Moderator Kenneth Schaffner - 2:22pm May 8, 1997 

The past few years have seen exciting challenges to widely accepted immunologic knowledge. Perhaps the most recent challenge is the "danger" theory championed by Fuchs and Matzinger, but other major alternatives have been advanced. These include extensions of Jerneian idiotype network theories by Coutinho and Bandeira, as well as "stranger," morphostasis, "integrity," and antigen-localization accounts by Janeway, Cunliffe, Dembic, and Zinkernagel, respectively. The associative recognition model of Cohn and Langman has also evolved in this context to respond to this debate in immunology. 

These contrasting views will probably see the main function(s) of the immune system, and how tolerance is effected, differently, but perhaps we can come to some agreement about the main function(s). Then we can move on to look at the specifics of the alternative approaches and their strengths and weaknesses. We will examine the import of these various approaches for understanding and controlling responses to pathogens, to organ transplantation, to autoimmune diseases, and perhaps to cancer may be significant.  

The HMS Beagle has recruited a most distinguished group of immunologists as participants, and I am privileged to work with you. Welcome to the forum! The first question is, What did the immune system evolve to do? In other words, what is (are) the main function(s) of the immune response? Can we reach any consensus on this? At this initial stage, please keep the answers fairly general; we will get into details and mechanisms (both cellular and molecular) later. Let the debate begin! 

Doug Green - 10:23pm May 8, 1997 (#1 of 13) 

What did the immune system evolve to do? Simple question, maybe a simple answer, but a fairly complex (although hopefully elegant) explanation.  

Informally speaking, the immune system "evolved" to maximally damage parasites while minimally damaging us. (In this and all my discussions, parasites include viruses, bacteria, fungi, protozoans, and any metazoans that may attempt to eat us while we're alive.) Somewhat more technically, we can state that an individual with an immune system that gives maximum protection against parasites while minimally damaging the individual will have a reproductive advantage over individuals with weaker immune responses (who have a greater risk of dying from infection) or those with more damaging responses (who have a greater risk of dying from the consequences of inflammatory diseases).  

Sounds okay so far, perhaps. But there's a fundamental problem with this that may give us our first insights into how such things as disease caused by the immune system can exist. The older we get, the less reproductive advantage we are afforded by virtue of the protective system. It's subtle, because it's not a consequence of aging, per se, but a potential cause of aging.  

The idea is based on notions that have been described in detail by R.A. Fisher, George Williams, and Peter Medawar (Medawar, 1957; see also Rose, 1991). Sir Peter's is the easiest and most entertaining to read (although a recent book by George Williams also looks pretty user-friendly). Since this is nicely described by these authors, I'll just give the bare bones. First, we have to imagine a sort of creature that doesn't age. Medawar's analogy was a population of test tubes, but I prefer something fuzzier, so I'll propose Generic Individual Thingies, or GITs. Not only don't GITs age, but the probability of any GIT reproducing at any age is about the same. The only thing is that GITs, like everything else, can die by simple accident, so the number of 10-year-old GITs is less than the number of 9-year-old GITs.  

There are three types of GITs, immunologically speaking. One population has a very kind and friendly immune response that gently nudges parasites to go elsewhere; unfortunately, these GITs generally die of infection at a young age (so a very large proportion are adversely affected). Another population has a very powerful immune response, and any parasites that enter their bodies are annihilated. They never get infections, but the damage to their bodies that accumulates as a consequence of these powerful responses (because some tissue is damaged when the system destroys the parasites) eventually cripples and kills them, again at a young age. The third population, just like Baby Bear's porridge, are just right. They fight off most parasites, but not all, and the slowly accumulating tissue damage doesn't adversely affect them until later in life, when - and this is a key point - there are fewer of them anyway, because of the natural attrition due to unrelated accidents. Thus, in this third population, the advantage of the response is obtained early in life, when there are statistically more individuals to reproduce, and the adverse consequences are put off until later, when there are statistically fewer individuals. The third group will produce more offspring, and this strategy of the immune response will be favored over the others. 

Medawar and the others mentioned above explicitly stated that a gene that benefits an individual early in life will be favored, even if it causes reduced fitness later in life. Note that the argument does not assume that older individuals are weaker or less capable of producing offspring; the differential age-related effects are based only on the numbers of individuals of any given age in a population. Indeed, this is the argument that is used to explain why aging might happen at all: Deleterious effects of genes that are beneficial earlier in life are pushed back in time until they affect a relatively small cohort of the population (those that have been around longer). Interestingly, although he was an immunologist, Medawar did not use the double-edged nature of the immune response as an example of this kind of age-dependent benefit/consequence scenario.  

The result, then, is that the immune system is built to keep us relatively free of disease until we can effectively reproduce, after which the consequences of that destructive process can take their toll. (We tend to reproduce earlier in life - unless we're Tony Randall - for the same reasons). And no, I don't consider anyone reading and disagreeing with this a GIT.  

Zlatko Dembic - 3:43am May 9, 1997 (#2 of 13) 

If we agree with Doug (#1) about the definition of the "parasites" (quotation marks denote his definition), then the assumption that evolution has selected for the maximal-damage model, provided the integrity of an organism is protected until the mature reproductive age, is very reasonable. However, I disagree completely concerning the immune response causing aging. It is a contradiction per se because the immune system would end up as young in older individuals as it was in the beginning of the life of the same individual, and here we have little evidence that this is so; or, in other words, the immune system is inherently damaging or suicidal neither for itself nor for an individual.  

Coming back to "parasites": How does the immune system distinguish between the beneficial "parasites" like E. coli, which can help us in digestion and provide some essential nutrients, and the non-beneficial ones (dangerous, nonself, not lethal, and nonpathogenic)? Or what about the immune responses to some molecules that are not necessarily derived from "parasites"? I suggest that the immune system evolved to better protect the integrity of each organism until its reproductive maturity (Dembic, 1996). Senescence decreases the level of integrity. If harmonious, it slides in parallel with the immune system's aging, keeping the balance between activation and tolerance. If a tissue (organ) gets old faster than others, then this might eventually trigger an immune response and be manifested as autoimmunity or autoimmune disease.  

Throughout evolution, preservation of the integrity of an individual organism until its reproductive age is an obvious phenomenon that comes as the first priority of survival. Apparently each organ contributes its share toward this common goal. As a defense department, the immune system has the unique ability to infiltrate almost all other organs and as such is likely to assume an indispensable guardian-of-integrity role within the boundaries of an organism. There are two boundaries to be distinguished: conceptual and real. The first is the complexity formed by the interaction of the highly organized, high-energy state of "self" and the chaos outside the protective integrity. The real border is the skin, epithelia of digestive, respiratory, and urogenital tissues, and the eyes. Both boundaries contain genetic and epigenetic elements. Like the defense apparatus of a fictitious state, the immune system neither loses time on trying to guess what the potential and unknown enemies ("nonself") might look like nor builds a range of potential weapons against them. Instead, it uses its available integrative communications (cross-talk) and developmental or cleaning tools (i.e., oxidative radicals, pore-making molecules, and apoptotic signals) for defense. Identification of each particular "enemy" (nonself) has evolved to a great extent and now includes specific immunity, which is a consequence of the main driving force - protection of integrity - as opposed to self-nonself discrimination for natural selection. The potential useful weapons against disintegrative elements are within the "memory of the past"; thus the initial immune repertoire must be germ-line encoded. The variability of the response is dependent on class (molecule, virus, bacterium, or parasite), and the specificity depends on the way (how the complexity of signaling is being affected) of the intrusion.  

According to the integrity hypothesis (Dembic, 1996; Dembic, submitted), "integrity" is defined as the complexity of signals within and among a living cell, tissue, organ, and organism. Signals between living cells make the boundary between the organized "self" and disorganized "nonself." The complexity of such cross-talk allows many combinations by which signals could substitute for each other, although the number of messengers is finite for any given species. The protection of integrity of an individual depends on two critical choices that each individual has in life: to eat or to be eaten. The nervous system protects by anticipating the threats that could lead to disintegration, including specimens larger than a particular individual. The immune system senses disruption of the integrity of tissues and cells (other than apoptosis, or normal programmed cell death) but not that of molecules, although it uses molecules as effector tools and messengers. 

William O. Weigle - 9:39am May 9, 1997 (#3 of 13) 

The primary function of the immune system is to confront pathogens and other invaders of the body and react with these agents, enabling them to be subsequently neutralized and eliminated. At this time in our debate, I will focus only on specific immune responses and not deal with nonspecific events such as innate immunity, phagocytosis, complement usage, etc. The immune system is a highly complex, finely tuned and regulated system that is capable of handling a variety of encounters. It is a diversified system with fine specificity involving numerous specialized effector cells, numerous intracellular signals, cytokines, etc., allowing for containment of parasites, viruses, bacteria, tumors, etc. Many, if not all, of the systems involved in defending the body against pathogens are also at play in autoimmune reactivity. 

Although the primary function is in the defense of the body against pathogens, etc., encumbered in this process is the necessity of the system to refrain from mounting detrimental immune responses to self components. Thus, the body must recognize self and not aggressively respond to it. In this fashion, both nonself and self antigens are recognized by the host, and immune response usually results from encounters with nonself, whereas tolerance results from encounters with self. It is assumed that these events occur first in the thymus, where the decision as to self or nonself is made. Although there are variations of how this occurs, some means of negative and positive selection are most acceptable and most likely take place early in life. 

The main question to be addressed is how tolerance to autoantigens that either escape negative selection or are exposed to the immune system for the first time in the periphery is maintained during adult life. Although we all recognize that both self and nonself are recognized by the immune system, my view may differ from that of some of the other debate members in questioning whether there is a true discrimination made by the system. How does one account for the failure of the immune system to mount a positive immune response to monomeric heterologous (foreign) gamma globulin as well as certain other serum proteins (Weigle, 1980)? It was demonstrated some time ago by Howard Dvorak (1970) that bovine serum albumin (and most likely other heterologous albumin) in the absence of endotoxin is recognized as a tolerogen and not as an immunogen. Why do neonatal animals respond to antigens such as viruses and those incorporated into adjuvants (Forsthuber et al., 1996) but develop tolerance to heterologous serum proteins (Dietrich and Weigle, 1963)? Is there another property of heterologous serum proteins that allows them to be recognized as self?  

Rod Langman - 9:58am May 9, 1997 (#4 of 13) 

I would like to keep the concept of an immune system as simple as possible - the multiple manifestations of the system in nature and the laboratory are what the concept must account for. Evolution selected against organisms that could be killed prematurely, and one class of selecting agents is pathogenic infectious agents. At one end of the spectrum are bacteria that can be prematurely killed by viruses, and at another point in the spectrum are the vertebrates that can similarly be killed by viruses, bacteria, fungi, and protozoa. There are defense mechanisms that evolved to protect the host against these pathogenic agents, and evolution simply rummaged around in the trash can of mutations to weed out the ones that didn't help as well as others. However, evolution does not take sides, and the infectious pathogens searched in their bag of mutations to likewise avoid premature death in inhospitable hosts.  

Starting from relatively simple bacteria, we find, for example, families of restriction endonucleases that, upon recognition of a rather unique tract of nucleotides in the DNA, snip the DNA, and the "owner" of the DNA dies. In this way, a restriction enzyme coevolves with the host to not recognize host DNA, and sometimes - often enough to be kept around - viral DNA sequences. Obviously there is a limit to this game of hide-and-seek because the advantages of more, and more diverse, DNA reduce the number of different restriction enzymes or make their recognition sequences so complex that the viruses that stay simple can often find a place to hide. Notice that the biodestructive effector function (an endonuclease in this case) makes what can be called a self-nonself discrimination because it recognizes only sequences that are absent from the host; whether these sequences can be found in the current crop of viruses determines the advantage conferred by this particular enzyme defense mechanism. This is an example of what I would call germ-line-selected self-nonself discrimination, meaning that errors in recognizing self and not recognizing nonself result in elimination of the host's particular germ line (i.e., it dies prematurely).  

As an aside: Just because I use "self" and "nonself" to describe a phenomenon (because I can't think of a better pair of words for the moment), there is no point in arguing about other meanings for these words; the criticism needs to be directed at the concept they stand for in this discussion. If a better pair of words can be found, the concept will not have changed. I hope that the concept and the code words can be kept clear throughout the debate.  

It is not too difficult to imagine the time when, for reasons irrelevant here, animals hopped out of the water and started colonizing the land. The first land animals were rather like incubators full of nutrient broth on wheels. The task of keeping the parasites out of the broth was moderately successful, but inevitably some found their way in. The most difficult problem faced by the land animals, with their relatively long life spans (reproduction frequency, fecundity, and rearing included), is their highly uneven exposure to pathogens, whether airborne, in food, or via the occasional breaches of entry barriers. Defense mechanisms with a germ-line-selected self-nonself discrimination are limited in individuals with long life spans because they are continuously exposed to a much more rapidly evolving and diverse pathogenic universe. When a large germ-line-encoded repertoire of pathogen recognition and elimination substances includes self recognition and elimination, one (the only?) way to increase the repertoire while maintaining species diversity of self is to make the self-nonself discrimination by some somatic selection process. It is this point in evolution that I find to make the cleanest break with all other defense mechanisms.  

In short, I think that the concept of an immune system requires at a minimum both a biodestructive means of defense against infectious pathogens and a somatically selected mechanism for making a self-nonself discrimination. Evolution will seek to maximize the ability of the immune system to eliminate infectious pathogens while minimizing the ability of the immune system to destroy itself (the host). This is the sort of balance that Doug dealt with in his imaginary thingies (#1).  

Zlatko's suggestion (#2) that the immune system is under selection to maintain inner integrity suffers from several weaknesses. For example, when individuals are deprived of an immune system, they invariably die from infectious diseases, not other inner disharmonies. And when individuals with an immune system are given syngeneic grafts, these are accepted fully, but when given allogeneic grafts, these are rejected; the immune system can tell by the antigenic difference, not the stresses of engraftment, that elimination of nonself is to be applied to the allogeneic graft. I know there are responses awaiting these criticisms, but they hinge on mechanistic detail best left until later.  

Ephraim Fuchs - 2:31am May 10, 1997 (#5 of 13) 

Fortunately, it appears that all of us agree that our answer to "What did the immune system evolve to do?" has to fit into the larger paradigm of evolution by natural selection. In other words, I completely agree with Doug Green and others who postulate that the immune system evolved to enhance reproductive fitness and that it does so essentially by maximizing protection against parasites while minimizing damage to the host. Indeed, the choice of the word "parasite" is an excellent one, because it implies that the immune system should not destroy symbionts. This raises the question of how the immune system distinguishes between parasites and symbionts, given that both are nonself. I suspect that this question will be a major emphasis of further discussions, so I shall hold off until later. 

In considering the evolutionary function of the immune system, I think it is very worthwhile to reflect on the studies of the first immunologist, Ilya Metchnikoff. I consider him to be the first immunologist because he was the first to propose that inflammation is actually an active response to infection (Metchnikoff, 1968a; Metchnikoff, 1968b), as opposed to theories [I believe of Connheim (see Tauber and Chernyak, 1991)] that inflammation was a passive transudation of white blood cells out of damaged vessels. How did Metchnikoff come to such a conclusion? Before becoming an immunologist, he was an invertebrate embryologist who was predominantly interested in the function of the "wandering mesodermal cells," which today we would call phagocytes or macrophages. He speculated that one of the functions of these cells was to maintain bodily integrity by scavenging dead, effete, or senescent cells. My guess is that this function of the phagocytic system is quite similar to the integrative function as postulated by Zlatko Dembic (#2). Yet when he noticed an accumulation of phagocytes around a thorn plunged into a starfish, it then occurred to Metchnikoff that the same phagocytes may have acquired, as an evolutionary overlay, the ability to eat invading parasites and, in so doing, incite inflammation. What Metchnikoff did not ask was how phagocytes distinguish senescent cells from invading parasites. Only the latter induce inflammation and immune responses. My guess is that the functions are mediated by two different cell types: Macrophages provide the integrative, or morphostatic, function by eating senescent cells, whereas it is up to dendritic cells to initiate the adaptive immune response. My guess is that there is no need for lymphocytes for the integrative function, and this is why SCID mice or other mice lacking lymphocytes do not have gross morphological or reproductive abnormalities (when kept in sterile environments).  

It is worth noting that invertebrates, which lack adaptive immune systems, have a number of defenses at their disposal to defend against a variety of pathogens. Neutrophils and macrophages are quite adept at eating and digesting organisms with evolutionarily conserved signatures of "foreignness." I suspect that programmed cell death, or apoptosis, represents a primordial defense against parasitism by viruses. That is, if a virus comes in and disrupts the orderly timing of the cell cycle, the cell knows that something is wrong and performs the altruistic act of saving other cells with identical genomes by committing suicide. Given that apoptosis is accompanied by scission of the nucleic acids within the cell, it is an efficient way of killing any resident parasitic viruses. The apoptotic cell is then scavenged by phagocytes so that the internal contents of the cell can be efficiently recycled, an energy-efficient process. The body can handle this without inducing an immune response. Only when the rate of apoptosis is so high that the scavenging capacity of the organ is exceeded is there real trouble, and this is when the adaptive immune system is called in to play. According to this model, the immune system has evolved to react to situations in which non-apoptotic death or cell stress is occurring, or when the rate of apoptosis in an organ exceeds the capacity of the innate immune system (macrophages) to handle it. This is the sign of a dangerous parasitic infection.