiac Dysfunction. SOJ Surg 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Biddle KD, d’Oleire Uquillas F, Jacobs HIL, Zide B, Kirn DR, Rentz DM, Johnson KA, Sperling RA, and Donovan NJ (2019). Social Engagement and Amyloid-beta-Related Cognitive Decline in Cognitively Normal Older Adults. Am J Geriatr Psychiatry 27, 1247–1256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Blandino-Rosano M, Alejandro EU, Sathyamurthy A, Scheys JO, Gregg B, Chen AY, Rachdi L, Weiss A, Barker DJ, Gould AP, et al. (2012). Enhanced beta cell proliferation in mice overexpressing a constitutively active form of Akt and one allele of p21Cip. Diabetologia 55, 1380–1389. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carey HV, Andrews MT, and Martin SL (2003). Mammalian hibernation: cellular and molecular responses to depressed metabolism and low temperature. Physiol Rev 83, 1153–1181. [DOI] [PubMed] [Google Scholar]
- Castellano G, Affuso F, Conza PD, and Fazio S (2009). The GH/IGF-1 Axis and Heart Failure. Curr Cardiol Rev 5, 203–215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cerf ME (2013). Beta cell dynamics: beta cell replenishment, beta cell compensation and diabetes. Endocrine 44, 303–311. [DOI] [PubMed] [Google Scholar]
- Corbett S, Courtiol A, Lummaa V, Moorad J, and Stearns S (2018). The transition to modernity and chronic disease: mismatch and natural selection. Nat Rev Genet 19, 419–430. [DOI] [PubMed] [Google Scholar]
- Cremer S (2019). Social immunity in insects. Curr Biol 29, R458–R463. [DOI] [PubMed] [Google Scholar]
- Curtis V, de Barra M, and Aunger R (2011). Disgust as an adaptive system for disease avoidance behaviour. Philos Trans R Soc Lond B Biol Sci 366, 389–401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Del Coso J, Hamouti N, Aguado-Jimenez R, and Mora-Rodriguez R (2009). Respiratory compensation and blood pH regulation during variable intensity exercise in trained versus untrained subjects. Eur J Appl Physiol 107, 83–93. [DOI] [PubMed] [Google Scholar]
- Duerr RL, Huang S, Miraliakbar HR, Clark R, Chien KR, and Ross J Jr. (1995). Insulin-like growth factor-1 enhances ventricular hypertrophy and function during the onset of experimental cardiac failure. J Clin Invest 95, 619–627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ertl G, Gaudron P, Eilles C, Schorb W, and Kochsiek K (1991). Compensatory mechanisms for cardiac dysfunction in myocardial infarction. Basic Res Cardiol 86 Suppl 3, 159–165. [DOI] [PubMed] [Google Scholar]
- Fraser ST, Midwinter RG, Berger BS, and Stocker R (2011). Heme Oxygenase-1: A Critical Link between Iron Metabolism, Erythropoiesis, and Development. Adv Hematol 2011, 473709. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Geng Y, Zhao W, Zhou F, Ma X, Yao S, Hurlemann R, Becker B, and Kendrick KM (2018). Oxytocin Enhancement of Emotional Empathy: Generalization Across Cultures and Effects on Amygdala Activity. Front Neurosci 12, 512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gillies ID, White YS, and White JM (1976). Compensatory mechanisms for the severe anaemia caused by haemoglobin Hammersmith. Eur J Clin Invest 6, 213–220. [DOI] [PubMed] [Google Scholar]
- Hart BL (2011). Behavioural defences in animals against pathogens and parasites: parallels with the pillars of medicine in humans. Philos Trans R Soc Lond B Biol Sci 366, 3406–3417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hoetzel A, Dolinay T, Schmidt R, Choi AM, and Ryter SW (2007). Carbon monoxide in sepsis. Antioxid Redox Signal 9, 2013–2026. [DOI] [PubMed] [Google Scholar]
- Jamieson AM, Pasman L, Yu S, Gamradt P, Homer RJ, Decker T, and Medzhitov R (2013). Role of tissue protection in lethal respiratory viral-bacterial coinfection. Science 340, 1230–1234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kappeler PM, Cremer S, and Nunn CL (2015). Sociality and health: impacts of sociality on disease susceptibility and transmission in animal and human societies. Philos Trans R Soc Lond B Biol Sci 370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Keubler LM, Buettner M, Hager C, and Bleich A (2015). A Multihit Model: Colitis Lessons from the Interleukin-10-deficient Mouse. Inflamm Bowel Dis 21, 1967–1975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kloting N, Koch L, Wunderlich T, Kern M, Ruschke K, Krone W, Bruning JC, and Bluher M (2008). Autocrine IGF-1 action in adipocytes controls systemic IGF-1 concentrations and growth. Diabetes 57, 2074–2082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Koshiyama H (2012). Explanation of the insulin paradox from the evolutionary point of view. Jpn Clin Med 3, 21–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Krause EG, de Kloet AD, Flak JN, Smeltzer MD, Solomon MB, Evanson NK, Woods SC, Sakai RR, and Herman JP (2011). Hydration state controls stress responsiveness and social behavior. J Neurosci 31, 5470–5476. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Louie A, Song KH, Hotson A, Thomas Tate A, and Schneider DS (2016). How Many Parameters Does It Take to Describe Disease Tolerance? PLoS Biol 14, e1002435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Luan HH, Wang A, Hilliard BK, Carvalho F, Rosen CE, Ahasic AM, Herzog EL, Kang I, Pisani MA, Yu S, et al. (2019). GDF15 Is an Inflammation-Induced Central Mediator of Tissue Tolerance. Cell 178, 1231–1244 e1211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Medzhitov R (2009). Damage control in host-pathogen interactions. Proc Natl Acad Sci U S A 106, 15525–15526. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Medzhitov R, Schneider DS, and Soares MP (2012). Disease tolerance as a defense strategy. Science 335, 936–941. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Metivier F, Marchais SJ, Guerin AP, Pannier B, and London GM (2000). Pathophysiology of anaemia: focus on the heart and blood vessels. Nephrol Dial Transplant 15 Suppl 3, 14–18. [DOI] [PubMed] [Google Scholar]
- Murapa P, Ward MR, Gandhapudi SK, Woodward JG, and D’Orazio SE (2011). Heat shock factor 1 protects mice from rapid death during Listeria monocytogenes infection by regulating expression of tumor necrosis factor alpha during fever. Infect Immun 79, 177–184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Murray MJ, and Murray AB (1979). Anorexia of Infection as a Mechanism of Host Defense. Am J Clin Nutr 32, 593–596. [DOI] [PubMed] [Google Scholar]
- Nairz M, Sonnweber T, Schroll A, Theurl I, and Weiss G (2012). The pleiotropic effects of erythropoietin in infection and inflammation. Microbes Infect 14, 238–246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Paaby AB, and Rockman MV (2013). The many faces of pleiotropy. Trends Genet 29, 66–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Park TJ, Reznick J, Peterson BL, Blass G, Omerbasic D, Bennett NC, Kuich P, Zasada C, Browe BM, Hamann W, et al. (2017). Fructose-driven glycolysis supports anoxia resistance in the naked mole-rat. Science 356, 307–311. [DOI] [PubMed] [Google Scholar]
- Pavlov VA, and Tracey KJ (2012). The vagus nerve and the inflammatory reflex-linking immunity and metabolism. Nat Rev Endocrinol 8, 743–754. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pearson-Stuttard J, and Gregg EW (2019). Decreasing mortality masks a growing morbidity gap in patients with heart failure. Lancet Public Health 4, e365–e366. [DOI] [PubMed] [Google Scholar]
- Pickard JM, Maurice CF, Kinnebrew MA, Abt MC, Schenten D, Golovkina TV, Bogatyrev SR, Ismagilov RF, Pamer EG, Turnbaugh PJ, et al. (2014). Rapid fucosylation of intestinal epithelium sustains host-commensal symbiosis in sickness. Nature 514, 638–641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pontzer H, Raichlen DA, Wood BM, Mabulla AZ, Racette SB, and Marlowe FW (2012). Hunter-gatherer energetics and human obesity. Plos One 7, e40503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Raberg L, Graham AL, and Read AF (2009). Decomposing health: tolerance and resistance to parasites in animals. Philos Trans R Soc Lond B Biol Sci 364, 37–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Raberg L, Sim D, and Read AF (2007). Disentangling genetic variation for resistance and tolerance to infectious diseases in animals. Science 318, 812–814. [DOI] [PubMed] [Google Scholar]
- Rao S, Schieber AM, O’Connor CP, Leblanc M, Michel D, and Ayres JS (2017). Pathogen-Mediated Inhibition of Anorexia Promotes Host Survival and Transmission. Cell 168, 503–516 e512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rauber A (1990). Growth and Development. In Clinical Methods: The History, Physical, and Laboratory Examinations, rd, Walker HK, Hall WD, and Hurst JW, eds. (Boston: ). [PubMed] [Google Scholar]
- Sanchez KK, Chen GY, Schieber AMP, Redford SE, Shokhirev MN, Leblanc M, Lee YM, and Ayres JS (2018). Cooperative Metabolic Adaptations in the Host Can Favor Asymptomatic Infection and Select for Attenuated Virulence in an Enteric Pathogen. Cell 175, 146–158 e115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schieber AM, and Ayres JS (2016). Thermoregulation as a disease tolerance defense strategy. Pathog Dis 74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schieber AM, Lee YM, Chang MW, Leblanc M, Collins B, Downes M, Evans RM, and Ayres JS (2015). Disease tolerance mediated by microbiome E. coli involves inflammasome and IGF-1 signaling. Science 350, 558–563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schlichting C, and Pigliucci M (1998). Phenotypic evolution : a reaction norm perspective (Sunderland, Mass.: Sinauer; ). [Google Scholar]
- Schneider DS (2011). Tracing personalized health curves during infections. PLoS Biol 9, e1001158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schneider DS, and Ayres JS (2008). Two ways to survive infection: what resistance and tolerance can teach us about treating infectious diseases. Nat Rev Immunol 8, 889–895. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schulkin J (2011). Social allostasis: anticipatory regulation of the internal milieu. Front Evol Neurosci 2, 111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Seifter JL, and Chang HY (2017). Disorders of Acid-Base Balance: New Perspectives. Kidney Dis (Basel) 2, 170–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Seixas E, Gozzelino R, Chora A, Ferreira A, Silva G, Larsen R, Rebelo S, Penido C, Smith NR, Coutinho A, et al. (2009). Heme oxygenase-1 affords protection against noncerebral forms of severe malaria. Proc Natl Acad Sci U S A 106, 15837–15842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shankar-Hari M, and Rubenfeld GD (2016). Understanding Long-Term Outcomes Following Sepsis: Implications and Challenges. Curr Infect Dis Rep 18, 37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sohn JW, Elmquist JK, and Williams KW (2013). Neuronal circuits that regulate feeding behavior and metabolism. Trends Neurosci 36, 504–512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Steppuhn A, B.I.T. (2008). Induced Defenses and the Cost-Benefit Paradigm In: Schaller A (eds) Induced Plant Resistance to Herbivory Springer, Dordrecht. [Google Scholar]
- Stollman TH, de Blaauw I, Wijnen MH, van der Staak FH, Rieu PN, Draaisma JM, and Wijnen RM (2009). Decreased mortality but increased morbidity in neonates with jejunoileal atresia; a study of 114 cases over a 34-year period. J Pediatr Surg 44, 217–221. [DOI] [PubMed] [Google Scholar]
- Suzuki T, Ueta YY, Inoue N, Xuan X, Saitoh H, and Suzuki H (2006). Beneficial effect of erythropoietin administration on murine infection with Trypanosoma congolense. Am J Trop Med Hyg 74, 1020–1025. [PubMed] [Google Scholar]
- Taner T, Iqbal CW, Textor SC, Stegall MD, and Ishitani MB (2015). Compensatory hypertrophy of the remaining kidney in medically complex living kidney donors over the long term. Transplantation 99, 555–559. [DOI] [PubMed] [Google Scholar]
- Torres BY, Oliveira JH, Thomas Tate A, Rath P, Cumnock K, and Schneider DS (2016). Tracking Resilience to Infections by Mapping Disease Space. PLoS Biol 14, e1002436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Trivers R (2002). Natural selection and social theory : selected papers of Robert L. Trivers (New York: Oxford University Press; ). [Google Scholar]
- Turturro A, Witt WW, Lewis S, Hass BS, Lipman RD, and Hart RW (1999). Growth curves and survival characteristics of the animals used in the Biomarkers of Aging Program. J Gerontol A Biol Sci Med Sci 54, B492–501. [DOI] [PubMed] [Google Scholar]
- Umberson D, and Montez JK (2010). Social relationships and health: a flashpoint for health policy. J Health Soc Behav 51 Suppl, S54–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wang A, Huen SC, Luan HH, Yu S, Zhang C, Gallezot JD, Booth CJ, and Medzhitov R (2016). Opposing Effects of Fasting Metabolism on Tissue Tolerance in Bacterial and Viral Inflammation. Cell 166, 1512–1525 e1512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Young LJ (2009). The neuroendocrinology of the social brain. Front Neuroendocrinol 30, 425–428. [DOI] [PubMed] [Google Scholar]