One cell–one antibody: prelude and aftermath G J V Nossal This year marks the 50th anniversary of Burnet’s clonal selection theory. Here Gustav Nossal recounts his pioneering work that supported Burnet’s theory and led to the death of the direct template hypothesis. The date 21 October 2007 marks the 50th anniversary of the publication of Frank Macfarlane Burnet’s clonal selection theory of antibody formation1. The Walter and Eliza Hall Institute of Medical Research will host an international symposium in celebration. As I had the good fortune to obtain the first evidence in favor of the theory2, and as it was my fate to become Burnet’s chief ambassador in its defense in international fora, it is a pleasure to center this essay on the heady events of 1957 and on the slow but relentless march of clonal selection toward general acceptance. Medical and PhD studies I am, as it were, an ‘accidental immunologist’. I wanted to be a doctor as far back as I can remember and studied medicine at The University of Sydney. Given that in those days anyone who matriculated from high school could enter medical school, and given further that the Australian Government supported university study for ex-servicemen and women, we were 600 in first-year medicine in 1948. The staff/student ratios were appallingly low; our teachers struggled mightily, but the instruction left a lot to be desired. For third-year biochemistry and physiology, a few of us ‘clever’ kids took the matter into our own hands, each studying selected topics in the library, then giving the group a seminar on the latest data. This fired my imagination for research, and I took advantage of a program that allowed students to take a year off and work in an academic department as research apprentices. In G.J.V. Nossal is in the Department of Pathology, The University of Melbourne, Victoria 3010, Australia. e-mail: [email protected] my case, I had a clear idea of what I wanted to do. Fascinated by biochemistry in the preWatson-Crick era, I wanted to discover the ‘secrets of life’ by studying the multiplication of the simplest forms of life, the viruses. In the Bacteriology Department, a senior lecturer, Patrick M. de Burgh, was investigating the effects of the ectromelia virus on the metabolism of infected liver cells. I joined him and, cognizant of the one-step growth curves of bacteriophages, I charted the growth of ectromelia in the liver after intravenous infection, recording the latent period and then successive waves of multiplication. Remarkably, when liver cells were infected with up to 20 median lethal doses of virus per cell, the viruses multiplied independently of each other3, showing that “the fundamental virus-synthesising centres must be sought at a sub-cellular level,”3 antedating John Cairns’ demonstration of vaccinia virus cytoplasmic factories by more than a decade. Pat de Burgh arranged for me to spend a week in Melbourne sitting at the feet of Australia’s greatest virologist, F.M. Burnet, and that experience at age 21 marked me for life. I finished my medical course and hospital residency years, and went to Melbourne in 1957 to pursue a PhD degree. Imagine my disappointment to learn that Burnet had switched his interests from virology to immunology, the area he now wanted me to follow. How had this come about? Burnet’s clonal selection theory Burnet had long had a theoretical interest in immunology, prompted by some early studies on the formation of antibodies to bacterial toxins. He had noted several features not explained by the then-fashionable ‘direct template’ theory of antibody formation4. These NATURE IMMUNOLOGY VOLUME 8 NUMBER 10 OCTOBER 2007 The Walter and Eliza Hall Institute of Medical Research © 2007 Nature Publishing Group http://www.nature.com/natureimmunology E S S AY Sir Gustav Nossal, successor to Sir Macfarlane Burnet as Director of the Walter and Eliza Hall Institute of Medical Research. included the exponential rise in antibody titers after immunization; the improvement in the quality or affinity of antibody after repeated immunization; the booster effect after secondary immunization; and, very significantly, the phenomenon of immunological tolerance that he had predicted on theoretical grounds5 and that Peter Medawar and associates proved experimentally6. He had made attempts to replace the direct template hypothesis with a rather clumsy ‘indirect template’ theory, but deep down he knew this was not quite right. He was most impressed with Astrid Fagraeus’ work implicating the family of rapidly dividing plasmablasts and plasma cells in antibody formation7. He was also intrigued with a case of a 1015 © 2007 Nature Publishing Group http://www.nature.com/natureimmunology E S S AY monoclonal gammopathy, macroglobulinemia, that was accompanied by a high titer of autoantibodies8. Above all, he kept chewing over the natural selection theory of antibody formation promoted by Neils Jerne9, in which millions of antibody specificities existed in the serum before any immunization, with antigen serving somehow to catalyze the production of much more of the antibody than happened to fit the antigen in question. So there was a picture of antigen causing cell division, of autoimmune clones gone wild, of antigen sometimes causing tolerance, and of antibody formation perhaps being selective rather than instructive. Into this charged atmosphere burst David Talmage. In a reflective essay covering many topics, he wondered whether natural antibodies were actually cell surface receptors and whether antigen selects cells for replication on the basis of their having the right receptor10. He did not come out and say that each cell has only one kind of receptor. This prompted Burnet to crystallize his ideas, and over one weekend in August 1957 Burnet drafted his short paper for the Australian Journal of Science1. He speculated that the small lymphocyte population of the body is actually a repertoire of specificities, with each cell bearing on its surface just one kind of natural antibody. During immunization, antigen selects cells with the corresponding specificity for multiplication and differentiation into antibody-secreting status. This accounts for an exponential rise in antibody concentrations and an increased number of the ‘right’ cells to respond to a second immunization. As the immune response progresses, somatic mutation of antibody genes allows cells of higher affinity to arise and to be selected by antigen, accounting for affinity maturation. If a clonally individuated cell encounters ‘its’ antigen while the system is still immature, whether a ‘self ’ antigen or one introduced artificially during embryonic life, the cell, rather than being activated, is actually killed, explaining natural self-tolerance as well as artificially induced tolerance. Occasionally the tolerance system goes wrong, allowing ‘forbidden clones’ to escape, resulting in autoimmune disease. The theory retains Jerne’s random generator of antibody diversity and Talmage’s idea of the cell as the unit of selection, but extends both ideas to cover the key inadequacies of the direct template theory. Antibody formation by single cells At that time I was still scanning the virus literature and was intrigued by Marguerite Vogt’s ability to grow poliovirus in isolated single cells11. Also, I read a fair proportion of Joshua Lederberg’s work, as he was expected at the Hall Institute for a short sabbatical leave. He had 1016 cultured single bacteria in microdroplets. It occurred to me that I could, in best Popperian fashion, disprove clonal selection by showing that one cell could simultaneously form two or more antibodies. Lederberg duly arrived and we discussed three possibly suitable sensitive antibody titration techniques: immobilization of motile salmonella bacteria by antibody to flagella; lysis of sheep erythrocytes in the presence of antibody and complement; and neutralization of bacteriophages. We chose the first, largely because of Lederberg’s deep knowledge of the H antigens of salmonella. We faced a few hurdles, however. There were few decent microscopes at the institute. We rustled up an old Bausch and Lomb, still with vertical rather than inclined eyepieces, and no darkfield condenser. I figured out that if I removed the condenser and inserted a large coin in its stead, permitting just an arcuate path of light to come through, I could create a fuzzy but adequate imitation of darkfield to visualize swimming bacteria. There was no micromanipulator to move cells around, but one was located in the Microbiology Department and borrowed. Lederberg taught me the essentials of micromanipulation and then unfortunately had to go back to the United States. By late 1957, my first few results trickled through2. Rats were immunized with two different strains of salmonella, and single lymph node cells were suspended in hanging microdroplets surrounded by mineral oil and cultured for 4 hours. Then, ten or so motile bacteria were instilled into each droplet, first of one strain and then of the other, and were monitored for immobilization. With a ‘good’ cell, the results were dramatic: immobilization was nearly instantaneous. Over the ensuing months, I improved the techniques considerably and obtained very clean results12. One cell always formed only one antibody. This was a first hint in the direction of clonal selection. Using a phage-neutralization system, a group headed by Mel Cohn and Ed Lennox obtained contrary results13, which made me stick with the problem longer than was really necessary. The end result14 was that of 3,628 single cells tested, only two seemed to secrete two antibodies, which could well have represented monoclonal antibodies with a rare crossreactivity or perhaps binucleate plasma cells, which are not uncommon. My confidence in my own results was increased by results from immunofluorescence studies (which of course measure antibody content and not secretion) that failed to find doubly specific cells15. My colleague Olavi Mäkelä did indeed pursue the problem using phage-neutralization techniques. He failed to find double producers16 but made the interesting observation that when cells were tested with a cross-reacting phage, each single cell gave its own unique pattern of cross-neutralization, some even making ‘heteroclitic’ antibodies that neutralized the cross-reactive phage better than the immunizing one, foreshadowing the precise qualities of monoclonal antibodies. Death of the direct template theory Selective theories made only slow headway, so Gordon Ada and I sought to embark on a test of the antigen content of single antibodyforming cells. Using flagellin heavily iodinated with 125I, and relatively low antigen doses to minimize background radioactivity, we micromanipulated single antibody-producing cells to marked spots on a glass slide and analyzed them by autoradiography. In circumstances that would have detected as few as four antigen molecules per cell, 216 single cells showed no antigen whatever in them17. Hugh McDevitt et al. reinforced this conclusion18. So there was no template to copy! Of course, studies of antibody-forming cells did not directly address the characteristics of the B lymphocyte that was their ancestor. The discovery of David Naor and Dov Sulitzeanu19 that only very rare lymphocytes bound a given antigen was critically important. The ‘hot antigen suicide’ experiment of Gordon Ada and Pauline Byrt20, in which very highly radioactive antigen specifically abrogated the antibody-forming potential in an unimmunized lymphocyte population, emphasized the probability of antibody receptors on clonally individuated cells. A further elegant experiment was that of Martin Raff et al.21, which showed that when an antigen was used to ‘cap’ the immunoglobulin receptors of a B cell, there was no detectable other immunoglobulin left as a ring on the lymphocyte surface; thus, all the immunoglobulin seemed to have the same specificity. Final proof was slow in coming. The isolation of antigen-specific B cells from unimmunized animals proved onerous, despite the fluorescence-activated cell sorter. Cloning single B cells in vitro to test such a cell’s antibody-forming potential also required methodological advances. When we finally succeeded in both endeavors and showed that the antibody secreted faithfully reflected the original specificity22, our paper caused little comment. By the time of the 1967 Cold Spring Harbor Symposium on immunology, most aspects of clonal selection had in fact been accepted through the weight of the indirect evidence. ‘One cell–one antibody’ had its most glorious aftermath in the work of Georges Köhler and César Milstein23 and the revolution in basic science and therapeutics ushered in VOLUME 8 NUMBER 10 OCTOBER 2007 NATURE IMMUNOLOGY © 2007 Nature Publishing Group http://www.nature.com/natureimmunology E S S AY by monoclonal antibodies. Furthermore, as molecular biology demonstrated the somatic ‘minigene’ translocation events underlying the generation of the B cell repertoire and the post-antigenic hypermutations refining antibody specificity, it was satisfying to be able to map these findings on to the original, insightful sketch1. That the T cell system also obeys clonal selection rules, though with some noteworthy differences, is an extra bonus. Rarely has one theoretical paper so comprehensively fertilized a discipline or been so thoroughly vindicated. COMPETING FINANCIAL INTERESTS The author delcares no competing financial interests. 1. Burnet, F.M. A modification of Jerne’s theory of antibody production using the concept of clonal selection. Aust. J. Sci. 20, 67–69 (1957). 2. Nossal, G.J.V. & Lederberg, J. Antibody production by single cells. Nature 181, 1419–1420 (1958). 3. Nossal, G.J.V. & de Burgh, P.M. Growth cycle of ectromelia virus in mouse liver. Nature 172, 671 (1953). 4. Pauling, L. A theory of the structure and process of forma- tion of antibodies. J. Am. Chem. Soc. 62, 2643–2657 (1940). 5. Burnet, F.M. & Fenner, F. The Production of Antibodies 2nd edn. (Macmillan, London, 1949). 6. Billingham, R.E., Brent, L. & Medawar, P.B. Actively acquired tolerance of foreign cells. Nature 172, 603–606 (1953). 7. Fagraeus, A. The plasma cellular reaction and its relation to the formation of antibodies in vitro. J. Immunol. 58, 1–14 (1948). 8. Mackay, I.R. & Gajdusek, D.C. An “autoimmune” reaction against human tissue antigens in certain acute and chronic diseases. Arch. Intern. Med. 101, 30–36 (1958). 9. Jerne, N.K. The natural selection theory of antibody formation. Proc. Natl. Acad. Sci. USA 41, 849–857 (1955). 10. Talmage, D.W. Allergy and immunology. Annu. Rev. Med. 8, 239–256 (1957). 11. Lwoff, A., Dulbecco, R., Vogt, M. & Lwoff, M. Kinetics of the release of poliomyelitis virus from single cells. Virology 1, 128–139 (1955). 12. Nossal, G.J.V. Antibody production by single cells. IV. Further studies on multiply immunized animals. Br. J. Exp. Pathol. 41, 89–96 (1960). 13. Attardi, G., Cohn, M., Horibata, K. & Lennox, E.S. On the analysis of antibody synthesis at the cellular level. Bacteriol. Rev. 23, 213–223 (1959). 14. Nossal, G.J.V. & Mäkelä, O. Kinetic studies on the incidence of cells appearing to form two antibodies. NATURE IMMUNOLOGY VOLUME 8 NUMBER 10 OCTOBER 2007 J. Immunol. 88, 604–612 (1962). 15. White, R.G. Antibody production by single cells. Nature 182, 1383–1384 (1958). 16. Mäkelä, O. The specificities of antibodies produced by single cells. Cold Spring Harb. Symp. Quant. Biol. 32, 423–430 (1967). 17. Nossal, G.J.V., Ada, G.L. & Austin, C.M. Antigens in immunity. IX. The antigen content of single antibodyforming cells. J. Exp. Med. 121, 945–954 (1965). 18. McDevitt, H.O., Askonas, B.A., Humphrey, J.H. & Sela, M. The localisation of antigen in relation to specific antibody-producing cells. I. Use of a synthetic peptide (T,G)-A-L labelled with iodine-125. Immunology 11, 337–351 (1966). 19. Naor, D. & Sulitzeanu, D. Binding of radioiodinated bovine serum albumin to mouse spleen cells. Nature 214, 687–688 (1967). 20. Ada, G.L. & Byrt, P.L. Specific inactivation of antigenreactive cells with 125I-labelled antigen. Nature 222, 1291–1292 (1969). 21. Raff, M.C., Feldmann, M. & de Petris, S. Monospecificity of bone marrow-derived lymphocytes. J. Exp. Med. 126, 443–454 (1973). 22. Nossal, G.J.V. & Pike, B.L. Single cell studies on the antibody-forming potential of fractionated, haptenspecific B lymphocytes. Immunology 30, 189–202 (1976). 23. Köhler, G. & Milstein, C. Derivation of specific antibodyproducing tissue culture and cell lines by cell fusion. Eur. J. Immunol. 6, 511–519 (1976). 1017
© Copyright 2026 Paperzz