Zlatko Dembic - 2:03pm May 10, 1997 (#6 of 13) 

I am glad to note that we seem to agree on the role of immunity as a protection against "parasites." This definition is not a problem for me, because "parasites" are the most abundant and the most common potential disintegrators of living organisms.  

I should be more clear about the integrity view of the world because, I believe, none of the weaknesses and criticisms mentioned by Rod and Ephraim really apply to the integrity model, or, at worst, not more than they do to any other. The central problem of the confusion lies in the distinction between the "integrative" and the "protection of integrity" functions ascribed to the immune system.  

An integrated being (AIB), as I would like to call a living organism, is the most fundamental definition of "self." AIB could be perceived through the prism of the self-nonself discrimination principle as "not-changing self." With the description of the integrity principle and AIB, I wanted to avoid mentioning "self" for the weaknesses in the definition of self, per se. For example, self changes its physicochemical texture during development and on the onset of adulthood together with alterations in its biological functions. Thus, the antigenic properties of self are changed during life of AIB. How can we then understand what self tolerance means, when the "self" is constantly slipping away? AIB is conceptually the unchanged part of that self. AIB imposes tolerance to its immune system (the mechanics will come later), and the changes in the integrity can cause either activation or tolerance. Not all disruptions of the integrity will result in the immune response - only those that can be sensed by specialized cells. To be effective, specific immunocytes are not directly "integrated" in an organism, and they do not play an integrative or organizing role. They receive commands (signals[1] and [2]) and are allowed to say no (for a very good reason). The link to AIB is maintained by specialized cells (including dendritic cells and possibly macrophages) that sense the disruption and disintegration of a tissue and non-apoptotic death of the cells. Here the role for the dendritic cell is very close to that explained by Ephraim (#5). In fact, the only difference I see is that I propose that the activation of a dendritic cell is caused by the disruption of preexisting signal(s); on the other hand, the "danger" model suggests that there is a single signaling pathway for the induction of activation of a dendritic cell.  

To specifically answer, then, Rod's criticism (#4) about not seeing overt inner disharmony in immunodeprived individuals, I should state that the infectious disease is (in part) the disharmony of inner integrity, and, as such, it is perceived by specialized cells that can stimulate the immune response by signaling. With the lack of specific immunity (immunodeficiency), however, the signals (commands) are useless. As the cells of the immune system do not organize the integrity (they just protect), the cohesiveness of tissues suffices for the maintenance of integrity until the organism finally becomes disrupted by "parasites" and succumbs.  

This can also explain the criticism by Ephraim (#5) that SCIDs and immunodeficient mice do not have gross morphological or reproductive abnormalities if kept in isolators. It remains open whether the lack of the "protection of integrity" under sterile conditions would lead to shortened life span. This could be tested.  

Rod's further comment, that the rejection of the allogeneic graft cannot be explained by the integrity model, is not true (probably due to my or his imprecise interpretation of the model), because it is easily explained by the lack of some crucial integrative signals that are sensed by specialized cells (the mechanics later).  

Antonio Bandeira - 5:19pm May 10, 1997 (#7 of 13) 

What did the immune system evolve to do? In simple terms, the classical answer is "to protect us against bacterial and viral pathogens." Most people are happy with this because, again in simple terms, it is unquestionable that if mammals lose the diversity of their immune system, they die of infectious disease. As people like Antonio Coutinho and John Stewart would argue (and I agree with them), this remains true. As extensively discussed recently in a forum on evolutionary origins of immunoglobulins and T-cell receptors (Stewart and Coutinho, 1996), no one could yet demonstrate that defense against pathogens was in fact the major function of the primordial immune system, as far as a system of variable-region molecules (VRM) is concerned. As an example, Ohno argues (in Stewart and Coutinho, 1996) that the VRM system of cold-blooded animals doesn't seem to add much to the already existing defense systems. I think that everybody would agree that as soon as diversity increases, self reactivities unavoidably occur. The generation of diversity to fight infections is therefore linked from the start to the establishment of tolerance to self. If it is true that mammals die of infections as the immune system is crippled in its diversity, it is also more and more documented that autoimmune diseases come together with immunodeficiencies. Most interesting is that mice lacking transforming growth factor-beta, an immunosuppressive interleukin, die very early (before reproduction) of autoimmune diseases. I disagree, therefore, with the notion that the immune system is optimized in evolution "to maximize protection against parasites while minimizing damage to the host." If we just consider the billions of bacteria that colonize our gut, it looks like the immune system cares to maintain the integrity of its own components as much as that of the intestinal flora. Actually, the thinking nowadays in the field of mucosal immunity is that the organism tries all the time to "cool down" immune responses against gut antigens, because too much inflammation may lead to life-threatening situations (e.g., inflammatory bowel disease). Ironically, one could argue that for several years, immunologists have been trying to tolerize against self antigens (to cope with autoimmune disease) using exactly the same strategy built up evolutionarily to tolerize against bacterial pathogens in the gut! Given this, I believe it would be more accurate to think, at the start, that immunology is the science of self-nonself discrimination, and the immune system before anything else, considered as a VRM system without an immediate connotation of "immune." Until we know more about the functions of the primordial VRM system, the issue of the evolutionary pressures for the immune system remains unsolved. Hence, so far I don't see a reason why nonself antigens would be more important than self antigens as a driving force for the immune system. 

Rod Langman - 11:40pm May 10, 1997 (#8 of 13) 

Antonio, can you add a short description of the VRM system to help the uninformed audience? It would also make my response to your arguments better if I start with your explanation rather than my interpretation of the VRM system, both then and now. I would like to know whether Antonio Coutinho agrees that the elimination of pathogenic infections (one very good reason to have an immune system today) requires coupling antigen recognition with a biodestructive effector mechanism. Although I would agree with Antonio that self and nonself antigens are equally important targets of immune system recognition, the problem is how the immune system treats these two classes of antigen differently when the recognition of some antigens kills the host and recognition of others enables the host to survive. When the immune elimination of antigen kills the host, with rare exceptions (e.g., adults infected with nonpathogenic lymphocytic choriomeningitis virus) these are self antigens, but not all self antigens would, if eliminated, kill the host (e.g., effete erythrocytes). What criteria would the immune system need to use to minimize self-inflicted death and minimize nonself- (pathogen) inflicted death? This is the driving force in the evolution of the immune system, the action - what to do to whom, not the mere passive description of who is who. 

Antonio Bandeira - 4:20pm May 11, 1997 (#9 of 13) 

Rod, thanks for helping to clarify my message. First, concerning the VRM system. Immunoglobulins (antibodies; abs) and gamma-delta T-cell receptors are proteins encoded by genes constructed by a process of somatic DNA rearrangement, worked out to generate huge diversities randomly, and this is apparent in the so-called hypervariable region of today's immune system. These proteins are the VRMs. They came from primordial variable-region molecules. The question is what were the functions of the primordial VRM system? The alternative view provided by Stewart and Coutinho is that the VRM system arose with nondestructive functions and only later was redeployed to be used also for destructive effector functions. What do they propose? They suggest that the original function may have been in relation to the general problem of conflict of interests between clonal cell lines and the multicellular organism as a whole. For the first vertebrates, with circulatory blood, this problem would be very important for the first lymphocyte-like cells because of VRM diversity. More precisely, because VRMs are related to cell-adhesion molecules, which play an important role in cellular differentiation and tissue organization, Stewart suggests that that the first VRMs functioned as self-reactive membrane-bound receptors that enabled lymphocyte-like cells to interact and to control their own growth. For more on this, see Stewart (1992). I would just add here a very brief note, as Stewart and Coutinho (1996) wrote: If the function of the primordial VRM system was based on regulated nondestructive self reactivity, then there was no need at that moment for a categorical self-nonself discrimination. Where I think they are going with this is that the strategies of today's immune system to deal with tolerance to self find their bases in the primordial VRM system's organization based on positive (nondeleterious) recognition of self and in a dynamic equilibrium of population sizes, tightly regulated.  

As is to be expected, the question "What did the immune system evolve to do?" cannot be dissociated from the discussion of the two alternative paradigms, the Burnetian and the Jerneian, on self-nonself discrimination. As we all know, this is the most open issue. The complexity of vertebrate immune systems requires an understanding of several interrelated features (cellular diversity, clonality, the existence of a resting state, programmed cell death, control of total lymphocyte numbers, and repertoire selection); given this, it seems to me that there are enough unknowns and alternative sound reasonings to the whole issue to date to make it not unreasonable to question the statement that the primary function of the immune system is to fight infections. If so, classical textbooks should take this into consideration and provide the reader with a stimulating discussion rather than a dogmatic statement.  

Now, with regard to your question to Antonio Coutinho, I think he would agree that elimination of pathogenic infections (yes, one of the good uses for today's immune system) requires coupling of antigen recognition with a given effector mechanism, but I guess that the word "coupling" should be better explained. Certainly he would agree that there must be at least two classes of immune responses, destructive and nondestructive, but this is not directly related to antigen recognition per se but, among other things, to the microenvironmental context in which the cell is activated.  

Now, just a few comments on what you said. First, we agree then that both self and nonself antigens are important targets of immune recognition. The key question I would ask here, which should be added to the Day 2 issue, is this: Do you allow or accept that, following recognition of self antigens, the immune system may develop a functional nondestructive immune response? Second, how does the immune system treat the two classes of antigens (self and nonself)? In my view, in contrast to what you wrote, there are not two classes of antigens. I guess that is not what you want to say, because everybody agrees that self and nonself cannot be distinguished from a biochemical point of view. So where is the difference? This is a key issue of self-nonself discrimination. Our first answer to this question is "the historical context" - that is, the presence or absence of a ligand during the ontogenetic development of the immune system and not during the development of the cell. 

Zlatko Dembic - 5:10pm May 11, 1997 (#10 of 13) 

Antonio, could you please clarify what you mean by "functional nondestructive immune response"?  

Doug Green - 7:13pm May 12, 1997 (#11 of 13) 

Zlatko's discussion (#6) has several very useful features. In particular, I agree that the barriers between our bodies and the outside world (e.g., skin and gut) are not only first lines of defense but also important in the signaling to initiate an immune response. The main problem I have, however, is in the concept of "integrity." In a general sense, the immune system maintains integrity just as by definition every system in the body does (because integrity is defined as the complexity of signals within and among a living cell, tissue, organ, and organism). I suggest, though, that the distinctions made by this model move us no closer to a mechanistic model of how the immune system works than do the distinctions of self and nonself. 

For much of this century, the function of the immune system was so mysterious that it seemed that no simple model (or even a fairly complex one) could explain its function. Self marking, universal principles of self-nonself discrimination, networks of interacting receptors that form an internal image of antigenic self (the homunculus) - these concepts evoked images of massive complexity from which immune function might emerge, and they had the "benefit" of being so loosely framed and flexible that they could accommodate nearly any piece of information while not really explaining anything. But for this reason these models were also not very satisfying. 

Zlatko objects to some of my statements (#2) on the grounds that the system has no way to discriminate between dangerous and beneficial parasites (or unrelated molecules). But of course it doesn't, not if all are presented to the system in the same way. As I mention in Day 2, message 12, as a first approximation, the immune system discriminates between molecules that are always present in the body (and available for presentation to the immune system) and those that are only sometimes present. There are simple, understandable, and concrete mechanisms that mediate this distinction, which can be defined at molecular levels (e.g., the mechanism of negative selection). Another thing that the body does is respond to tissue damage: Mast cells release their granule contents and tissues produce inflammatory cytokines in response to damage. These are important in initiating responses. If a tissue becomes damaged by some other means (as in the premature aging that Zlatko suggests), it is easy to imagine that this could help to trigger a response to proteins that are newly exposed on the damaged tissue. 

From these points of view, I think we are in agreement. Certainly this is akin to what Ephraim says, and I agree with him. Necrotic cell death (occurring because cells have been extensively damaged, from either without or within) is extremely pro-inflammatory, although relatively little is known about precisely what the damaged cells produce that mediates this response. Zlatko also strongly objects (#2) to what he thinks is my suggestion that the immune system causes aging. However, I didn't suggest that; what I said was that the immune system - and I include inflammatory responses as an integral part - damages host tissue in the course of making a response, and this damage can accumulate. The basic principle I invoked was the idea that this can be tolerated if the accumulated damage does not result in disease until later in life. This is similar, but not the same, as saying "after reproductive maturity." In humans, it is true that senescence decreases reproductive maturity, but this is by no means a necessary state of affairs (as Fisher, Medawar, and Williams all noted in their rigorous treatments of the evolution of aging). In some organisms, older equals fitter and more fecund: large corals, for example. The evolution of senescence based on the simple models I discussed here helps to explain the existing relationships between age and a variety of functions (and diseases).  

Rod Langman - 8:56pm May 12, 1997 (#12 of 13) 

In response to Antonio (#7 and #9): The VRM concept is different in important ways from the markers of inflammation and disintegration that characterize in a general way the views of Doug, Ephraim, and Zlatko. The Stewart-Coutinho view of the origin of the world of the immune system is quite different from mine, but that is not of interest until the two world views start to describe the same thing: a protective immune response. Although Antonio Coutinho is not a participant, he claims to have his eye to the screen and may want to pass along any corrections to my interpretations.  

I understand the VRM model to have evolved a set of specificity elements in order to maintain a kind of overview of the physiology of the individual - a kind of housekeeping. Then, under the influence of withering bacterial and viral assaults, these VRMs were retooled and linked to the biodestructive effector mechanisms we associate today with the expression of immunity to infectious diseases. So long as the housekeeping VRMs are separated and non-overlapping with the immune VRMs, there are no difficulties that arise from one VRM being under two evolutionary selection pressures to perform two different tasks. The problem I have is when a self-recognizing set of VRMs is used to establish a dominantly suppressive system of unreactivity to self and another set of VRMs takes over the biodestructive protective duties. If unreactivity is dominant, then immunity is recessive, and antigens having a mixture of both self and nonself epitopes are treated as self and unreactivity ensues. Any smart bug would make sure it had one self epitope and be protected, and grafts that shared one histocompatibility marker would take instead of being rejected.  

You wanted an answer (#9) to the question of whether the recognition of self antigens may lead to a functional nondestructive response. I'd say no. Of course, this depends on how self and nonself antigens are to be classified, and you are quite correct to point out that there are no such antigens per se; "self" and "nonself" are postscripts that describe how the immune system actually behaved toward these antigens, with self antigens those that would lead to the death of the host if an immune response were mounted against them and nonself the remainder. This conclusion is, we agree, the result of a historical learning process. The historical process in the VRM world view leads to dominant suppression toward self epitopes, whereas in the associative antigen recognition (AAR) view, dominant immunity develops toward nonself epitopes because recognition of self epitopes has been recessively deleted.  

The retort to the AAR view has been along the lines that, because antigens known to be self by genetic criteria are on occasion recognized by the immune system, all self-reactive recognition is not deleted, and the AAR view is disproven. However, we just had to agree that self is not defined by genetics but by the immune system, and because we are not all dead of autoimmunity, all self reactivity that kills is gone. 

The question on everyone's agenda is why some genetic self components are treated differently from others. There are two parts to the answer that are going to have to be faced in any model. First, there is no way to induce epitope-specific unreactivity (by any mechanism) in the absence of that epitope. In other words, some epitopes are simply never accessible to immune recognition. Second, there is a threshold of detection, below which epitopes are invisible. In other words, the immune system does not respond to antigens present at a concentration of one molecule per mouse, and not all self components are in a concentration range that the immune system could detect, and hence have reactivity eliminated. There are numerous examples of this, including many idiotopes. Thus, we need to list the "self" antigens that are not accounted for by these almost inescapable limitations to all mechanisms of immune regulation.  

Antonio Bandeira - 7:56am May 13, 1997 (#13 of 13) 

Zlatko, "functional nondestructive immune response" means a non-aggressive immune response - in general, one that would not lead to the destruction of the target recognized. Several cytokines released by T cells are actually considered as hemopoietic growth factors or to enhance survival or even proliferation of the targets. Suppressor or regulatory T cells, for example, are able to inhibit graft rejection or autoimmune disease induced by nontolerant T cells. There is evidence that interleukins described as immunoregulatory (e.g., transforming growth factor-beta) could well be mediators of such a process. In this case, the destruction could be actually directed against the aggressive cells. The Th1/Th2 type of dichotomy is actually being used to explain tolerance, a concept now called "immunodeviation." "Facilitating" antibodies have been described linked to the survival of transplanted tissues. Interleukin- (IL) 2, for example, is a growth factor for T cells, but the possibility now emerges that it regulates T-cell numbers and particularly prevents death of effector cells. IL-2 receptor knockout develops autoimmune diseases. The point I want to make, following Stewart and Coutinho's propositions concerning the primordial VRM system, is that there is enough room to envisage several other fundamental functions for these cells linked to normal regulation of tissue development and organization. For example, IL-1, IL-6, and tumor necrosis factor-alpha seem to participate in this respect in early phases of development, yet they are taken as aggressive T-cell-derived lymphokines.