Every biologist is familiar with the fact that injection of an antigen into an animal induces an immune response that results, among other things, in the production of specific antibodies. So, if a mouse is given antigen X, it will produce the corresponding anti-X antibody. If we now purify this anti-X antibody and inject it into a mouse of the same strain, (ie, expressing the same allotypes as the first mouse), the second mouse will produce an antibody that will react with the anti-X antibody and that represents an anti-anti-X antibody. This defines the idiotypic specificities, which were first described independently in the 1960s by Oudin (1) in the rabbit and by Kunkel (2) in humans. These specificities are located on the variable regions of both the heavy and the light chains and are the result of epitope(s) [in this case we use the term of idiotope(s)], related at least in part to the antibody combining site of anti-X. It was shown subsequently that this game could be played one step further: If we now inject a third mouse with the anti-anti-X molecule, the mouse will make anti-anti-anti-X antibodies. Immunologists proposed a simplified nomenclature that reads
X-Abl-.Ab2-.Ab3"
where Ab1 is the idiotype, Ab2 is the anti-idiotype, and Ab3 is the anti-anti-idiotype. This constitutes the so-called idiotypic cascade. One observation of particular interest, independently made by the groups of Oudin in Paris and Urbain (3) in Bruxelles, was that some antibodies in the Ab3 population behaved like Ab1, in that they could interact with antigen X, which the animal had never "seen." This led them to consider that Ab2 had structures resembling X, because it could induce the formation of antibodies (Ab3) that behaved as anti-X. Ab2 was thus considered an "internal image" of the antigen. The observation was generalized by Jerne (4), once it was shown that this cascade could take place spontaneously within the same animal and that the level of expression of each antibody could be modulated (up or down, depending upon the experimental conditions). Jerne proposed that the entire population of antibodies of a given individual were organized as a huge network of interactions, leading to a dynamic equilibrium of B lymphocytes. This was called the idiotypic network.
Modulation of the repertoire of B cells expressed at any given time by injection of either partner of the cascade was thoroughly investigated during a number of years. Particular attention was paid to the diverse roles exerted by the Ab2 antibodies, because it was demonstrated that two subpopulations, termed Ab2b and Ab2a, had completely antagonistic regulatory effects, resulting either in the production of Ab3 antibodies that were Ab1-like, or in the down regulation of Ab1. Ab2b are anti-idiotype antibodies that contain a typical internal image and are therefore able to stimulate the production of anti-X antibodies, without disposing of the original antigen. This led to the development of numerous attempts to make "idiotypic vaccines." The second type of population, Ab2a, recognizes idiotopes that are outside the antibody combining site of Ab1, and they appeared quite suitable to repress undesirable antibodies (ie, autoantibodies that have a pathogenic effect in autoimmune diseases).
The theoretical interest of idiotypic vaccines is to function as surrogate for antigens that are difficult to use for immunization, either because they are difficult to isolate, are poorly immunogenic, or present a potential hazard. Many attempts have been made, especially for viruses, including HIV, and parasites such as schistosomiasis. Antibodies were generally produced, but it turned out that in most (if not all) cases they were poorly neutralizing, and the protective effect in vivo was too low relative to the high efficiency that is expected from an acceptable vaccine.
Down-regulation of pathogenic autoantibodies encountered in autoimmune disease has also been extensively investigated, and it is still currently used in therapeutics. In fact, one uses immunoglobulin preparations assembled from very large pools of donors that are administered intravenously (IVIg). Although experimental arguments are compatible with the presence of antiidiotypes, many other factors may account for the down regulation of the patient antibodies.
Finally, an interesting application of the internal image is the isolation of cellular receptors by Ab2b that mimic the corresponding ligand. Numerous examples have been reported and successfully used; the first was the Ab2b of the insulin cascade, which was shown to activate the insulin receptor of the b islet of pancreas. Adrenergic receptors or T3 and T4 thyroid hormone receptors are other examples. This could facilitate receptor isolation by the use of immunoadsorbents prepared with these anti-idiotypic internal image antibodies.
