Lederberg, J. (1959) Genes and antibodies. Do antigens bear instructions for antibody specificity or do they select cell lines that arise by mutation? Science 129, 1649-1653.

Eisen, H. N. & Siskind, G. W. (1964) Variations in affinities of antibodies during the immune response. Biochemistry 3, 996-1008.

Weigert, M. G., Cesari, I. M., Yonkovich, S. J. & Cohn, M. (1970) Variability in the lambda light chain sequences of mouse antibody. Nature 228, 1045-1047.

Kaartinen, M., Griffiths, G. M., Markham, A. F. & Milstein, C. (1983) Nature 304, 320-324. Griffiths, G. M., Berek, C., Kaartinen, M. & Milstein, C. (1984) Nature 312, 271-275. Foote, J. & Milstein, C. (1991) Kinetic maturation of an immune response. Nature 352, 530-532.

Doubts about affinity maturation. Antigen capture is clearly a key event in B-cell ontogeny, but the theory that maturation is affinity-based has come into conflict with new data.

Hybridomas derived from different stages of the murine immune response to the hapten 2-phenyl-5-oxazolone (Ox) exclusively used a single pair of antibody heavy and light chain variable genes in the early response, and these genes mutated systematically to give higher affinity. This "canonical" anti-Ox structure was replaced in later response stages by heterogeneous gene usage. This repertoire shift occurred without a concomitant increase in hapten affinity, suggesting the action of a different selective factor. JF, while a postdoc in César Milstein's lab, measured the hapten binding kinetics of each monoclonal antibody in the Ox collection, and found that many of the antibodies with mediocre affinity that emerged in the late response had an order-of-magnitude faster rate constant than the starting repertoire. The repertoire shift could then be explained by postulating that binding an antigen rapidly, as well as with high affinity, was a factor in lymphocyte selection driving a heretofore unrecognized process of "kinetic" maturation.

A number of subsequent studies on protein antigens have exposed difficulties with affinity-based selection:

Newman, M. A., Mainhart, C. R., Mallet, C. P., Lavoie, T. B. & Smith-Gill, S. J. (1992) J. Immunol. 149, 3260-3272.

•Smith-Gill's laboratory found that anti-lysozyme hybridomas made at different times after immunization showed no evidence of increasing affinity over the course of the immune response.

Mantovani, L., Wilder, R. L. & Casali, P. (1993) J. Immunol. 151, 473-488.

•Casali's group, studying human rheumatoid factors, found little affinity difference between molecules that clearly had a high number of antigen-selected somatic mutations.

Mukherjee, J., Casadevall, A. & Scharff, M. D. (1993) J. Exp. Med. 177, 1105-1116.

•Scharff's group, working on immunity to cryptococcal infection, again found discrepancies between somatic mutation and affinity increase.

Roost, H.-P., Bachman, M. F., Haag, A., Kalinke, U., Pliska, V., Hengartner, H. & Zinkernagel, R. M. (1995) Proc. Natl. Acad. Sci. USA 92, 1057-1061. Bachmann, M. F., Kalinke, U., Althage, A., Freer, G., Burkhart, C., Roost, H., Aguet, M., Hengartner , H. & Zinkernagel, R. M. (1997) Science 276, 2024-2027.

•Zinkernagel's laboratory, looking at the murine immune response to vesicular stomatitis virus infection, made a seminal finding: that on-rate closely correlated with virus neutralization, moreso than did affinity. In this case also, affinity maturation was not evident. A subsequent paper showed that affinity was only a marginal factor in virus neutralization; any antibody above a certain affinity threshold was potentially neutralizing.

Goldbaum, F. A., Cauerhff, A., Velikovsky, C. A., Llera, A. S., Riottot, M. M. & Poljak, R. J. (1999) J. Immunol. 162, 6040-6045.

England, P., Nageotte, R., Renard, M., Page, A. L. & Bedouelle, H. (1999) J. Immunol. 162, 2129Ð2136.

Some clarification has emerged from recent studies on the response to hen egg lysozyme. Roberto Poljak's group, which has generated many anti-lysozyme monoclonals, did a comprehensive survey of kinetic and affinity properties of their collection, and again found that high affinity was the rule for IgGÕs, regardless of immune response stage. Bedouelle's group examined one of these antibodies, identified germline genes from which it originated, and re-created the probable precursor, devoid of any somatic point mutations. The germline precursor proved to have a 60-fold lower affinity than the mature antibody, attributable largely to a single residue difference. The implication of Bedouelle's paper is that maturation does occur, and that the foregoing studies of the immune response to protein antigens simply failed to detect it, because the process happens in a shorter time window than anyone realized.

(Sure. Why not call it that.)

Foote, J. & Eisen, H. N. (1995) Kinetic and affinity limits on antibodies produced during immune responses. Proc. Natl. Acad. Sci. USA 92, 1054-1056.

Foote-Eisen Model. The failure of affinity maturation to explain the course of immune responses to protein antigens led JF and Herman Eisen to propose a new model, which appeared as a Commentary along with Zinkernagel's paper. In the model, affinity does not increase without limit, but is governed by separate kinetic limits on two biological events:

1. The on-rate constant (k_on) determines the efficiency of initial antigen capture, and is in turn limited by the rate of antigen diffusion. Somatic mutation can improve antigen capture by generating antibodies with kon up to the diffusion limit, but not beyond. Numerically, maximum k_on is about 10⁶ M^-1 s^-1 for antibodies binding small protein antigens, with exceptional cases (probably utilizing electrostatic interactions) up to 10⁷.

2. The off-rate constant (k_off) controls whether after initial capture an antigen is more likely to be released or internalized. Decreasing k_off increases antigen residence time on a B cell's surface, hence increases the probability of internalization. However, beyond a time period determined by membrane dynamics, very approximately 1 hour, internalization becomes a certainty, and is unaffected by further decreases in off-rate. Thus, the optimum kinetic phenotype of an antigen-specific B cell should also show a k_off slow enough to guarantee that antigen capture wll be followed by internalization.

Hapten and protein antigens. Haptens are small, generally aromatic, molecules that form limited antigenic determinants when conjugated to a protein carrier. Antibodies to conjugated hapten structures are easily raised, and the interaction of these antibodies with free hapten is readily quantified by spectrometric or radioactive assays. Unfortunately, conventional hapten conjugates do not allow meaningful analysis of the interaction of an antibody with the actual immunogen used to induce that antibody. Conventional conjugates are heterogeneous ensembles of carrier molecules with different numbers of haptens distributed over structurally diverse attachment sites. This polymorphism precludes assigning a preparation an unambiguous chemical concentration that could be used with the Mass Action Law to interpret physical data. Such analysis is crucial to interpret changes in the antibody repertoire that occur during an immune response. Use of native protein antigens in place of conjugates does not offer a solution. Detection and kinetic measurement of the interaction of native proteins with antibody is problematical, and the lack of a detachable hapten ensures that analysis of the immune response to the entire protein surface will be a task of great complexity.