We are searching data for your request:
Upon completion, a link will appear to access the found materials.
Humans get sick on such a regular basis and animals will hardly ever get sick. Animals are typically exposed to the same pathogens as humans, yet a dog won't have a cold or the flu twice a year, with other miner health issues occurring regularly in between. I understand that humans are very structurally flawed due to some evolutionary processes, and I'm just wondering why our immune systems seemed to have taken a pretty big hit as well.
The question probably has a mistaken premise. Human immune systems are pretty similar to other species, at least in major structural elements. Our differences in adaptive niche have led to differences in some aspects; humans don't fare well with rotting meat, for example, while a scavenger can eat it without a problem. Many scavengers have unique adaptations for dealing with this. Dogs do not, however. Humans too have a unique niche, which is highly social, and thus have a well-developed social immunity (sometimes called a behavioral immune system).
Many animal populations have endemic diseases at fairly high levels. Koalas, for example, famously have a very high level of chlamydia. Sea lions rookeries often have infections, bats can have high levels of rabies, and so forth. Lots of animals do get lots of infections.
However your question specifically on dogs is interesting, and probably gets to a risk factor which is very human. Humans come into contact with far more other humans (and thus, likely more infected humans) than dogs encounter other dogs. Diseases which spread through close proximity, such as airborn upper respiratory viruses, spread more readily in humans than these relatively isolated animal populations. Combine that with a concerted effort to wipe out in animals any diseases which can transmit to humans, and pet infections are indeed relatively rare. This is not due to the immune system per se, but the social factors associated with each organisms' niche.
Primate Immune Systems Offer Clues to Combating Disease
Humans may be smarter than our primate cousins, but research has shown their immune systems may trump ours, in some ways. We tend to be more susceptible to HIV, cancer, hepatitis and other infectious diseases than some of our closest relatives.
Because humans share the majority of our genetic makeup with other primates, scientists at the University of Chicago set out to uncover some of the small differences in the innate immune systems of humans, macaque monkeys and chimpanzees – the latter being humans' closest evolutionary relative. Changes that have arisen in our immune systems since we split from a common ancestor might give clues for disease research.
"We believe these differences in susceptibility are probably differences in immune response," said Luis Barreiro, co-author of a study published today (Dec. 16) in the journal Public Library of Science.
Barreiro and his colleagues took blood samples from six macaque monkeys, six people and six chimpanzees. They tested the effects of diseases on each species by adding a chemical made by bacteria to the blood samples, and then observed the gene activity in the blood cells in response.
At first, all three species had the same "core" initial immune response to lipopolysaccharide (LPS), a molecule that's found on the surface of bacteria.
But after kick-starting the immune system, researchers found differing gene activity across the species. In humans, more gene sets were activated that have been linked to cancer biology, cell death and immune disorders. In chimpanzees, researchers saw greater activity in genes related to warding off HIV.
Evolving new immune systems
Macaques and humans branched off from a common ancestor 13 million years ago, Barreiro said. But chimpanzees and humans diverged from a common ancestor much later, about 5 million years ago.
By comparing blood from all three, Barreiro and his colleagues built a picture of which immune responses were originally shared by all, and which evolved along the way within each species. Barreiro said he was pleasantly surprised that the differences between the three species that the experiment revealed correlated with observed differences between the species in the real world.
"There are a few differences in medical traits between human and non-human primates, one is [humans' greater susceptibility to] HIV, a second one is humans tend to develop cancer at a much higher rate than other primates," Barreiro said.
Narrowing down the genes related to these immune response differences may lead to new research targets among people infected with these diseases, he told MyHealthNewsDaily.
Can comparing humans with other primates fight disease?
Dr. Warner C. Greene, director of the Gladstone Institute of Virology and Immunology at the University of California, San Francisco, said researching primates who are resistant to HIV is a known tactic in the fight against AIDS.
Greene said the Sooty Mangabey monkey is a prime example of a monkey that's intrigued HIV researchers.
"When HIV is injected into the Sooty Mangabeys, they develop very high viral loads," Greene said. "But they won't develop AIDS."
By studying the Sooty Mangabey and other primates who have immunity to HIV, Greene said researchers aim to discover how to mimic this immunity in humans.
Yet for all the intriguing connections between gene activity and real-world differences between the species, Greene and Barreiro say the study's findings are preliminary. More research would be needed to start new infectious disease experiments.
Instead of using immunity activators such as LPS, Barreiro said the next step is to " actually infect the cells with different pathogens.
In the meantime, Barreiro's paper has caught the eye of other scientists as a new research approach.
Comparing animal genomes for medical research
"It's a nice piece of work," said Dr. Duncan Odom, a principal investigator at the University of Cambridge Cancer Center in the U.K.
Odom was intrigued by the HIV and cancer implications the study found. But he also noted that the LPS – the chemical used in Barreiro's experiment – is made by bacteria. Using bacteria-like immune system stimulants may not be the best way to get information about the body's reaction to viruses such as HIV or hepatitis.
More importantly than its HIV implications, Odom said, the paper opened up research into the evolution of immune systems.
He also sees the research as symbol of the power of emerging genomic research.
"Traditionally, we thought that a mouse was our model organism," Odom said, explaining that most medical research has been done with mice as a model for humans.
Now, with more animal genomes mapped and experiments such as Barreiro's, Odom said, "We can get a lot of information from species that we normally would not think of as a source of information."
"Five years ago, this would be an impossible study do to, full stop," Odom said. "This is going to become a very increasing theme in the next 20 years."
The behavioural immune system and the psychology of human sociality
Because immunological defence against pathogens is costly and merely reactive, human anti-pathogen defence is also characterized by proactive behavioural mechanisms that inhibit contact with pathogens in the first place. This behavioural immune system comprises psychological processes that infer infection risk from perceptual cues, and that respond to these perceptual cues through the activation of aversive emotions, cognitions and behavioural impulses. These processes are engaged flexibly, producing context–contingent variation in the nature and magnitude of aversive responses. These processes have important implications for human social cognition and social behaviour—including implications for social gregariousness, person perception, intergroup prejudice, mate preferences, sexual behaviour and conformity. Empirical evidence bearing on these many implications is reviewed and discussed. This review also identifies important directions for future research on the human behavioural immune system—including the need for enquiry into underlying mechanisms, additional behavioural consequences and implications for human health and well-being.
Humans and other animals have a long history of living in proximity to parasitic organisms—bacteria, viruses, helminths—that cause infectious diseases. This proximity imposed substantial selection pressures on ancestral populations, resulting in many different adaptations that, in a variety of ways, mitigate the potential fitness costs posed by these pathogens. Most obviously, there evolved the sophisticated suite of physiological mechanisms that define immunological defence systems, which are designed to detect the presence of pathogens within the body and, when detected, to mobilize physiological responses that encapsulate, kill or otherwise eliminate these pathogenic intruders. Immunological defence against infection has obvious fitness benefits, but can be substantially costly too . An immune response is metabolically costly (consuming caloric resources that might otherwise be devoted to other important physiological systems) and can be temporarily debilitating (because of fever, fatigue and other physiological consequences of an aggressive immunological response). And, of course, immunological defence is merely reactive—triggered only after the pathogenic infection has occurred within the body.
Given these limits and costs associated with immunological defence against pathogens, additional fitness benefits would have accrued from an additional set of proactive mechanisms that—by guiding organisms' behaviour—inhibit contact with pathogens in the first place. These mechanisms offer a sort of behavioural prophylaxis against infection . Indeed, it is not merely metaphorical to suggest that these mechanisms comprise a behavioural immune system that is separate from, and complementary to, the ‘real’ immune system [3–5].
Behavioural defence against pathogens has been observed across a wide variety of animal species [6,7]. Some forms of behavioural defence—such as cytokine-induced sickness behaviour [8,9] and self-medication [10,11]—are reactive, rather than proactive. But there is also abundant evidence of proactive behavioural defence as well: wood ants collect pieces of coniferous resin as a prophylactic defence against bacteria and pathogenic fungi  bullfrog tadpoles selectively avoid swimming near infected tadpoles  female mice respond aversively to the odours of male mice infected with nematode parasites  chimpanzees avoid social contact with (and may even respond aggressively towards) other chimpanzees infected with polio . In short, just as the ‘real’ immune system is characterized by mechanisms that facilitate adaptive immunological responses to pathogens that enter the body, the behavioural immune system is characterized by mechanisms that facilitate adaptive psychological responses to perceptual cues connoting the presence of pathogens in the immediate perceptual environment—including the presence of pathogens in conspecifics. The specific nature of the perceptual detection and behavioural response mechanisms may vary across species, but the existence of these detection and response mechanisms is common across species.
In recent years, the behavioural immune system has received considerable attention in the study of human behaviour, with an emphasis on the specific psychological mechanisms (pertaining to attention, perception, cognition and emotion) that guide human behaviour. Much of this work has focused on one specific emotion—disgust—that is associated with disease-avoidance behaviour, on the specific kinds of perceptual things that elicit disgust, and on the specific circumstances under which a disgust response is either exaggerated or reduced [2,16,17]. This work has been reviewed extensively elsewhere [18,19]. My focus here is on a set of complementary programmes of research that focus less on emotion and more on social cognition and social interaction—lines of enquiry that explore how the behavioural immune system guides people's perceptions of, thoughts about and behaviour towards other individuals within their immediate social ecologies. The upshot is an emerging literature documenting many subtle but important linkages between anti-pathogen defence and the contours of human social life.
2. The behavioural immune system: detection, response and functional flexibility
Like the ‘real’ immune system, the behavioural immune system includes both detection and response mechanisms. Detection mechanisms employ sensory systems (e.g. olfaction, vision) to detect things that appear to pose some infection risk. These things include inanimate objects (e.g. a pile of faeces, a piece of putrid meat). These things also include conspecifics (i.e. other people) who may be inferred to pose an infection risk either because (i) they appear already to be infected or because (ii) they tend to behave in ways that increase the likelihood that infections will be spread to others (e.g. by failing to observe customary hygiene practices). When a superficial cue connoting infection risk is detected, it triggers a cascade of adaptive psychological responses. These responses include not only the emotional experience of disgust but also the activation of aversive cognitions into working memory, and the arousal of a motivational system that guides decision-making strategies and motor movements in ways that minimize the infection risk (e.g. behavioural avoidance of and social condemnation of people who appear to pose an infection risk).
Of course, just as immunological response has costs (as well as benefits), the detection and response mechanisms that characterize the behavioural immune system can also be costly (as well as adaptively beneficial). These costs arise because disease-avoidant behavioural responses can consume considerable metabolic resources, and because these responses may inhibit the satisfaction of other fitness-relevant goals (e.g. avoidance of interpersonal contact can interfere with valuable opportunities for social exchange, mating, etc.). Therefore, just as with other adaptive psychological systems, the mechanisms that comprise the behavioural immune system are ‘functionally flexible’: They are sensitive to contextual information bearing on cost–benefit ratio, with predictable context–contingent variation in the nature and magnitude of response [5,20,21].
The benefits offered by the behavioural immune system (reduction of infection risk) are a direct function of individuals' actual vulnerability to infection: these benefits are minimal under conditions in which perceivers are invulnerable to infection, and are relatively greater under conditions in which vulnerability to infection is also relatively greater. Consequently, the detection and response mechanisms that characterize the behavioural immune system are sensitive to any kind of information that suggests increased vulnerability to the transmission of infectious diseases. This information may arise from sources either internal to the perceiver (e.g. chronic anxieties and worries) or in the external environment (e.g. context-specific perceptual reminders of the threat posed by infectious diseases). This information may be veridical (e.g. correctly implying that the perceivers' immunological defences are compromised) but it need not be. Indeed, regardless of the source or veracity of an individual's subjective perception of vulnerability to infection, that subjective perception is likely to influence the activation of the behavioural immune system. Under conditions in which individuals perceive themselves to be more vulnerable to infection, they are expected to be more perceptually sensitive to things (including people) who appear to pose an infection risk  and when those things (including people) are detected, those perceivers are expected to exhibit more exaggerated aversive responses—greater disgust, greater activation of aversive cognitions into working memory, greater motivation for behavioural avoidance and so forth.
The functional flexibility of the behavioural immune system has many implications for human social interaction, and for human sociality more broadly. Many kinds of human social behaviour that serve disease-irrelevant goals (e.g. acquisition of resources, sexual reproduction) also have potential implications for disease transmission too. Consequently, these social behavioural tendencies may vary, depending on the extent to which individuals are (or merely perceive themselves to be) vulnerable to infectious disease. This has important consequences for a wide variety of social attitudes, social perceptions and social activities.
3. Social gregariousness
Research on the structure of human personality reveals a small handful of fundamental trait dimensions that characterize individual differences in psychological functioning. One of those dimensions is extraversion—the extent to which individuals are socially gregarious or not.
Gregariousness is typically considered to be beneficial. Indeed, empirical research shows that extraversion is associated with many positive outcomes, including higher levels of happiness and increased opportunities for sexual reproduction [23,24]. But gregariousness may have infection-specific costs as well. People who are more gregarious tend to come into interpersonal contact with a relatively larger number of people, with the implication that they are more likely to be exposed to interpersonally transmitted pathogens [25,26]. These costs of gregariousness are relatively greater (and more likely to outweigh the social benefits) under conditions in which individuals are more vulnerable to infection. Therefore, when people feel relatively invulnerable to infection, they may show a natural tendency towards gregariousness. However, under conditions in which people perceive themselves to be vulnerable, it follows that they will be less sociably inclined.
Two recent experiments tested and supported this hypothesis . Both experiments included an experimental manipulation designed to make some participants (compared with those in a control condition) especially aware of the threat posed by infectious pathogens. One experiment assessed participants' personality traits, and found that the pathogen-salience manipulation caused participants to rate themselves as relatively less extraverted. The second experiment measured actual motor movements in response to the visual perception of other people, and found that the pathogen-salience manipulation caused participants to engage in relatively more socially avoidant motor movements. Together, these findings indicate that the perceived threat of infectious disease has predictable implications for individuals' basic behavioural tendency towards social gregariousness.
A parallel pattern of variation in social gregariousness is found when treating entire populations—rather than individuals—as units of analysis. There is considerable worldwide ecological variability in the historical prevalence of infectious diseases. There is also considerable worldwide cultural variability in dispositional tendencies towards extraversion. These two variables are linked: among populations living in regions that historically have had a high prevalence of pathogens, the mean level of extraversion is lower .
4. Discriminatory sociality (prejudice)
The inhibition of social gregariousness may offer one means of reducing infection risk, but it is a rather blunt tool for doing so. In many animal species, behavioural disease avoidance is indicated not so much by unsociable behaviour in general, but by discriminatory unsociable behaviour: the use of diagnostic perceptual cues to selectively avoid particular conspecifics that appear to pose a particularly high risk of infection [13–15]. The same sort of discriminatory sociality is observed in humans. Aversive emotional and cognitive responses are aroused by the perception of other individuals who are known to be diseased, or who are judged to be at greater risk of being diseased and these aversive responses are especially pronounced when the diseases are perceived to be especially infectious [29,30]. Thus, as a result of the behavioural immune system and its implications for discriminatory sociality, many people suffering from infectious diseases also suffer from prejudice and social stigmatization as well.
Importantly (and troublingly), the evolved design of the behavioural immune system can not only lead to the social stigmatization of people who truly are infectious but also to equally pernicious prejudices directed against people who are not. Here is why:
Most disease-causing organisms (e.g. bacteria) are so tiny as to be imperceptible to human perceptual processes, and so it is largely impossible for people to directly detect the presence of pathogen infection in others. Because of this fact, the behavioural immune system responds to the inferred presence of parasites as indicated by superficial sensory cues (e.g. coughing spasms, skin discolorations). These cues may be probabilistically predictive of the presence of infection, but are still imperfectly diagnostic. This results in a signal-detection problem, with the potential to make both false-positive errors (a healthy person is erroneously perceived to be infectious) and false-negative errors (an infectious person is erroneously perceived to be healthy). From an adaptive perspective, one would expect a particular form of signal-detection bias to emerge: a bias that minimizes the likelihood of making the error with the greatest potential fitness cost, even though that bias inevitably leads to many errors of the opposite kind [31,32]. As with other psychological systems designed for self-protection, false-negative errors are likely to be especially costly. And so the behavioural immune system errs on the side of making false-positive errors instead [4,5,21]. The upshot is that the behavioural immune system is perceptually sensitive to any superficial cue that appears likely to be a symptom of infection—even if it is objectively not.
Furthermore, there is no finite category of superficial cues to which the behavioural immune system is sensitive. This is because there is no finite set of symptoms associated with infection. (Different kinds of parasitic organisms produce different kinds of symptoms. Different people may show somewhat different symptoms even if they are infected with the same species of parasite. And parasitic species can evolve rapidly, with the consequence that their symptomatic manifestations may be highly variable over time.) Thus, to avoid costly false-negative errors, the behavioural immune system must be sensitive to a very broad range of cues that might be potential indicators of infection. Indeed, it has been suggested that any perceived deviation from prototypical human morphology and motor behaviour may implicitly connote potential infection risk [4,5,33].
Therefore, just as the ‘real’ immune system responds not only to actual pathogenic infection but also to intrusion by benign organic matter (as in the case of organ transplants, for example), the behavioural immune system also responds to an over-general set of superficial social cues. The result is a set of predictable prejudices directed at people who may be objectively non-infectious, but who simply have some sort of non-prototypical physical appearance. Furthermore (in keeping with the principle of functional flexibility), these prejudices are likely to be especially pronounced under conditions in which perceivers feel especially vulnerable to infection.
This line of reasoning is supported by many empirical studies, some of which employ the methodological tools of cognitive psychology to assess the automatic activation of semantic concepts into individuals' working memory [4,5]. It has been found, for example, that aversive semantic concepts (such as ‘disease’) are more readily activated into working memory upon encountering people with physical disabilities and people bearing the characteristic features of old age—and that these implicit prejudices occur especially strongly among individuals who feel especially vulnerable to infection [34,35]. Similarly, people who feel more vulnerable to infection express stronger anti-fat attitudes and are more likely to implicitly associate obese people with aversive concepts connoting disease . This latter finding is perhaps especially revealing about the manner in which the behavioural immune system guides discriminatory sociality. Obesity is not objectively diagnostic of pathogen infection (if anything, the opposite is more likely to be true: infectious diseases are more likely to cause weight loss than weight gain) but it does represent a substantial deviation from prototypical human morphology. This finding therefore attests to the behavioural immune system's sensitivity to a very broad category of superficial cues connoting non-normative physical appearance.
An additional programme of research reveals that the behavioural immune system produces a somewhat different form of discriminatory sociality as well: aversive responses to subjectively foreign peoples. There are, of course, many different psychological sources of xenophobia and ethnocentrism, and some of these psychological processes have nothing to do with infectious disease still, disease-avoidant processes apparently contribute to these discriminatory outcomes. There are at least two distinct reasons why subjective ‘foreign-ness’ may implicitly connote an increased infection risk. First, exotic peoples may be host to exotic pathogens that can be especially virulent when introduced to a local population. Second, exotic peoples may be more likely to violate local behavioural norms (in domains pertaining to hygiene, food preparation, etc.) that serve as barriers to pathogen transmission. Thus, perceivers are likely to be hypersensitive to inferential cues that discriminate between familiar and foreign peoples and, when those cues are detected, they are likely to trigger the aversive, discriminatory responses associated with the behavioural immune system. This is especially likely to occur when perceivers feel especially vulnerable to infection.
Many studies now support this hypothesis. One provocative study revealed that women in their first trimester of pregnancy—when the ‘real’ immune system is naturally suppressed—reported exaggerated ethnocentrism and xenophobia . Similar exaggerations in xenophobia occur among people who merely perceive themselves to be especially vulnerable to infection . In one experiment, students at the University of British Columbia (in Canada) watched one of two slide shows: in a control condition, the slide show made salient the threat posed by accidents and mishaps (e.g. electrocution) in the other condition, the slide show made salient the threat posed specifically by infectious pathogens. Participants then completed a task that assessed their interest in attracting, to Canada, immigrants from a variety of countries that were either subjectively familiar (e.g. Poland and Taiwan) or subjectively foreign (e.g. Mongolia and Peru). The pathogen-salience manipulation influenced responses on the immigration attitudes task: compared with the accident-salient control condition, when the threat of pathogen infection was salient, participants indicated a stronger preference for immigrants from familiar places, to the exclusion of those from subjectively foreign places .
Intriguingly, there is also cross-cultural evidence linking xenophobia and intergroup prejudice to worldwide ecological variation in the prevalence of pathogenic diseases. Ecological variation in pathogen prevalence is correlated with the percentage of people in a population who explicitly express intolerance for ‘people of a different race’ in their neighbourhood , and with regional frequency of ethnopolitical warfare . Additionally, collectivistic value systems—which emphasize sharp boundaries between ‘us’ and ‘them’—are especially likely to exist in social ecologies characterized historically by especially high levels of pathogen prevalence . Thus, just as with sociality in general, discriminatory sociality is predicted by infection risk not only at an individual level of analysis, but also at a population level of analysis.
5. Mate preferences and mating behaviour
Specific forms of social behaviour increase individuals' susceptibility to pathogen transmission. Mating behaviour is one obvious example.
Sexual contact with others exposes individuals to a much higher risk for contracting sexually transmitted diseases (e.g. syphilis). In addition, given the intimate physical proximity associated with mating behaviour, it also facilitates transmission of other pathogens as well. Sexual promiscuity therefore poses a potential problem: the more sexual encounter partners an individual has, the greater is that individual's infection risk. Of course, these infection-specific costs of promiscuity must be balanced against potential fitness benefits associated with multiple mating partners (benefits that may be especially pronounced among men ). Following the principle of functional flexibility, one might expect an attitudinal disposition towards sexual promiscuity to be inhibited among people who feel vulnerable to pathogen infection, whereas a preference for promiscuity may be more pronounced among people who feel relatively invulnerable to infection. Consistent with this hypothesis is a negative correlation between individuals' perceived vulnerability to infection and their endorsement of an ‘unrestricted’ (i.e. more promiscuous) sociosexual style .
This inverse relation between vulnerability to infection and preference for promiscuity may apply not just only to one's own sexual behaviour but also to preferred behavioural dispositions of potential mates. This is because one's risk of infection is a function not only of one's own sexual promiscuity but also a function of the sexual promiscuity of anyone with whom one has sexual contact (i.e. a sexually monogamous woman is at minimal risk of infection if her one sexual partner is also monogamous, but at greater risk of infection if her one partner is sexually promiscuous). Therefore, in mating contexts, the sexual promiscuity of other people can be considered a dispositional trait connoting infection risk, and, as such, it is likely to trigger an aversive response from the behavioural immune system—especially under conditions in which perceivers feel especially vulnerable to disease. Recent empirical evidence provides preliminary support for this hypothesis: people who feel more chronically vulnerable to infection indicate a stronger preference for non-promiscuous mates, and this effect itself is especially pronounced under conditions in which the threat of infectious diseases is psychologically salient (D. R. Murray, D. N. Jones & M. Schaller 2010, unpublished raw data).
The behavioural immune system may also influence the extent to which other kinds of traits are valued in a mate. Physical attractiveness is one such trait. Subjective assessments of facial attractiveness are influenced by specific aspects of facial physiognomy that are associated with genetic quality . Consequently, subjective assessments of physical attractiveness may be somewhat diagnostic not only of a potential mate's own immunological competence but also diagnostic of the immunological competence of any offspring produced by that potential mate. This line of reasoning offers a partial explanation for the high value that people place on the physical attractiveness of a mate. It also implies that the typical preference for physically attractive mates may be exaggerated even further under conditions in which people feel more vulnerable to infectious diseases. Results from a recent experiment support this hypothesis: romantic interest in physically attractive (compared with unattractive) opposite-sex individuals is exaggerated under circumstances in which the threat posed by infectious diseases is temporarily salient (A. Beall & M. Schaller 2010, unpublished raw data).
Cross-national analyses reveal conceptually similar linkages between ecological variability in pathogen prevalence and cultural differences in mating behaviour. In countries that historically have had a relatively higher prevalence of pathogenic diseases, people (especially women) report attitudes endorsing relatively more ‘restricted’ (i.e. less promiscuous) strategies of mating behaviour . Ecological differences in pathogen prevalence also predict cultural variation in mate preferences, with physical attractiveness emerging as an especially prized attribute in populations characterized historically by an especially high prevalence of pathogens [45,46].
6. Normative and counter-normative behaviour
The implications of the behavioural immune system for xenophobia, discussed above, are predicated in part on the possibility that foreign peoples may be especially likely to violate local behavioural practices (e.g. hygiene rituals, food preparation norms) that inhibit the spread of infectious diseases. The potential for norm violation is not specific to foreign peoples anyone might potentially engage in non-normative behaviour. And while non-normative behaviour can have substantial infection-specific costs (e.g. a free thinker who violates normative practices pertaining to defaecation may increase the infection risk of the entire local population), it is also potentially beneficial (especially when that non-normative behaviour produces technological innovations and novel solutions to enduring problems). The ratio of costs to benefits is likely to be a function of the threat posed by infectious pathogens. Under circumstances in which the threat posed by pathogens is rather modest, the benefits of non-normative behaviour may outweigh the costs. But under circumstances in which people are more highly vulnerable to pathogen infection, the costs of non-normative behaviour increase accordingly, and may outweigh the benefits. Drawing on the principle of functional flexibility, it follows that the extent to which individuals favour normative versus non-normative tendencies is likely to vary depending on the extent to which those individuals feel vulnerable to infectious disease. Compared with circumstances in which people feel relatively invulnerable, when people feel more vulnerable to infection they may be relatively more conformist in their own behavioural tendencies and also less tolerant of others' non-normative behaviour. (In addition, these effects may be more pronounced in behavioural domains that are more clearly linked to the transmission of infectious diseases.)
This hypothesis has yet to be rigorously tested in laboratory research, but some preliminary evidence is supportive: people who report higher levels of chronic concern with infection also report more conformist attitudes, and people also show an increased tendency to conform to majority opinion under conditions in which the threat of infectious disease is temporarily salient (D. R. Murray & M. Schaller 2011, unpublished raw data).
In addition, there is now ample evidence linking ecological variation in pathogen prevalence to cultural variation in conformity relevant dispositions and values. The linkage between pathogen prevalence and collectivistic value systems is indirectly supportive, given that collectivism is defined in part by a higher value placed on the conservation of cultural traditions . Also indirectly supportive are results revealing that higher levels of pathogen prevalence are associated with lower population-level scores on the personality trait ‘openness to experience’, which is associated with novelty-seeking and tolerance for inconsistency . More convincing support emerges from recent findings on more focused measures of cultural conformity pressure and tolerance for non-conformity: the results reveal that, in places characterized historically by a higher prevalence of pathogens, there exist stronger cultural pressures towards obedience and conformity, as well as a reduced tolerance for non-normative behaviour .
7. Questions, speculations and directions for future research
These lines of research reveal that the mechanisms of behavioural disease-avoidance influence a diverse set of social psychological phenomena. The behavioural immune system matters not only because of its implications for anti-pathogen defence but also because of its implications for social perception, social cognition and the social lives of human beings.
Along with these wide-ranging implications, there also emerges a diverse set of scientific questions that can only be answered through further scientific investigation.
(a) Underlying mechanisms
Most discussions of the human behavioural immune system assume that the psychological mechanisms that define it are likely to have been adaptive throughout long stretches of human evolutionary history [4,19,33]. This assumption does not, however, imply that each mechanism is an adaptation specific to selection pressures posed solely by infectious diseases. Many of the psychological systems employed by the behavioural immune system are likely to have evolved in response to additional selection pressures as well—some of which may have predated threats posed by infectious pathogens. An obvious example: the human behavioural immune system employs sensory organs (such as the eyes) in the service of detecting infection-connoting cues, but this hardly implies that these organs evolved specifically as adaptations ‘for’ anti-pathogen defence. Similarly, while the emotional experience of disgust is integral to the suite of adaptive responses associated with the behavioural immune system (and is causally linked to social outcomes such as xenophobia and the moral condemnation of norm violators [48,49]), the physiological substrates of disgust may have evolved originally as a means of facilitating the expulsion of harmful things that have been ingested orally . So, while it is sensible to assume evolutionary origins for the various mechanisms that define the behavioural immune system, it is probably sensible also to assume that some of these mechanisms have ancient evolutionary origins that predate the behavioural immune system, and were adaptively ‘re-purposed’ in response to selection pressures imposed by infectious diseases.
Any speculations about the evolutionary origins of these psychological mechanisms, and the manner in which they have been adaptively coordinated in the service of anti-pathogen defence, can be usefully buttressed by enquiry into the underlying physical substrates—at anatomical, neurochemical and genetic levels of analysis. It will therefore be useful for future research on the human behavioural immune system to follow the lead of behavioural neuroscientists who study disease-avoidant behaviour in other species. For instance, many mammals are sensitive to olfactory cues that are diagnostic of infection risk these olfactory cues are employed to identify infected conspecifics and trigger avoidant responses. Correlates of these behavioural processes have been identified at both genetic and neurochemical levels of analysis, in the form of specific genes coding for neuropeptide, oxytocin and oestrogenic mechanisms .
Of course, while it is instructive to examine disease-avoidant behaviour in other animals to inform enquiry into the human behavioural immune system, it is also important to recognize elements of human psychology that are unique, and thus may have unique implications for disease-avoidant behaviour. Compared with other animals, humans have unusually massive neocortical brain structures consequently, humans have relatively greater capacities for perspective-taking, deliberative thought and the intentional inhibition of behavioural impulses. Thus, while the perception of another person's infection-connoting features may automatically trigger aversive affective and cognitive states, these affective and cognitive states may not necessarily manifest in avoidant motor behaviour. These behavioural impulses may be muted by more deliberative cognitive processes—which, for instance, allow people to sympathize with conspecifics who are sick or disabled, and facilitate behaviour that is nurturant rather than neglectful . A fuller understanding of the behavioural immune system, and its impact on human sociality, must consider its highly automated mechanisms in conjunction with the more deliberative and controllable cognitive processes that also influence human behaviour.
An additional kind of enquiry into underlying mechanisms is implied by results that test disease-avoidance predictions at a cultural rather than at an individual level of analysis. As reviewed above, many of the findings observed in laboratory experiments (e.g. the psychological salience of infectious diseases leads individuals to be less gregarious and more xenophobic) have parallels in comparisons between different cultural populations (e.g. in geographical regions characterized by high pathogen prevalence, human populations tend to be characterized by lower mean levels of extraversion and more xenophobic cultural values). It is perhaps tempting to assume that these cultural differences simply represent population-level outcomes of the same psychological processes that account for the laboratory findings (e.g. the operation of functionally flexible neurocognitive systems). But this single explanation cannot account for the full pattern of cross-cultural findings (including the fact that these cultural differences are typically predicted less strongly by contemporary pathogen prevalence than by historical pathogen prevalence [28,41]). A variety of additional explanatory mechanisms must therefore be considered as well, including cultural transmission processes, developmental processes and genetic selection processes too. These different kinds of explanatory mechanisms are mutually compatible with one another and each is associated with some empirical support, either direct or indirect . For example, recent evidence suggests that the relationship between pathogen prevalence and cultural collectivism may be partially mediated by population-level differences in the frequency of a specific genetic polymorphism in the serotonin transporter gene-regulatory region .
(b) Additional consequences for additional kinds of social behaviour
A more complete understanding of the mechanisms that underlie disease-avoidant behaviour in humans is likely also to be accompanied by a fuller appreciation for the varieties of social behaviour that may serve a disease-avoidant function.
For instance, while chimpanzees have been observed to act aggressively towards a diseased conspecific , no rigorous experimental research has yet explored potential connections between the behavioural immune system and overt aggressive behaviour (in either humans or in other animal species). Aggression is a particularly intriguing behavioural response to perceived infection. To aggress physically against another individual typically requires some sort of approach rather than avoidance behaviour and so if that individual is infected, one's own risk of infection may be temporarily increased during the act of aggression. However, aggression may be an effective means by which to compel the infected individual to leave the immediate area, or to stay at a distance, which may functionally reduce infection risk (for oneself, and for others within the immediate vicinity) in the long term. These considerations suggest that aggression may sometimes be a behavioural outcome of the behavioural immune system, but this behavioural consequence may depend greatly on additional context-specific variables that influence the relative salience of these short-term costs and long-term benefits. In addition, given that aggressive responses to infection risks impose costs primarily on the actual aggressors, while potentially benefiting others within the immediate vicinity, aggressive responses may be highly influenced by social pressures. For example, when the threat of disease is great, normative prohibitions against interpersonal aggression may be relaxed instead, people may be especially likely to encourage others within their social community to act aggressively against anyone who appears to pose an infection risk, or to be especially tolerant (and perhaps even rewarding) of such acts of aggression when they occur.
More broadly, within any social community, one might expect the emergence and persistence of societal norms that encourage specific individuals to engage in any kind of approach-oriented behaviour that increases the infection risk of those particular individuals while simultaneously reducing the infection risk of others in the community. For this reason, perhaps, unusually high levels of prestige and economic resources are accorded to those individuals who habitually undertake the task of attempting to cure others' infections (i.e. physicians and other ‘healers’). People may be especially supportive of these societal inequities under conditions in which they personally feel especially vulnerable to infection.
(c) Implications for human health and well-being
Following the principles of evolutionary psychology, research on the human behavioural immune system is guided by logical considerations regarding specific forms of social behaviour that are likely to have either amplified or reduced individuals' risk of pathogen infection, and thus had implications for reproductive fitness within ancestral social ecologies. The bulk of this research reveals that vulnerability to infection (either real or perceived) has consequences for social cognitive and behavioural outcomes. In contrast, very little research has tested whether these social cognitive and behavioural outcomes actually do have consequences for reproductive fitness, or even whether they have measurable consequences for individuals' health. Of course, there is no necessary reason to assume that, just because a particular behavioural tendency reduced infection risk within the context of ancestral ecologies, that it also reduces infection risk in contemporary ecologies. Indeed, given the substantial changes that have occurred in social ecologies during recent centuries (e.g. demographic shifts, advances in public health infrastructure), there are abundant reasons to presume that behavioural strategies that once functioned as effective defences against pathogen infection may no longer do so . It is an empirical question as to whether the behavioural outcomes produced by the behavioural immune system reduce infection risk in contemporary environments—and if so, which specific behavioural outcomes might do so. There is some evidence pertaining to the infection risks associated with gregariousness and sexual behaviour [26,56], but little is yet known about possible risk-reducing consequences of xenophobia, conformity or other behaviours implicated in this line of enquiry.
To the extent that these behavioural responses do reduce individuals' risk for infection, there are likely to be large-scale epidemiological implications as well. The size, scope and speed with which many infectious diseases become epidemic depends, in part, on the geometric properties of the social networks through which those diseases are transmitted. These geometric properties are themselves products of the social behaviour of the individuals within those networks (the number of sexual partners that people tend to have, the number of acquaintances with whom people interact socially, the frequency of those interactions, etc.). Therefore, as these behaviours vary (in response to vulnerabilities both real and imagined), epidemiological outcomes are likely to vary predictably as well. One intriguing implication is that, because of existing cultural differences in transmission-relevant behavioural attitudes , there may be predictable differences in the epidemic spread of emerging infectious diseases within different cultural populations. More broadly, this research suggests that the psychological mechanisms that characterize the behavioural immune system might fruitfully be included into the mathematical models employed in the service of epidemiological prediction.
Finally, it is worth considering the possibility that there may be important functional connections between psychological mechanisms employed by the behavioural immune system and the biochemical processes through which immunological defences respond to actual infection . One plausible hypothesis is that sensory perception of infection-connoting stimuli may trigger a more aggressive immunological response. Results from two recent experiments provide preliminary support for this hypothesis. One experiment showed that the subjective experience of disgust influences several markers of oral immune function (e.g. increased salivary tumour-necrotizing factor alpha) . The other experiment revealed that immediately after people see photographs depicting symptoms of infectious disease in others, their own white blood cells produce higher levels of the pro-inflammatory cytokine interleukin-6 in response to a bacterial stimulus . The provocative implication (which raises many additional questions requiring further investigation) is that activation of the behavioural immune system may influence the functioning of the ‘real’ immune system too.
Evolution of the immune system in humans from infancy to old age
This article reviews the development of the immune response through neonatal, infant and adult life, including pregnancy, ending with the decline in old age. A picture emerges of a child born with an immature, innate and adaptive immune system, which matures and acquires memory as he or she grows. It then goes into decline in old age. These changes are considered alongside the risks of different types of infection, autoimmune disease and malignancy.
And one man in his time plays many parts,
His acts being seven ages.
More than 1600 genes are involved in innate and adaptive immune responses . These genes are of great importance for sustaining life in a hostile environment. Yet the immune system is relatively immature at birth and has to evolve during a life of exposure to multiple foreign challenges through childhood, via young and mature adulthood (including pregnancy), to the decline of old age (figure 1).
Figure 1. (a) The seven ages of woman. (b) Schematic graph of excess deaths from seasonal or pandemic influenza over the lifetime of an individual represented as number of deaths per 1000 persons (adapted from ). Note that while pregnancy increases the risk of severe influenza, in severe pandemics such as 1918/1919 there were also excess deaths in previously healthy young adults who were not pregnant. (c) Schematic graph of the different arms of the immune response to influenza over the lifetime of an individual.
2. Ontogeny of the immune system in early life
Mewling and puking in the nurse's arms.
In utero, the fetal environment demands that the immune system remains tolerant to maternal alloantigens. After birth, the sudden enormous exposure to environmental antigens, many of them derived from intestinal commensal bacteria, calls for a rapid change to make distinct immune responses appropriate for early life.
(a) The innate immune system
The innate immune system provides an early first line of defence against invading pathogens. The cells involved are neutrophils, monocytes, macrophages and dendritic cells, which all interact with the adaptive immune system. These cells develop and mature during fetal life, but at different times, and the function of all components of innate immunity is weak in newborns compared with later life.
Mature neutrophils are present at the end of the first trimester and steeply increase in number, stimulated by granulocyte-colony-stimulating factor, shortly before birth. Their number then returns to a stable level within days, but they show weak bactericidal functions, poor responses to inflammatory stimuli, reduced adhesion to endothelial cells and diminished chemotaxis . These deficits are more striking in preterm infants, which also have lower serum IgG and complement. Consequently, the newborn, and especially premature infants, have impaired neutrophil functions , putting the child at risk of bacterial infections.
In preterm and newborn infants, classical monocytes and macrophages are also immature. They have reduced TLR4 expression  with impaired innate signalling pathways [6–8], resulting in diminished cytokine responses compared with adults. Consequently, there is poor tissue repair, impaired phagocytosis of potential pathogens and poor secretion of bioactive molecules. However, while there is a reduced frequency of pulmonary macrophages in premature and term infants, adult levels of these cells are reached within days after birth .
Compared with blood from children or adults, cord blood contains fewer myeloid-type dendritic cells (mDC). They express lower cell surface levels of HLA class II, CD80 and CD86 than adult mDC . They secrete low concentrations of IL-12p70 in response to activating innate stimuli . Thus priming of Th1 and CD8 T-cell responses is diminished compared with adults, correlating with an increased susceptibility to infections caused by viruses, Mycobacterium tuberculosis and Salmonella spp. In contrast, newborn mDC stimulated via TLR4 secrete adult-like concentrations of pro-inflammatory cytokines  that promote Th17 immune responses.
Plasmacytoid dendritic cells (pDC) release high concentrations of type I interferon (IFN) in response to TLR7 and TLR9 stimulation in adults. However, newborn pDC are severely limited in secreting interferon α/β upon exposure to different viruses, despite expressing levels of TLR7 and TLR9 that are similar to adults . Consequently, innate immune responses to viruses such as respiratory syncytial virus, herpes simplex virus and cytomegalovirus are poor compared with later in life.
Natural killer (NK) cells in adults restrain viral replication and dissemination before adaptive immunity is established . They are regulated by inhibitory receptors that recognize HLA-A, B, C and E, and therefore contribute to self-tolerance. In early gestation, NK cells are hypo-responsive to target cells lacking major histocompatibility complex (MHC) class I molecules (such as trophoblast ) and are highly susceptible to immune suppression by transforming growth factor-β (TGF-β). NK cytolytic function increases during gestation but is still only half of adult level at birth. Neonatal NK cells are less responsive to activation by IL-2 and IL-15, and produce limited IFN-γ concentrations. However, the cells' threshold for activation is lower, which provides some anti-viral protection .
The three independent pathways that activate the complement system are critical to host defence and inflammation. Complement components facilitate opsonization, are chemo-attractants for innate cells, mediate cell lysis and influence antibody production. Newborn serum concentrations of almost all circulating components are 10–80% lower than in adults , with diminished biological activity. Complement levels increase after birth, with some serum factors reaching adult concentration within a month (e.g. Factor B), but others evolve more slowly . Because infants have low immunoglobulin concentrations, complement effector functions depend on the alternative and lectin-binding activation pathways, triggered by polysaccharides and endotoxins.
Overall, the innate immune system is muted at birth, a price probably paid by the fetus not only to tolerate non-shared maternal antigens but also to ignore the considerable amount of stress and remodelling that takes place during development. This makes the newborn, and particularly the premature baby, relatively susceptible to bacterial and viral infections.
(b) The adaptive immune system
T cells develop in the thymus, which is largest at birth and during the first years of life. Mature single CD4 + and CD8 + positive T cells are first detected in the thymus at week 15 and abundant in the periphery well before birth [17,18]. However, neonatal T cells differ significantly from adult cells, reflecting the fetal life, where exposure to foreign antigens is largely restricted to non-inherited maternal alloantigens. The function of early-life T cells is different from adult T cells. For example, though fetal naive CD4 + T cells respond strongly to alloantigens, they tend to develop towards Foxp3 + CD25 + regulatory T cells (Treg) through the influence of TGF-β , and thus actively promote self-tolerance. Peripheral Treg represent around 3% of total CD4 + T cells at birth  and these cells persist for an extended period of time , giving the early-life immune response an anti-inflammatory profile .
Foreign antigen activation of late fetal or neonatal T cells results in a response skewed towards Th2 immunity , which is reinforced by neonatal dendritic cells and epigenetic features [24,25]. Very early-life adaptive T-cell immunity is thus characterized by tolerogeneic reactivity, reduced allo-antigen recognition and poor responses to foreign antigens.
In the newborn, in addition to conventional T cells that recognize peptide antigens in the context of classical MHC molecules, there are populations of γδ T-cell receptor (TCR)-positive and innate-like αβ TCR-positive T cells. These include functionally competent iNKT cells that rapidly produce IFN, mucosal-associated invariant T (MAIT) cells  and the newly described interleukin-8 (CXCL8)-secreting naive T cells that bridge innate and adaptive immunity . MAIT cells develop in the thymus, but their maturation can take place in fetal mucosal tissues before microbial colonization. The CXCL8-producing T cells produce important effector functions in human newborns as they have the potential to activate antimicrobial neutrophils and γδ T cells. They appear to be particularly active at the mucosal barriers of premature and term infants, though their frequency decreases with age. In contrast to adult blood, where the repertoire of γδ TCR is restricted, neonatal blood γδ T cells display a variety of receptor chain combinations that change with gestation . γδ T cells can produce significant amounts of IFN-γ, after brief polyclonal stimulation, compensating for the immaturity of the more classical Th1-type T-cell response to neonatal infections [28,29].
Two types of B cell arise via distinct developmental pathways . B1 cells spontaneously secrete low-affinity IgM with a limited range of antigen specificities (including common bacterial polysaccharides), have fewer somatic mutations and serve as a first line of defence . B1 cells secrete IL-10 and TGF-β, and thus promote a Th2 response. At birth, B1 cells comprise 40% of peripheral blood B cells and this frequency remains high for a few months . Conventional B cells (designated B2 cells) originate from a multi-linage CD34 + common lymphoid progenitor and generate a broad repertoire of immunoglobulin specificities due to their expression of terminal deoxynucleotidyl transferase, which enhances diversity in V-D-J immunoglobulin gene segment joining. B cells are typically present in secondary lymphoid organs and in the bone marrow, where they contribute to the humoral response of the adaptive immune system.
Most antibody responses, including those to bacterial proteins, bacterial polysaccharides and to polysaccharide–protein conjugate vaccines, are dependent on T-cell help. They rely on interactions between the TCR and the engagement of co-receptors including CD28 and CD40 ligand on Th2 or follicular T helper cells with their corresponding binding partners HLA-peptide, CD80/86 and CD40 on antigen-specific B cells. However, neonatal B cells express low levels of these co-receptors, limiting their capacity to respond . Furthermore, low levels of the receptor for complement C3d fragment (CD21) impede responses to polysaccharide–complement complexes . Together, these features contribute to blunted humoral immune responses with incomplete immunoglobulin class switching , although memory B cells are generated . B cells from neonates and infants aged less than 2 months show decreased somatic hypermutation compared with adults, limiting affinity maturation of antibodies . Finally, there is a failure of early-life bone marrow stromal cells to support long-term plasmablast survival and differentiation to plasma cells, so that any IgG antibodies elicited rapidly decline after immunization, unlike in older children and adults . Hence, the efficiency of the adaptive immune system to respond to T-cell-dependent antigens early is markedly impaired in neonates compared with older children and adults. This physiological behaviour is particularly relevant to vaccination programmes. Together with the impaired innate immunity, the weak Th1 and antibody responses amply explain why neonatal mortality can be high under conditions of increased pathogen exposure.
3. From childhood to adulthood
Then, the whining schoolboy with his satchel
And shining morning face, creeping like snail
The young human child, even as the innate and adaptive immune systems start to mature, is at risk from many pathogenic viruses, bacteria, fungi and parasites. Nevertheless, he or she has a good chance of survival in developed countries. Before there was good nutrition, hygiene and comprehensive vaccination, there was a high mortality in infants and young children. In 1900, the UK infant mortality rate was 140 per 1000, falling to 7 per 1000 by 2000 . This reduction in mortality was proportionally greater in infants and children compared with other age groups . Better prevention and control of infections accounts for most of this fall. However, in many countries, infant mortality rates remain above 50 per 1000, giving some indication of the evolutionary pressure that must have selected a working protective immune system. Furthermore, such pressure has selected the extreme genetic polymorphism in the MHC, which through peptide presentation to T cells and NK cells is a key regulator of almost all immune responses.
The immune system gradually matures during infancy. Critical early protection against many infectious diseases previously experienced by the mother is given by the passive IgG antibody transferred from the mother transplacentally and in milk. Once that fades away, young children become more vulnerable to infections, though by then better armed with the maturing innate and adaptive immune systems. The risks are now much reduced by vaccinations, which stimulate protective immune responses in the maturing immune system. Nevertheless, children may still acquire viral, bacterial and parasitic infections that have to be fought off and controlled by immune responses. Besides promoting recovery, such antigen stimulation results in immunological memory [41,42]. Thus, over time, protection provided by the immune response increases, and young adults suffer fewer infections. This accumulation of immunological memory is an evolving feature of the adaptive immune response. The memory persists into old age  but then may fade.
Besides frank infections and vaccinations, the newborn is exposed to other antigens. He or she comes from a relatively sterile environment in utero and is then rapidly exposed to multiple microbes . The first major exposure to bacteria is during passage through the birth canal, and then as soon as he/she makes oral, skin and respiratory contact with the exterior. From then on, exposure to microorganisms is continuous. Many of the bacteria that colonize the gut and other mucosal sites are essential for healthy life, including digestion of food and acquisition of vital nutrients. They also impact on the development of the immune system .
Approximately 20% of all lymphocytes reside in the gut , exposed to many possible foreign immunogens. Gut immune cells monitor the boundary with a potentially dangerous source of infections. Gut bacteria influence the development of Th17 cells , Treg cells  and memory T cells [48–50]. At birth, nearly all T cells carry the CD45RA glycoprotein, typical of naive T cells, which have never encountered foreign antigen. There are also relatively abundant Tregs within the CD45RA negative CD4 T cells. During childhood, Treg cell numbers decline, and memory Th1, Th17 and Th2 cells gradually increase to equal the number of naive T cells . Although some of these memory T cells could have been stimulated by infections with specific pathogens and by vaccinations, many may be primed by the microbiome, not only in the gut but also in the respiratory tract and skin. These primed memory T cells may respond to subsequent infections through cross-reactions [48,52,53]. For example, adults who have never been exposed to HIV-1 have memory T cells in their repertoire that react with HIV peptides presented at the cell surface by HLA proteins these T cells are likely to be reawakened should HIV infection occur [48,50], similarly to other microbes . The cross-reactivity arises from the discrete short (8–15 amino acids) peptides (epitopes) which fit into peptide-binding grooves on the HLA class I or II molecules at the cell surface and are then recognized by T cells. Within the microbiome sequences, there are numerous perfect and near-perfect matches to known virus peptide epitopes, such as those from HIV-1 [48,50]. These could easily be responsible for generating the memory T cells specific for pathogen epitopes the person has never encountered.
Segmented filamentous bacteria in the gut are necessary for the development of Th17 cells  and Clostridium spp. induce colonic Treg cells [54,55]. Germ-free mice have immunological defects, including fewer Peyers patches, smaller lymphoid follicles and abnormal germinal centres in the small intestine lymphoid tissue . This immuno-deficiency can be corrected in a few days by adding a single mouse with normal gut flora to a cage of germ-free animals [56,57]. Thus animal data support the notion that the microbiome shapes the development of both memory T and B cells.
Similar events occur for B cells. The carbohydrate antigens of the ABO blood groups cross-react with gut bacterial antigens and stimulate IgM antibody responses. Antibodies to the gp41 protein of HIV-1 may be derived from B cells whose antibody receptors cross-react with a protein in Escherichia coli .
As the child grows, the immune repertoire is also shaped by intercurrent infections and vaccinations . Pathogenic infections can be documented by symptomatic illnesses suffered by the child or adult, but for many viruses, such as influenza, infection may be subclinical, but still sufficient to stimulate or boost immune responses . Generally, the protection offered by the immune response, both by antibodies and T cells, is very potent. Most childhood infections happen only once and then protection is lifelong.
The maintenance of long-term B-cell memory is remarkable given that IgG immunoglobulin has a half-life in vivo of around 25 days . The antibody-producing plasma cells that develop during an immune response migrate to the bone marrow, where they are very long lived. In addition, there may be continuous regeneration of memory B cells in contact with persisting antigen and helper T cells. Particulate antigens persist for years in lymph nodes, held by follicular dendritic cells . Antigen persistence and cross-reactive antigens probably help to keep these B cells alive, dividing occasionally and secreting antibodies.
It is remarkable that a mother can transfer sufficient antibody to protect her infant when she was infected 20–30 years previously. The transmission of protective antibody protection from a mother to her child is hugely important, especially in environments where 15% or more infants and children die of infection. Paradoxically, a mother who avoided a dangerous childhood infection, through herd immunity, may actually put her child at risk by being unable to transfer specific protective antibodies.
There are a large number of asymptomatic chronic infections, mostly viral, that provoke immune responses. Examples are cytomegalo virus (CMV), Epstein–Barr virus (EBV) and Mycobacterium tubercolosis (Mtb), but the full list is long and expanding . EBV, CMV and Mtb provoke very strong CD4 and CD8 T-cell responses in humans. The CMV-specific CD8 T-cell response can result in oligoclonal T-cell expansions reaching more than 10% of circulating CD8 T cells. These T cells are important because they control the virus and their depletion, for instance by immunosuppressive therapy, can activate the infection (e.g. Mtb, EBV, CMV), with devastating consequences.
The evolution of antibody responses in B lymphocytes has been reviewed elsewhere in detail . In brief, naive B cells with antibody receptors specific for the immunogen bind antigen in the germinal centre of lymph nodes and receive a partial signal. The bound antigen is internalized and digested in lysosomes. A few resulting peptides bind to the HLA class II molecules of that cell and are then presented on the cell surface where T follicular helper cells with appropriate T-cell receptors respond and deliver further signals, including IL-21, to the B cell. These signals trigger B-cell division, class switching of the antibody genes and somatic hypermutation. B cells that express mutated antibody that binds immunogen with higher affinity are then favoured. Selection for better binding antibodies continues over months, ultimately resulting in high-affinity antibody coming from highly mutated germ line genes. High-affinity antibodies are more effective at neutralizing or opsonizing invading microbes and their pathogenic products.
The somatic hypermutation process does not occur in T cells, even though they have antibody-like T-cell receptor genes, because there is no advantage in having a high-affinity T-cell receptor. The T-cell receptor binding to the peptide–HLA complex on an antigen presenting cells has low affinity. It is enhanced by several co-receptor–ligand pairs that are not antigen-specific, giving the T cell the signal to divide and function.
As a result of an immune challenge, the responding T and B cells may expand transiently to very high numbers , sometimes more than 10% of all circulating T cells, but these decline rapidly as a result of activation-induced cell death and from attrition over a longer time period. Thus as the pathogen is controlled and disappears, some memory T and B cells persist for a long time in numbers that far exceed the number of naive and ‘naive-memory’ T cells that were there before infection.
As the individual gets older, he or she develops an expanding repertoire comprising memory T and B cells triggered by previous infections and vaccinations, but also a naive-memory repertoire shaped by exposure to the microbiome, food antigens and inhaled antigens. Given the great complexity of the T- and B-cell repertoires and a large stochastic element in choosing which cells will respond to a given stimulus, and somatic mutations in B cells, the precise composition will differ in each individual, even in monozygotic twins . Add to this considerable genetic variability in how individuals respond, determined by the highly polymorphic HLA genes  and by the genes of innate immunity, and it is not surprising that the immune responses of any single adult vary considerably.
It is beyond the scope of this review to explore the immunology of pregnancy in detail (reviewed in [68,69]). However, successful reproduction is of central evolutionary importance and there are immunological issues. How the newborn retains mechanisms by which the fetus minimizes its immune responses to the mother has been discussed above. A bigger puzzle is how the mother tolerates a semi-allogeneic graft without rejecting it and without the immunosuppression necessary to accept an organ transplant . There are features at the trophoblast maternal interface at the site of initial implantation and in the placenta that subvert the normal graft rejection immune response. These include expression only of non-polymorphic non-classical HLA antigens on the trophoblast , local immune suppression mediated by infiltrating NK cells , monocytes and T regulatory cells [69,73], and inhibition of T-cell activation by tryptophan catabolism . Around the time of implantation, a local inflammatory response sets up the stable placental site . There is evidence that the mother changes the balance of her T-cell responses to Th2 rather than Th1 . Thus pregnant women can show remissions of autoimmune disease , and are more susceptible to severe complications of influenza  and some other infections. This immune modulation, necessary for the well-being of the fetus, can occasionally be harmful to the mother.
(b) Malignancy and autoimmunity
The primary role of the immune system is probably to protect against infections. Other roles such as destruction of mutated cells may be very important, though more so in old age after reproduction. Many tumours turn off T cells specific for tumour antigens by binding to ‘check-point’ receptors such as PD-1 or CTLA4, and new treatments that block these receptor–ligand interactions have great therapeutic potential [77,78]. However, the side effects of such therapy and of the passive transfer of anti-cancer T cells include autoimmune reactions, suggesting a balance between anti-self-immune reactions preventing cancer and causing autoimmunity . In adult life, the balance usually works, but one-third of Western humans develop cancer, usually later in life, while 5–10% develop clinical autoimmune disease, so the balance is finely set and may shift over time. The fading immune system in old age (see below) may ameliorate autoimmunity but at the expense of increased cancer risk.
Microorganisms cause about a quarter of all cancers (e.g. EBV, hepatitis B and C viruses, human papilloma virus and Helicobacter pylori). Specific T-cell responses normally hold these microbes in check. However, if immunity is impaired through ageing (see below), immunosuppressive therapy or certain infections, particularly HIV-1, these cancers emerge .
Therefore, having developed a fully effective immune response in early childhood, this matures as memory accumulates and maintains the health of the individual during critical periods of life, including child bearing. It not only protects against potentially lethal infections but also controls a number of persisting infections, some of which have the potential to cause cancer. It can also deal with mutant cells that have potential for becoming malignant. It can be over-reactive and cause autoimmune disease or allergy, a price paid for the overall benefit.
Your Environment Is Cleaner. Your Immune System Has Never Been So Unprepared.
Don’t laugh. Scientifically, it’s an interesting question.
Should your children pick their noses? Should your children eat dirt? Maybe: Your body needs to know what immune challenges lurk in the immediate environment.
Should you use antibacterial soap or hand sanitizers? No. Are we taking too many antibiotics? Yes.
“I tell people, when they drop food on the floor, please pick it up and eat it,” said Dr. Meg Lemon, a dermatologist in Denver who treats people with allergies and autoimmune disorders.
“Get rid of the antibacterial soap. Immunize! If a new vaccine comes out, run and get it. I immunized the living hell out of my children. And it’s O.K. if they eat dirt.”
Dr. Lemon’s prescription for a better immune system doesn’t end there. “You should not only pick your nose, you should eat it,” she said.
She’s referring, with a facetious touch, to the fact our immune system can become disrupted if it doesn’t have regular interactions with the natural world.
“Our immune system needs a job,” Dr. Lemon said. “We evolved over millions of years to have our immune systems under constant assault. Now they don’t have anything to do.”
She isn’t alone. Leading physicians and immunologists are reconsidering the antiseptic, at times hysterical, ways in which we interact with our environment.
Why? Let us turn to 19th-century London.
The British Journal of Homeopathy, volume 29, published in 1872, included a startlingly prescient observation: “Hay fever is said to be an aristocratic disease, and there can be no doubt that, if it is not almost wholly confined to the upper classes of society, it is rarely, if ever, met with but among the educated.”
Hay fever is a catchall term for seasonal allergies to pollen and other airborne irritants. With this idea that hay fever was an aristocratic disease, British scientists were on to something.
More than a century later, in November 1989, another highly influential paper was published on the subject of hay fever. The paper was short, less than two pages, in BMJ, titled “Hay Fever, Hygiene, and Household Size.”
The author looked at the prevalence of hay fever among 17,414 children born in March 1958. Of 16 variables the scientist explored, he described as “most striking” an association between the likelihood that a child would get hay fever allergy and the number of his or her siblings.
It was an inverse relationship, meaning the more siblings the child had, the less likely it was that he or she would get the allergy. Not just that, but the children least likely to get allergies were ones who had older siblings.
The paper hypothesized that “allergic diseases were prevented by infection in early childhood, transmitted by unhygienic contact with older siblings, or acquired prenatally from a mother infected by contact with her older children.
“Over the past century declining family size, improvements in household amenities, and higher standards of personal cleanliness have reduced the opportunity for cross infection in young families,” the paper continued. “This may have resulted in more widespread clinical expression of atopic disease, emerging in wealthier people, as seems to have occurred for hay fever.”
This is the birth of the hygiene hypothesis. The ideas behind it have since evolved and expanded, but it provides profound insight into a challenge that human beings face in our relationship with the modern world.
Our ancestors evolved over millions of years to survive in their environments. For most of human existence, that environment was characterized by extreme challenges, like scarcity of food, or food that could carry disease, as well as unsanitary conditions and unclean water, withering weather, and so on. It was a dangerous environment, a heck of a thing to survive.
At the center of our defenses was our immune system, our most elegant defense. The system is the product of centuries of evolution, as a river stone is shaped by water rushing over it and the tumbles it experiences on its journey downstream.
Late in the process, humans learned to take steps to bolster our defenses, developing all manner of customs and habits to support our survival. In this way, think of the brain — the organ that helps us develop habits and customs — as another facet of the immune system.
We used our collective brains to figure out effective behaviors. We started washing our hands and took care to avoid certain foods that experience showed could be dangerous or deadly. In some cultures, people came to avoid pork, which we now know is highly susceptible to trichinosis in others, people banned meats, which we later learned may carry toxic loads of E. coli and other bacteria.
Ritual washing is mentioned in Exodus, one of the earliest books in the Bible: “So they shall wash their hands and their feet, that they die not.”
Our ideas evolved, but for the most part, the immune system did not. This is not to say that it didn’t change. The immune system responds to our environment. When we encounter various threats, our defenses learn and then are much more able to deal with that threat in the future. In that way, we adapt to our environment.
We survived over tens of thousands of years. Eventually, we washed our hands, swept our floors, cooked our food, avoided certain foods altogether. We improved the hygiene of the animals we raised and slaughtered for food.
Particularly in the wealthier areas of the world, we purified our water, and developed plumbing and waste treatment plants we isolated and killed bacteria and other germs.
The immune system’s enemies list was attenuated, largely for the good. Now, though, our bodies are proving that they cannot keep up with this change. We have created a mismatch between the immune system — one of the longest surviving and most refined balancing acts in the world — and our environment.
Thanks to all the powerful learning we’ve done as a species, we have minimized the regular interaction not just with parasites but even with friendly bacteria and parasites that helped to teach and hone the immune system — that “trained” it. It doesn’t encounter as many bugs when we are babies. This is not just because our homes are cleaner, but also because our families are smaller (fewer older children are bringing home the germs), our foods and water cleaner, our milk sterilized. Some refer to the lack of interaction with all kinds of microbes we used to meet in nature as the “old friends mechanism.”
What does the immune system do when it’s not properly trained?
It can overreact. It becomes aggrieved by things like dust mites or pollen. It develops what we called allergies, chronic immune system attacks — inflammation — in a way that is counterproductive, irritating, even dangerous.
The percentage of children in the United States with a food allergy rose 50 percent between 1997–1999 and 2009–2011, according to the Centers for Disease Control and Prevention. The jump in skin allergies was 69 percent during that period, leaving 12.5 percent of American children with eczema and other irritations.
Food and respiratory allergies rose in tandem with income level. More money, which typically correlates with higher education, has meant more risk of allergy. This may reflect differences in who reports such allergies, but it also springs from differences in environment.
These trends are seen internationally, too. Skin allergies “doubled or tripled in industrialized countries during the past three decades, affecting 15–30 percent of children and 2–10 percent of adults,” according to a paper citing research from the Journal of Allergy and Clinical Immunology.
By 2011, one in four children in Europe had an allergy, and the figure was on the rise, according to a report by the World Allergy Organization. Reinforcing the hygiene hypothesis, the paper noted that migration studies have shown that children born overseas have lower levels of some types of both allergy and autoimmunity than migrants whose children are born in the United States.
There are related trends in inflammatory bowel disease, lupus, rheumatic conditions and, in particular, celiac disease. The last results from the immune’s system overreacting to gluten, a protein in wheat, rye and barley. This attack, in turn, damages the walls of the small intestine.
This might sound like a food allergy, but it is different in part because of the symptoms. In the case of an autoimmune disorder like this one, the immune system attacks the protein and associated regions.Image
Allergies can generate a more generalized response. A peanut allergy, for instance, can lead to inflammation in the windpipe, known as anaphylaxis, which can cause strangulation.
In the case of both allergy and autoimmune disorders, though, the immune system reacts more strongly than it otherwise might, or than is healthy for the host (yeah, I’m talking about you).
This is not to say that all of these increases are due to better hygiene, a drop in childhood infection, and its association with wealth and education. There have been many changes to our environment, including new pollutants. There are absolutely genetic factors as well.
But the hygiene hypothesis — and when it comes to allergy, the inverse relationship between industrialized processes and health — has held up remarkably well.
As our bodies strive for balance, Madison Avenue has made a full-court press for greater hygiene, sometimes to our detriment.
We’re fed a steady diet of a hygiene-related marketing that began in the late 1800s, according to a novel study published in 2001 by the Association for Professionals in Infection Control and Epidemiology. Scientists at Columbia University who did the research were trying to understand how we became so enamored of soap products.
The Sears catalog in the early 1900s heavily advertised “ammonia, Borax, and laundry and toilet soap.”
“During the early to mid-1900s, soap manufacturing in the United States increased by 44 percent,” coinciding with “major improvements in water supply, refuse disposal and sewage systems.”
The marketing trailed off in the 1960s and 1970s as antibiotics and vaccines were understood to be the answer to infectious agents, with less emphasis on “personal responsibility.”
But then, starting in the late 1980s, the market for such hygiene products — home and personal — surged 81 percent. The authors cite a “return of public concern for protection against infectious disease,” and it’s hard not to think of AIDS as part of that attention. If you’re in marketing, never waste a crisis, and the messages had an impact.
The study cites a Gallup poll from 1998 that found that 66 percent of adults worried about virus and bacteria, and 40 percent “believed these microorganisms were becoming more widespread.” Gallup also reported that 33 percent of adults “expressed the need for antibacterial cleansers to protect the home environment,” and 26 percent believed they were needed to protect the body and skin.
They were wrong. And even doctors have been wrong.
They have vastly overprescribed antibiotics. These may be a huge boon to an immune system faced with an otherwise deadly infection. But when used without good reason, the drugs can wipe out healthy microbes in our gut and cause bacteria to develop defenses that make them even more lethal.
A scientist who led efforts at the World Health Organization to develop global policy to limit use of antibiotics told me that, philosophically, this is a lesson that runs counter to a century of marketing: We’re not safer when we try to eliminate every risk from our environment.
“We have to get away from the idea of annihilating these things in our local environment. It just plays upon a certain fear,” said the scientist, Dr. Keiji Fukuda.
Has much of our hygiene been practical, valuable, life-preserving? Yes.
Have we overcorrected? At times. Should you pick your nose? Or put another way: Might that urge to pick be part of a primitive strategy to inform your immune system about the range of microbes in your environment, give this vigilant force activity, and train your most elegant defense?
In short, from a cultural standpoint, you still probably shouldn’t pick — not in public. But it is a surprisingly fair scientific question.
Passive immunity is provided when a person is given antibodies to a disease rather than producing them through his or her own immune system.
A newborn baby acquires passive immunity from its mother through the placenta. A person can also get passive immunity through antibody-containing blood products such as immune globulin , which may be given when immediate protection from a specific disease is needed. This is the major advantage to passive immunity protection is immediate, whereas active immunity takes time (usually several weeks) to develop.
However, passive immunity lasts only for a few weeks or months. Only active immunity is long-lasting.
Published October 18. 2020 12:05AM
Ben Guarino, The Washington Post
Early in the coronavirus outbreak, hospital data from China revealed a startling disparity: COVID-19, the disease caused by the virus, was killing far more men than women.
That difference persisted in other Asian countries, such as South Korea, as well as in European countries, such as Italy. Then, it appeared in the United States.
By mid-October, the coronavirus had killed almost 17,000 more American men than women, according to data from the Centers for Disease Control and Prevention. For every 10 women claimed by the disease in the United States, 12 men have died, found an analysis by Global Health 50/50, a U.K.-based initiative to advance gender equality in health care.
That disparity was one of many alarming aspects of the new virus. It bewildered those unfamiliar with the role of gender in disease.
But the specialized group of researchers who study that relationship was not surprised. It prepared an array of hypotheses. One possible culprit was male behavior. Perhaps men were more likely to be exposed to the virus due to social factors a disproportionately male workforce, for instance, could place more men in contact with infected people. Or men's lungs might be more vulnerable because they were more likely to smoke in the earliest countries to report the differences.
What has become more evident, 10 months into this outbreak, is that men show comparatively weaker immune responses to coronavirus infections, which may account for those added deaths.
"If you look at the data across the world, there are as many men as women that are infected. It's just the severity of disease that is stronger in most populations in men," Franck Mauvais-Jarvis, a Tulane University physician who studies gender differences in such diseases as diabetes. In such cases, biology can help explain why.
- The male immune response. Women generally have stronger immune systems, thanks to sex hormones, as well as chromosomes packed with immune-related genes. About 60 genes on the X chromosome are involved in immune function, Johns Hopkins University microbiologist Sabra Klein told The Washington Post in April. People with two X chromosomes can benefit from the double helping of some of those genes.
Akiko Iwasaki, who studies immune defenses against viruses at Yale University, wanted to see how sex differences might play out in coronavirus infections. She and her colleagues cast a proverbial net into the immune system to fish out schools of microscopic fighters.
"We did a holistic look at everything we can measure immunologically," Iwasaki said, listing a litany of the molecules and cells that form the body's bulwark against pathogens: "cytokines, chemokines, T cells, B cells, neutrophils. Everything that we had access to."
In male patients, the T-cell response was weaker, the scientists found. Not only do T cells detect infected cells and kill them, they also help direct the antibody response. "It's like a master regulator of immune response. And when you have a drop in T cells, or in their ability to become activated, you basically lose the conductor of an orchestra," Iwasaki said.
The power of the immune system wanes as people age, regardless of sex. But what is a gentle decline for women is an abrupt dive off a cliff for men: Iwasaki's work indicates the T-cell response of men in their 30s and 40s is equivalent to that of a woman in her 90s.
And T cells aren't the only immune feature disproportionately impaired in men. Another paper, published in September in PLOS Biology, examined anonymous human genetic material collected along with viruses in nasal swabs.
That study found throttled defense signals in men. When a cell detects a virus, it performs the molecular equivalent of yanking the fire alarm, said one of the study's author, Nicole Lieberman, a research scientist at the University of Washington. That alarm is manifest in genetic messengers, called RNA, which react almost immediately.
The reaction should cause cells to churn out the first lines of defense, such as interferons, immune system molecules that, as the name suggests, interfere with the virus's ability to reproduce. Other molecules summon specialized immune cells to destroy the pathogens. "You want the fire alarm to go off for long enough that you can get the fire department there," Lieberman said.
Lieberman and her co-authors, however, found that in men and some older populations, the fire alarm shuts off early - maybe even before the firefighters have arrived. "That, I think, is the functional consequence, potentially, of what we're seeing here," she said.
- Harmful autoantibodies. Not only is the immune system in men weaker, but in some severe cases of the coronavirus, it may hobble itself. A study of nearly 1,000 patients with life-threatening COVID-19, published in Science in September, found evidence of molecular self-sabotage. Immune system fighters were acting against the body's defenses, like rebellious castle guards splintering their own gates. This flaw was much more prevalent in men than women.
Specifically, the researchers detected what are called autoantibodies, molecules that bind and neutralize parts of the immune system. Those neutralizers disabled a subset of defender molecules known as type-1a interferon. Simply put, having autoantibodies led to more viral replication.
Ninety-five of 101 people with autoantibodies against interferon were male. "Somehow males are probably more prone to develop such autoantibodies, but we do not know why," said study author Petter Brodin, a pediatrician at Sweden's Karolinska Institute who studies the immune system.
Interferon molecules come in several types, so it's possible these patients could be treated with another flavor of interferon, Brodin said. But that may be difficult, he acknowledged, because interferons are most helpful early in the course of an infection, before the disease progresses to life-threatening stages.
The lack of killer T cells, coupled with neutralizing antibodies, is "like a double whammy," Iwasaki said, "that would then ultimately increase the viral load in these men."
What's unusual about this result is that most autoantibody immune disorders appear in women, as is the case with the chronic disease lupus.
Iwasaki's research is examining whether female immune systems may play a role in people with long-lasting COVID-19, nicknamed long-haulers.
"There are thousands of people suffering from chronic symptoms," which may be debilitating, Iwasaki said. Many long-haulers are young and the majority of them, though not all, are women.
Beyond these biological differences, it would be simplistic to ignore how gender's other aspects, such as behavior and social norms, may also influence the pandemic.
Broadly speaking, men may be less likely to be worried about COVID-19 than women, fitting the pattern that women generally treat health risks more seriously. Women took a more cautious approach to the disease, a recent poll found, expressing more concern they could return to workplaces safely. Women are also more likely to follow expert advice such as mask-wearing and social distancing, according to another study that included surveys and observations of pedestrians' behavior in New York, Connecticut and New Jersey.
Sarah Hawkes, a professor of global public health at University College London who, with her husband, co-directs Global Health 50/50, said that the image of men as risk-takers extends back hundreds of years to John Graunt, one of the first people to participate in the field now known as epidemiology.
After he reviewed England's death records, Graunt postulated in 1662 that "men, being more intemperate then women, die as much by reason of their Vices" - that is, male behavior was to blame. Hawkes argues that "350 years later," Graunt's point still stands. "It is undoubtedly a mixture of both biology and behavior" responsible for the health differences in men and women, she said.
The share of coronavirus deaths in women also rises with their share of the full-time workforce, according to a report by University of Oxford economist Renee Adams that used Global Health 50/50 data.
"The more you have women participating in the workforce, the smaller your sex difference becomes," Hawkes said. That lines up with gender inequalities - men are more likely to work in environments where they are exposed to air pollution and other harms, Hawkes said. When women start to enter those traditionally masculine spaces, she said, it "turns out, women can get as sick as men."
The gender disparities discovered in the response to COVID-19 have sparked a surge of interest in such differences more broadly. "Almost nobody, apart from the people working in the field, were interested in that difference between men and women in disease until February or March," when the first results showed that more men were dying, Mauvais-Jarvis said.
Even agencies at the forefront of public health, such as the CDC, were initially slow to reveal sex-disaggregated coronavirus data, Hawkes said. The U.K. public health surveillance system was similarly late. Hawkes took those delays as a sign of just how unimportant people considered this data, since it is so readily available: When people die, their death certificates state whether they were male, female or, in some places, nonbinary.
The CDC data finally made that information accessible in mid-April. The male-skewed patterns revealed in those deaths conform to what was seen in earlier outbreaks of Middle East respiratory syndrome (MERS) and severe acute respiratory syndrome (SARS), both within the family of coronaviruses. And it is in line with other viral responses. "We know that women develop much better antibody response to flu vaccines," Iwasaki said.
Some of those experts are hoping to capitalize on this moment to shine a spotlight on other gender differences in health. The coronavirus, after all, isn't the only problem to afflict men and women unequally - so, too, do cancer, asthma, heart disease and other common illnesses, as Mauvais-Jarvis noted in a recent paper in the Lancet.
"The kinds of differences that we're seeing and outcomes in COVID-19 are not unexpected. They're not exceptional," Hawkes said. If there's surprise, it only demonstrates the widespread underestimation of the differences in men and women that persist even among physicians, she said.
Mauvais-Jarvis referred to this faulty approach as "bikini medicine" - in which clinicians view female patients as interchangeable with male ones, except for the organs covered by swimwear.
The coronavirus has helped accelerate the trend away from that outdated view. The "one positive that's come out of the pandemic," Hawkes said, is the sudden realization that gendered social factors and biology "may have a relationship with your life expectancy, your experience with illness, your risk of illness. It has made that conversation a little bit more real."
Immunology Is Where Intuition Goes to Die
Which is too bad because we really need to understand how the immune system reacts to the coronavirus.
Updated at 10:36 a.m. ET on August 5, 2020.
There’s a joke about immunology, which Jessica Metcalf of Princeton recently told me. An immunologist and a cardiologist are kidnapped. The kidnappers threaten to shoot one of them, but promise to spare whoever has made the greater contribution to humanity. The cardiologist says, “Well, I’ve identified drugs that have saved the lives of millions of people.” Impressed, the kidnappers turn to the immunologist. “What have you done?” they ask. The immunologist says, “The thing is, the immune system is very complicated …” And the cardiologist says, “Just shoot me now.”
The thing is, the immune system is very complicated. Arguably the most complex part of the human body outside the brain, it’s an absurdly intricate network of cells and molecules that protect us from dangerous viruses and other microbes. These components summon, amplify, rile, calm, and transform one another: Picture a thousand Rube Goldberg machines, some of which are aggressively smashing things to pieces. Now imagine that their components are labeled with what looks like a string of highly secure passwords: CD8+, IL-1β, IFN-γ. Immunology confuses even biology professors who aren’t immunologists—hence Metcalf’s joke.
Even the word immunity creates confusion. When immunologists use it, they simply mean that the immune system has responded to a pathogen—for example, by producing antibodies or mustering defensive cells. When everyone else uses the term, they mean (and hope) that they are protected from infection—that they are immune. But, annoyingly, an immune response doesn’t necessarily provide immunity in this colloquial sense. It all depends on how effective, numerous, and durable those antibodies and cells are.
Immunity, then, is usually a matter of degrees, not absolutes. And it lies at the heart of many of the COVID-19 pandemic’s biggest questions. Why do some people become extremely ill and others don’t? Can infected people ever be sickened by the same virus again? How will the pandemic play out over the next months and years? Will vaccination work?
To answer these questions, we must first understand how the immune system reacts to SARS-CoV-2 coronavirus. Which is unfortunate because, you see, the immune system is very complicated.
It works, roughly, like this.
The first of three phases involves detecting a threat, summoning help, and launching the counterattack. It begins as soon as a virus drifts into your airways, and infiltrates the cells that line them.
When cells sense molecules common to pathogens and uncommon to humans, they produce proteins called cytokines. Some act like alarms, summoning and activating a diverse squad of white blood cells that go to town on the intruding viruses—swallowing and digesting them, bombarding them with destructive chemicals, and releasing yet more cytokines. Some also directly prevent viruses from reproducing (and are delightfully called interferons). These aggressive acts lead to inflammation. Redness, heat, swelling, soreness—these are all signs of the immune system working as intended.
This initial set of events is part of what’s called the innate immune system. It’s quick, occurring within minutes of the virus’s entry. It’s ancient, using components that are shared among most animals. It’s generic, acting in much the same way in everyone. And it’s broad, lashing out at anything that seems both nonhuman and dangerous, without much caring about which specific pathogen is afoot. What the innate immune system lacks in precision, it makes up for in speed. Its job is to shut down an infection as soon as possible. Failing that, it buys time for the second phase of the immune response: bringing in the specialists.
Amid all the fighting in your airways, messenger cells grab small fragments of virus and carry these to the lymph nodes, where highly specialized white blood cells—T-cells—are waiting. The T-cells are selective and preprogrammed defenders. Each is built a little differently, and comes ready-made to attack just a few of the zillion pathogens that could possibly exist. For any new virus, you probably have a T-cell somewhere that could theoretically fight it. Your body just has to find and mobilize that cell. Picture the lymph nodes as bars full of grizzled T-cell mercenaries, each of which has just one type of target they’re prepared to fight. The messenger cell bursts in with a grainy photo, showing it to each mercenary in turn, asking: Is this your guy? When a match is found, the relevant merc arms up and clones itself into an entire battalion, which marches off to the airways.
Some T-cells are killers, which blow up the infected respiratory cells in which viruses are hiding. Others are helpers, which boost the rest of the immune system. Among their beneficiaries, these helper T-cells activate the B-cells that produce antibodies—small molecules that can neutralize viruses by gumming up the structures they use to latch on to their hosts. Roughly speaking—and this will be important later—antibodies mop up the viruses that are floating around outside our cells, while T-cells kill the ones that have already worked their way inside. T-cells do demolition antibodies do cleanup.
Both T-cells and antibodies are part of the adaptive immune system. This branch is more precise than the innate branch, but much slower: Finding and activating the right cells can take several days. It’s also long-lasting: Unlike the innate branch of the immune system, the adaptive one has memory.
After the virus is cleared, most of the mobilized T-cell and B-cell forces stand down and die off. But a small fraction remain on retainer—veterans of the COVID-19 war of 2020, bunkered within your organs and patrolling your bloodstream. This is the third and final phase of the immune response: Keep a few of the specialists on tap. If the same virus attacks again, these “memory cells” can spring into action and launch the adaptive branch of the immune system without the usual days-long delay. Memory is the basis of immunity as we colloquially know it—a lasting defense against whatever has previously ailed us.
This account is what should happen when the new coronavirus enters the body, based on general knowledge about the immune system and how it reacts to other respiratory viruses. But what actually happens? Well … sigh … the thing is, the immune system is very complicated.
In general, the immune system’s reaction to SARS-CoV-2 is “what I would expect if you told me there was a new respiratory infection,” says Shane Crotty from the La Jolla Institute of Immunology. The innate immune system switches on first, and the adaptive immune system follows suit. In several studies, most people who are infected develop reasonable levels of coronavirus-specific T-cells and antibodies. “The bottom line is that there are no big surprises,” says Sarah Cobey, an epidemiologist from the University of Chicago.
Still, “any virus that can make people sick has to have at least one good trick for evading the immune system,” Crotty says. The new coronavirus seems to rely on early stealth, somehow delaying the launch of the innate immune system, and inhibiting the production of interferons—those molecules that initially block viral replication. “I believe this [delay] is really the key in determining good versus bad outcomes,” says Akiko Iwasaki, an immunologist at Yale. It creates a brief time window in which the virus can replicate unnoticed before the alarm bells start sounding. Those delays cascade: If the innate branch is slow to mobilize, the adaptive branch will also lag.
Many infected people still clear the virus after a few weeks of nasty symptoms. But others don’t. Maybe they initially inhaled a large dose of virus. Maybe their innate immune systems were already weakened through old age or chronic disease. In some cases, the adaptive immune system also underperforms: T-cells mobilize, but their levels recede before the virus is vanquished, “almost causing an immunosuppressed state,” Iwasaki says. This dual failure might allow the virus to migrate deeper into the body, toward the vulnerable cells of the lungs, and to other organs including the kidneys, blood vessels, and the gastrointestinal and nervous systems. The immune system can’t constrain it, but doesn’t stop trying. And that’s also a problem.
Immune responses are inherently violent. Cells are destroyed. Harmful chemicals are unleashed. Ideally, that violence is targeted and restrained as Metcalf puts it, “Half of the immune system is designed to turn the other half off.” But if an infection is allowed to run amok, the immune system might do the same, causing a lot of collateral damage in its prolonged and flailing attempts to control the virus.
This is apparently what happens in severe cases of COVID-19. “If you can’t clear the virus quickly enough, you’re susceptible to damage from the virus and the immune system,” says Donna Farber, a microbiologist at Columbia. Many people in intensive-care units seem to succumb to the ravages of their own immune cells, even if they eventually beat the virus. Others suffer from lasting lung and heart problems, long after they are discharged. Such immune overreactions also happen in extreme cases of influenza, but they wreak greater damage in COVID-19.
There’s a further twist. Normally, the immune system mobilizes different groups of cells and molecules when fighting three broad groups of pathogens: viruses and microbes that invade cells, bacteria and fungi that stay outside cells, and parasitic worms. Only the first of these programs should activate during a viral infection. But Iwasaki’s team recently showed that all three activate in severe COVID-19 cases. “It seems completely random,” she says. In the worst cases, “the immune system almost seems confused as to what it’s supposed to be making.”
No one yet knows why this happens, and only in some people. Eight months into the pandemic, the variety of COVID-19 experiences remains a vexing mystery. It’s still unclear, for example, why so many “long-haulers” have endured months of debilitating symptoms. Many of them have never been hospitalized, and so aren’t represented in existing studies that have measured antibody and T-cell responses. David Putrino of Mount Sinai tells me that he surveyed 700 long-haulers and a third had tested negative for antibodies, despite having symptoms consistent with COVID-19. It’s unclear if their immune systems are doing anything differently when confronted with the coronavirus.
We should expect such mysteries to build. The immune system’s reaction to the virus is a matter of biology, but the range of reactions we actually see is also influenced by politics. Bad decisions mean more cases, which means a wider variety of possible immune responses, which means a higher prevalence of rare events. In other words, the worse the pandemic gets, the weirder it will get.
A few patterns offer easier possible explanations. “Kids have very trigger-happy innate immune systems,” says Florian Krammer of Mount Sinai’s Icahn School of Medicine, which might explain why they rarely suffer severe infections. Elderly people are less fortunate. They also have smaller standing pools of T-cells to draw from, as if the mercenary-filled bar from the earlier metaphor is only sparsely packed. “It takes longer for the adaptive response to mobilize,” Farber says.
There are also preliminary hints that some people might have a degree of preexisting immunity against the new coronavirus. Four independent groups of scientists—based in the U.S., Germany, the Netherlands, and Singapore—have now found that 20 to 50 percent of people who were never exposed to SARS-CoV-2 nonetheless have significant numbers of T-cells that can recognize it. These “cross-reactive” cells likely emerged when their owners were infected by other, related coronaviruses, including the four mild ones that cause a third of common colds, and the many that infect other animals.
But Farber cautions that having these cross-reactive T-cells “tells you absolutely nothing about protection.” It’s intuitive to think they would be protective, but immunology is where intuition goes to die. The T-cells might do nothing. There’s an outside chance that they could predispose people to more severe disease. We can’t know for sure without recruiting lots of volunteers, checking their T-cell levels, and following them over a long period of time to see who gets infected—and how badly.
Even if the cross-reactive cells are beneficial, remember that T-cells act by blowing up infected cells. As such, they’re unlikely to stop people from getting infected in the first place, but might reduce the severity of those infections. Could this help to explain why, politics aside, some countries had an easier time with COVID-19 than others? Could it explain why some people incur only mild symptoms? “You can go pretty crazy pretty quickly with the speculations,” says Crotty, who co-led one of the studies that identified these cross-reactive cells. “A lot of people have latched onto this and said it could explain everything. Yes, it could! Or it could explain nothing. It’s a really frustrating situation to be in.”
“I wish it wasn’t,” he adds, “but the immune system is really complicated.”
One of the most pressing mysteries is what happens after you’re infected—and whether you could be again. Crucially, researchers still don’t know how much protection the leftover antibodies, T-cells, and memory cells might offer against COVID-19, or even how to measure that.
In July, a team of British researchers released a study showing that many COVID-19 patients lose substantial levels of their coronavirus-neutralizing antibodies after a few months. An earlier Chinese study, published in June, found similar results. Both prompted cascades of alarming headlines, which raised concerns that people could be infected repeatedly, or even that a vaccine—many of which work by readying neutralizing antibodies—won’t provide long-term protection. But many of the immunologists I spoke with weren’t too concerned, because—and reassuringly this time—the immune system is really complicated.
First, declines are expected. During an infection, antibodies are produced by two different groups of B-cells. The first group is fast and short-lived, and quickly unleashes a huge antibody tsunami before dying off. The second group is slower but long-lasting, and produces gentler antibody swells that continuously wash over the body. The transition from the first group to the second means that antibody levels usually decline over the course of an infection. “There’s nothing scary about it,” Krammer says.
Taia Wang of Stanford is a little less sanguine. She tells me several studies, including upcoming ones, consistently show that many people seem to lose their neutralizing antibodies after a couple of months. “If you asked me to guess six months ago, I would have thought that they would last longer,” she says. “The durability is not what we’d like.”
But “the fact that you don’t have measurable antibodies doesn’t mean that you aren’t immune,” Iwasaki says. T-cells could continue to provide adaptive immunity even if the antibodies tap out. Memory B-cells, if they persist, could quickly replenish antibody levels even if the current stocks are low. And, crucially, we still don’t know how many neutralizing antibodies you need to be protected against COVID-19.
Wang agrees: “There’s a common notion that antibody quantity is all that matters, but it’s more complicated than that,” she says. “The quality of the antibody is as important.” Quality might be defined by which part of the virus the antibodies stick to, or how well they stick. Indeed, many people who recover from COVID-19 have low levels of neutralizing antibodies overall, but some of them neutralize very well. “Quantity is easier to measure,” Wang adds. “There are more ways to characterize quality and we don’t know which ones are relevant.” (This problem is even worse for T-cells, which are much harder than antibodies to isolate and analyze.)
These uncertainties strengthen the need for large, careful vaccine trials: Right now it’s hard to know whether the promising signs in early trials will actually lead to substantial protection in practice. (Developing and deploying vaccines is a subject for another piece, which my colleague Sarah Zhang has written.) Scientists are trying to work out how to measure COVID-19 immunity by studying large groups of people who have either been infected naturally or taken part in a vaccine trial. Researchers will repeatedly measure and analyze the volunteers’ antibodies and T-cells over time, noting if any of them become infected again. Krammer expects that results will take a few months, or possibly until the end of the year. “There’s no way to speed that up,” he says. Because … well, you know.
In the meantime, anecdotal reports have described alleged reinfections—people who apparently catch COVID-19 a second time, and who test positive for the coronavirus again after months of better health. Such cases are concerning, but hard to interpret. Viral RNA—the genetic material that diagnostic tests detect—can stick around for a long time, and people can test positive for months after they’ve cleared the actual virus. If someone like that caught the flu and went to their doctor, they might get tested for coronavirus again, get a positive result, and be mistakenly treated as a case of reinfection. “It’s really hard to prove reinfection unless you sequence the genes of the virus” both times, Iwasaki says. “No one has that data, and it’s unreasonable to expect.”
Immunity lasts a lifetime for some diseases—chickenpox, measles—but eventually wears off for many others. As the pandemic drags on, we should expect at least a few instances in which people who’ve beaten COVID-19 must beat it again. So far, the fact that reinfections are still the subject of smattered anecdotes suggests that “it’s happening at a very low rate, if at all,” Cobey says. But remember: A bigger pandemic is a weirder pandemic. When there are almost 5 million confirmed cases, something that occurs just 0.1 percent of the time will still affect 5,000 people.
If people endure a second bout with COVID-19, the outcome is again hard to call. For some diseases, like dengue, an antibody response to one infection can counterintuitively make the next infection more severe. So far, there’s no evidence this happens with SARS-CoV-2, says Krammer, who expects that any reinfections would be milder than the first ones. That’s because the coronavirus has a longer incubation time—a wider window between infection and symptoms—than, say, the flu. That could conceivably provide more time for memory cells to mobilize a new force of antibodies and T-cells. “Even if there’s some immunity loss in the future, it’s not that we’d have to go through this pandemic again,” Cobey says.
What will determine our future with the virus is how long protective immunity lasts. For severe coronaviruses like MERS and the original SARS, it persists for at least a couple of years. For the milder coronaviruses that cause common colds, it disappears within a year. It’s reasonable to guess that the duration of immunity against SARS-CoV-2 lies within those extremes, and that it would vary a lot, much like everything else about this virus. “Everyone wants to know,” says Nina Le Bert from the Duke-NUS in Singapore. “We don’t have the answer.”
Most people still haven’t been infected a first time, let alone a second. The immediate uncertainty around our pandemic future “doesn’t stem from the immune response,” Cobey says, but from “policies that are enacted, and whether people will distance or wear masks.” But for next year and beyond, modeling studies have shown that the precise details of the immune system’s reactions to the virus, and to a future vaccine, will radically affect our lives. The virus could cause annual outbreaks. It might sweep the world until enough people are vaccinated or infected, and then disappear. It could lie low for years and then suddenly bounce back. All of these scenarios are possible, but the range of possibilities will narrow the more we learn about the immune system.
That system may be vexingly complex, but it is also both efficient and resilient in a way that our society could take lessons from. It prepares in advance, and learns from its past. It has many redundancies in case any one defense fails. It acts fast, but has checks and balances to prevent overreactions. And, in the main, it just works. Despite the multitude of infectious threats that constantly surround us, most people spend most of the time not being sick.
Materials and Methods
In this work, we analyzed human immune cell frequency data from three cohorts. The Stanford cohort consisted of 398 individuals, both men and women, between ages 8 y and 89 y of age, sampled at the Stanford University clinical trials unit. The study protocol was approved by the Stanford University Administrative Panels on Human Subjects in Medical Research, and written informed consent was obtained from all participants. A total of 226 of these individuals were twin subjects from both MZ and DZ pairs from the Twin Research Registry at SRI International as previously described (4). A total of 23 additional subjects were included from a longitudinal study of aging in the immune system (11) and 149 previously unpublished subjects were included as well (combined in Datasets S1–S3). All demographic information of the subjects in the Stanford cohort is available in Dataset S1. The Stanford cohort comprises individuals from the following ClinicalTrials.gov study IDs: NCT01827462, NCT03020498, NCT03022396, NCT03022422, and NCT03022435. The Roederer et al. (10) cohort consists of 669 female twins sampled in the United Kingdom. The Carr et al. (6) cohort consists of 670 nontwin subjects, both men and women sampled in Belgium.
Immune Cell Population Frequency Measurements.
All three cohorts analyzed PBMCs, using either flow cytometry or mass cytometry. The immune cell composition data for the Stanford cohort were generated by the Human Immune Monitoring Center, using mass cytometry and antibody panels (4, 11). All cells were classified by manual gating, using the FlowJo Software (ThreeStar Inc.), and gating schemes are presented in the original publications. The PBMCs from the Roederer et al. (10) and Carr et al. (6) cohorts were analyzed by a combination of consecutive flow cytometry panels and manual gating.
Functional Response Measurements.
Frozen PBMCs were thawed and stimulated, with one of seven different cytokines (IL-2, IL-6, IL-7, IL-10, IL-21, IFN-α, or IFN-γ) for 30 min fixed and stained with antibodies to surface antigens (identifying eight responding cell populations: monocytes, B cells, CD4 + , CD4 + CD45RA + , CD4 + CD45RA – , CD8 + , CD8 + CD45RA + , and CD8 + CD45RA – T cells) as well as intracellular phosphor-STAT1, -3, and -5, respectively, as described in detail previously (4). The 90th percentile fluorescence intensity of stimulated samples was compared with the 90th percentile in PBS-treated cells (unstimulated) as a fold-change ratio used for downstream analyses (Dataset S2). In addition to the cell-signaling responses, we also include an analysis of antibody responses to seasonal flu vaccines (years 2010–2012) (Dataset S3). The vaccine administered was the Fluzone quadrivalent vaccine (Sanofi-Pasteur Inc.) and the vaccine-induced antibody responses were measured in a strain-specific manner, using HAI assays, and fold changes day 28/day 0 were used for analyses as described in detail in ref. 4. The flu-vaccine data included eight strains: A/Texas/50/2012, B/Massachusetts/2/2012, B/Brisbane/60/2008, A/California/07/2009(H1N1), A/Perth/16/2009(H3N2), A/South Dakota/06/2007(H1N1), A/Uruguay/716/2007(H3N2), and Wisconsin. The A/Texas/50/2012, B/Massachusetts/2/2012, and Wisconsin strains were removed due to small sample size (Data Preprocessing).
Of the 136 cell population frequencies measured in the Stanford cohort, 34 were considered in this study. We removed 89 cell populations because they were measured in only 10.3%, 29.1%, 39.4%, or 70.8% of the 398 patients. We also removed the following 13 subtypes because they either are rare or are the complement in a set of rare cell populations: plasmablasts, HLADR + NK cells, CD161 + CD45RA + Tregs, CD161 + CD45RA − Tregs, CD85j + CD4 + T cells, HLADR + CD38 + CD4 + T cells, HLADR + CD38 + CD8 + T cells, HLADR + CD38 − CD4 + T cells, HLADR + CD38 − CD8 + T cells, HLADR − CD38 + CD4 + T cells, HLADR − CD38 + CD8 + T cells, HLADR − CD38 − CD4 + T cells, and HLADR − CD38 − CD8 + T cells. To remove measurement dependence on measurement year (2009–2012), we applied the ComBat algorithm (38), an empirical Bayes method for removing nonbiological experimental variation that has been found to outperform other batch effect correction methods according to many metrics (39). Patient 398 was removed due to missing batch information, leaving a total of 397 patients studied from the Stanford cohort. A summary of the immune cell population frequencies and demographics data for the Stanford cohort is available in Dataset S1.
Functional responses having fewer than 34 patients (equal to the number of retained cell populations) were removed from consideration: three influenza strains (A/Texas/50/2012, B/Massachusetts/2/2012, and Wisconsin), as well as all 21 cytokine stimulation responses measured in the “Non-BT” cell subtype.
The Stanford, Roederer et al. (10), and Carr et al. (6) cohorts included composition measurements of nonidentical immune cell populations (Table S1 for a list of cell subtypes in each cohort), thus precluding a combined clustering analysis of the full dataset. The “monocytes” cell population was removed from the Roederer et al. (10) cohort as it was missing in 7.6% of the 729 patients. Of the remaining 32 populations, 11 had missing data in at most 1.6% of patients. For the Carr et al. (6) cohort, patients containing >20% missing measurements were removed, leaving a total of 449 patients. Cell populations with >10% missing values across patients were removed from consideration. Of the remaining 32 populations, 5 had missing data in at most 4% of patients. Missing values in both cohorts were imputed to the average across all patients.
Clustering analysis was performed on the PC scores (projections of immune cell composition data onto the eigenvectors of the correlation matrix—this places all immune cell population frequencies on the same scale). Clustering of individuals was performed on the top PCs with a k-means clustering algorithm implemented using the sklearn.cluster.KMeans() class in scikit-learn v0.17 (40), using interpoint Euclidean distance in PC space. The quality of clusters from this algorithm was quantified using two metrics:
i) The silhouette coefficient (41) is defined for each data point, i , as ( b i − a i ) / max < a i , b i >in which a i is the average pairwise Euclidean distance between data point i and all other points within the same cluster as i , whereas b i is the lowest average pairwise Euclidean distance between data point i and all data points in any other cluster. The silhouette coefficient ranges from −1 to 1, with a high value indicating good clustering.
ii) The between-cluster fractional explained variance is defined as the variance in the data where each point is replaced by the average of the points in the same cluster, divided by the total variance, V t o t , in the data (equal to the sum of between-cluster and within-cluster variance, V b / w and V w / i , respectively). More formally, let x i = ( x i ( 1 ) x i ( 2 ) … x i ( d ) ) T be a vector of PC scores (PCs 1 − d ) for patient i and μ = ∑ i = 1 N x i / N be the mean over all N patients. Now, let the clustering algorithm assign each patient to one of C clusters, c i ∈ < 1 , … , C >, and let X c be a matrix composed of the subset of rows of X corresponding to patients assigned to cluster c i.e., the rows of X c are the set < x i ∶ c i = c >. Finally, let μ c = 1 T X # < i ∶ c i = c >be the cluster average, where 1 is a vector of ones. Then, the between-cluster fractional explained variance, V b / w , f r a c , is calculated as follows:
V t o t = tr [ ( X − μ ) T ( X − μ ) ] V w / i = ∑ c = 1 C tr [ ( X c − μ c ) T ( X c − μ c ) ] V b / w , f r a c = V T o t − V w / i V t o t = V b / w V t o t .
Because these measures of cluster performance are difficult to interpret alone, they are compared with a null model formed by the following: (i) For each PC, i , in the original data, calculate the difference, a i , between the 97.5th and 2.5th percentiles of the scores of that PC. This gives a robust-to-outliers measure of the spread in each PC. (ii) Choose a number of PCs, d , to include in the clustering. (iii) Generate a uniform random sample of profiles within the PC space contained within a d -dimensional ellipsoid with principal axes equal to the respective differences calculated in i. If x i represents the randomly generated value for the PC i score, the null model satisfies ∑ i = 1 d x i 2 ( a i / 2 ) 2 < 1 .
This null model was chosen because the PC projections appear ellipsoid with apparent uniform density of points except at the outskirts (although with a fatter tail than a multivariate Gaussian). Thus, using a uniform density for the null model is conservative (i.e., it will tend to give a larger number of clusters than the real data).
These analyses are presented in Fig. S1, along with corresponding t-distributed stochastic neighbor embedding (t-SNE) plots, implemented using the sklearn.manifold.TSNE() class in scikit-learn v0.17 (40) with two components, perplexity of 30.0, and learning rate of 1,000.0. The t-SNE method for visualizing high-dimensional data overcomes some of the limitations of the linear PCA method (20).
Formally, we sought to develop a set of regression models that separately predict the values of each functional response, y ( k ) , given the immune cell compositions, X , of a group of individuals. Here, element y i ( k ) of vector y ( k ) represents the value of response k for individual i , whereas element X i j refers to the immune cell population frequency j in individual i . Linear models assume that the responses and predictors are related through the functional form y ^ ( k ) = X β ( k ) , where the “hat” distinguishes the predicted responses from the experimentally measured ones, and β ( k ) is the regression vector for functional response k . The elements of the regression vector are the weights of each cell population that define a linear combination that best predicts the functional response.
For the regression models, the functional responses were first transformed using the log-modulus transformation (42) to preserve rank ordering of responses while minimizing the effect of large outliers. For each functional response, PLS (Dataset S4) is used to find a set of linear combinations (termed the LVs) of cell populations that have the highest covariance with the functional response. These LV signatures, r j ( k ) , are the directions in the space of possible immune cell compositions that explain a particular functional response, ordered from best to least, as columns of the LV matrix R ( k ) . In this way it is possible that only a small number of such directions need to be identified to predict functional responses well. The advantages of using PLS are its simplicity, the avoidance of multicollinearity issues resulting from highly correlated immune cell populations (SI Materials and Methods, Regression Methods), and the conceptual advantage afforded by grouping cell populations together into latent variables that collectively explain a response of interest. We compared the performance of PLS with PCR, in which directions in the space of possible immune cell compositions are chosen to explain the greatest variance in immune cell composition across patients.
The regression models are evaluated using learning curves of 10-fold cross-validated (CV) negative Spearman correlation coefficients between observed and predicted response measurements for both PLS and PCR regression models. As described in ref. 43, cross-validation estimates out-of-sample model performance for smaller datasets by fitting a model on only a fraction of the available data and testing predictions on the remaining data left out. We used a 10-fold cross-validation scheme based on recommendations by Hastie et al. (43). The dataset is randomly partitioned into 10 groups and a PLS model separately fitted on each possible combination of 9/10 of those groups. The predictions on the left-out groups are compared with their experimental values, using a negative Spearman correlation coefficient. Learning curves are made by calculating CV error as a function of the number of LVs included in the model. The default error (when no LVs are used) is a correlation coefficient of zero. Error bars on learning curves are generally reported as SEs (44). However, because our measure of error combines errors over the entire dataset, we report error bars as SD of error calculated over 100 random instantiations of the CV learning curve (corresponding to different random 10-fold partitioning of the dataset).
We wish to verify that the PLS model is fitting meaningful correlations between both the immune markers and response functions and that improved model performance is not an artifact of the model-fitting procedure. To address these concerns, we evaluate the performance of a “null model,” created by random reassignment of responses across patients. This randomization eliminates the relationship between a patient’s immune cell composition and the value of the corresponding response, and the model-fitting procedure should correctly detect that there is no underlying structure to be learned from the data. We calculate 95% confidence intervals of the null model learning curve, using 100 instantiations of the null model to summarize how variable the estimates from the null model are. Responses for which the minimum error plus error bar come within 0.05% of the 95% confidence interval of the null model learning curve are removed from further consideration. Six signaling responses (IL-6/Mono/STAT5, IL-7/CD4 + CD45RA + /STAT1, IL-7/CD4 + CD45RA – /STAT1, IL-7/CD4 + /STAT1, IFN-γ/CD8 + CD45RA – /pSTAT3, and IFN-γ/CD4 + CD45RA + /pSTAT3) and four of the five flu-vaccine responses [A/California/07/2009(H1N1), A/Perth/16/2009(H3N2), A/South Dakota/06/2007(H1N1), and A/Uruguay/716/2007(H3N2)] were so removed.
Finally, to indicate the amount of front loading of explanatory power into the top LVs, we also report a normalized error defined as follows: If the minimum observed error for a given functional response in either PLS or PCR is denoted e m i n , and the raw model error is denoted e , the normalized error is defined as e n o r m = ( e − e m i n ) / ( 0.0 − e m i n ) , where an error of 0.0 indicates a degenerate model in which the mean of the response measurements is predicted regardless of an individual’s immune cell composition.
As discussed in the main text, the PLS model for the HAI response to the influenza B/Brisbane/60/2008 strain was fitted on only the subset of individuals who had a measurement increase in HAI titer upon vaccination (31% of patients with a post- to prevaccination HAI titer ratio of 1 were not included in the fit). Furthermore, the fit benefits from further outlier moderation seven patients with log-modulus HAI titer ratios greater than 4.0 (corresponding to HAI titer ratios >53.6) were adjusted to this 4.0 threshold for further outlier moderation. These seven individuals had HAI titer ratios equal to 64, 80 (for three individuals), 160, 320, and 640. The average HAI titer ratio for the remaining individuals was 11.3.
We highlight three linear regression methods in this section. The main regression method used in this study is PLS regression. However, ordinary least squares (OLS) and PCR are discussed first to illustrate the shortcomings that the PLS method addresses. Furthermore, the OLS results are used repeatedly in the development of the PCR and PLS methods.
12. Immunotherapy is on the cutting edge of immune system research.
Over the last few years, research in the field of immunology has focused on developing cancer treatments using immunotherapy. This method engineers the patient's own normal cells to attack the cancer cells. Vance says the technique could be used for many more conditions. "I feel like that could be just the tip of the iceberg," he says. "If we can understand better what the cancer and immunotherapy is showing, maybe we can go in there and manipulate the immune responses and get good outcomes for other diseases, too."