|Home | About | Journals | Submit | Contact Us | Français|
The question under analysis in this commentary is, what was the evolutionary selection pressure that necessitated the ectopic expression of a subset of peripheral self-antigens in the thymus and by peripheral APC? The suggestion is that antigen expression is delayed until after the immune system is responsive.
There is considerable literature on the ectopic expression of peripheral self-antigens in the thymus and by APC, some of which are under the control of the transcription regulator Aire (reviewed in ). It is established that deletion of Aire results in autoimmune responses to this subset of peripheral self-antigens. Given that not all self-antigens expressed exclusively in the periphery are ectopically expressed in thymus (i.e., peripheral tolerance exists), the question arises as to why does the inactivation of one of the controlling factors, Aire, result in autoimmunity directed against the Aire-regulated subset of antigens? The answer impinges on all theories of the sorting of the repertoire commonly referred to as the self-nonself discrimination.
For reasons of clarity, let me begin by stating the conclusion. Self-antigens must be expressed depending on their role in the physiology of the organism. Therefore, it is to be expected that some of them will be expressed after the immune system becomes responsive. The responsive immune system cannot distinguish newly arising self from nonself. Those peripheral self-antigens that are delayed in expression must be ectopically expressed as negatively selecting ligands for T-helpers before the developing immune system becomes responsive. Delayed appearance of a peripheral T-helper self-ligand requires early expression in the thymus in order to avoid autoimmunity.
I will first develop the basic argument, then discuss the limitations to the data, even though they do not fault the principle.
This conclusion, that in order to avoid autoimmunity early ectopic expression of delayed appearing self-antigens is obligatory, is a logical consequence of the “associative recognition of antigen” (ARA) model for the self/nonself discrimination. This model has been discussed many times [2-5] and needs no detailing here. To recall the elements essential to this conclusion:
An example of delayed expression resulting in autoimmunity is owed to Adams et al. . Two transgenic murine lines were analyzed that express the T-antigen of the Simian Virus (SV40) in the β-cells (insulin producing) of the pancreas. One line expresses the T-antigen early in fetal life while the developmental time window is open. The other line expresses the T-antigen delayed until after the window closes. The early expressor treats the T-antigen as self and is unresponsive to it. The late expressor treats the T-antigen as nonself, the humoral response to which results in diabetes.
If the T-antigen provokes a response when its expression is delayed in the periphery, then the antigen could not have been expressed ectopically as a Ps-RII negatively selecting ligand in the thymus. Assuming this to be also the case for the early expressor, then the immune system’s unresponsiveness must be due to negative selection in the periphery while the developmental window was open (insufficiency of eTh).
The conclusion from this experiment suggests a reinterpretation of the experiments of Le Douarin et al. [9-11]. A quail limb bud is grafted onto a chick embryo before the immune system arises. The chick hatches with a healthy integrated quail wing but roughly one week after birth acute rejection occurs. In the developmental time framework, the explanation would be that the quail limb expresses a new self-component after the developmental window closes and the system is responsive (sufficiency of eTh).
Given this assumption, why is the quail limb not rejected in the quail? There are two answers:
The two explanations were tested by grafting embryonic quail thymus onto a chick embryo, resulting in a chimeric quail-chick thymus. In such an animal, the grafted quail limb is accepted (as is the quail thymus), showing that the quail limb is accepted in quail because the target of rejection is ectopically expressed in the thymus.
These two experiments supported the ARA model and set the stage for the hypothesis that that a subset of delayed expression peripheral self-antigens are ectopically expressed in thymus as negatively selecting ligands, Ps-RII [3, 4, 6]. It might be well to stress that:
Not all peripheral autogenously generated components that are ectopically expressed in the thymus are regulated by Aire [13-15]. This does not impact on my argument which applies only when the failure to ectopically express the component results in a debilitating immune response to it. The mechanism regulating the ectopic expression is not an issue here. I, therefore, separated to some extent ectopic expression and Aire regulation (leaving a link in the title) as I have little reason to doubt that Aire plays a necessary, if not sufficient role in early ectopic expression in the thymus .
Ectopic expression must occur when the developmental window is open, meaning that the immune system is unresponsive due to an insufficiency of eTh. During this period when tolerance is being established to self, the thymic and periphery are indistinguishable. Therefore, deletional tolerance mediated by extrathymic Aire-expressing cells  made a great deal of sense. The apparent contradiction  needs resolution but does not affect the principle that delayed expression of self requires a special mechanism.
Kyewski and colleagues [19-21] have proposed that ectopic gene expression covers antigens that arise late in life. However, they do not place the observations in a general conceptual framework. Consequently they use pregnancy-associated antigens and those assumed to be expressed during puberty as examples. The mechanisms that protect the fetus against rejection cannot involve ectopic expression in the mother as paternal gene products are involved as the major target. As for puberty and the lactating female, there is only an intuitive reason to suspect that “self-antigens” are involved. No debilitating autoimmunity to antigens unique to lactation or puberty have been revealed. Self-antigens are defined by the immune system during a learning process, not by the immunologist. Auto-genously-generated antigens that have no debilitating consequence when attacked by the immune system, are not a selective pressure defining the immune self. It is only within the frame-work of the ARA Model [3-5] that ectopic expression makes sense.
A major contribution to the question of the role for ectopic expression has recently come to my attention. Guerau-de-Arellano et al.  showed that Aire expression is essential, necessary and sufficient, in the “perinatal” period to prevent generalized autoimmunity. Further, they show the shut down of Aire after weaning had few deleterious consequences and overt autoimmunity did not result. If Aire was not expressed until after birth, its presence did not protect against autoimmunity. The straightforward explanation for this observation under the ARA Model is that once tolerance to a given antigen is established then the continued presence of the antigen maintains the tolerance. The ectopic expression in the thymus as Ps-RII permits establishing tolerance in the T-helper class and once established the continued expression of Ps-RII in the periphery maintains the tolerance in the T-helper class. In the absence of the eTh anti-Ps-RII, the interaction of all other T/B cells with ligand (Signal 1) results in their deletion, thus maintaining tolerance in all classes. This predicts that self-antigens involved in this phenomenon are delayed in expression as Ps-RII. Clearly the ectopic expression in the thymus, after the delayed peripheral antigen has been expressed, cannot stave off autoimmunity, as the immune system cannot distinguish newly arising self from nonself.
Guerau-de-Arellano et al.  view the requirement for thymic tolerance to be due to an initial postnatal lymphopenia. Why does lymphopenia break tolerance? In order to break tolerance (establish autoimmunity), eTh specific for the self-antigen must be induced to an auto-sustaining level. We have postulated for some years now [6, 7, 23-26] an antigen-independent pathway that primes induction of eTh. In the race between deletional inactivation by interaction with self and the appearance of eTh anti-self by the antigen-independent pathway, the latter dominates in the rapidly dividing cells of the lymphopenic animal. Lymphopenia is an independent phenomenon. Normally the rate of deletion of anti-self T-helpers keeps the level of eTh anti-self sufficiently low to maintain tolerance. Ectopic expression may allow the individual to pass safely through the period of perinatal lymphopenia but this would be a role for Aire in addition to that of coping with autoimmunity to the late expression of self-antigens. The two autoimmune phenomena may be concomitantly regulated by Aire but they are of different origins.
An Aire−/− C57BL/6 mouse develops a humoral autoimmune response to a variety of peripheral self-components [27-29]. A hybridoma library from such an autoimmune mouse should permit characterization of the family of target self-antigens responsible for the autoimmunity. Many of the autogenous ligands observed could be irrelevant (e.g., housekeeping). However, those targets expressed on the surface of viable cells from WT C57BL/6 mice would be good candidates to explore the time of their expression as Ps-RII and as surface components physiologically active. One example of a delayed expression component under Aire control would validate this hypothesis and encourage a more extensive search. The isotype of the humoral response to a variety of self-antigens would, as a fringe benefit, tell us a great deal about the regulation of effector class because it would be uninfluenced by pathogen-host interaction signals (“danger,” “harm,” “pathogenicity,” etc.).
This work was supported by a grant (RR07716) from the National Center For Research Resources (NCRR), a component of the National Institutes of Health (NIH) and its contents are solely the responsibility of the authors and do not represent the official view of NCRR or NIH. I would like to thank the reviewers for their perceptive criticisms and request for precision in the description of the phenomenon of ectopic expression which led to the inclusion of the section “An in-depth view of the data”
Conflict of interest: The author declares no financial or commercial conflict of interest.