George Snell's First Foray Into the Unexplored Territory of the Major Histocompatibility Complex

Corresponding author: Jan Klein, Abteilung Immungenetik, Corrensstr. 42, 72076 Tübingen, Germany., jan.klein{at}tuebingen.mpg.de (E-mail)

THE expression "x number years ago" is a moving target. Right now it is 50 years ago, but next year it will be more than 50 and a year later more again. Historians avoid this relativity by agreeing on a date that serves as a fixed point in the ceaseless flow of time. Year 0 divides the recent past into B.C. and A.D., whereas the year 1950 sets the present apart from the time "before the present," B.P. In both instances, "before" trails off into the mists of antiquity.

Curiously, 1950 can also be taken as a dividing point in the history of the major histocompatibility complex (Mhc), separating the "before" from the "present" times, for it was in the first year of the present era that George D. Snell published two papers that marked the beginning of a modern inquiry into the nature of the Mhc. These works also mark the beginning of an interminable interest in the Mhc shared by an ever-growing coterie of students. To an immunogeneticist, 1951 is the annus mirabilis, a remarkable year in which the first hints of what later would be recognized as two essential properties of the Mhc—its polymorphism and its genetic complexity—were revealed. But to grasp the true significance of Snell's contribution, we must place it into its proper historical context.

In 1951, the only Mhc known was that of the laboratory mouse, the histocompatibility 2 or H2 system. (The designation Mhc was not to be introduced until the early 1970s, when it became known that systems genetically homologous to H2 existed in many other vertebrates, and the mouse and human Mhcs were already changing the course of more than one discipline.) In 1951, Snell in Bar Harbor and Peter A. Gorer in London were the only two researchers with a vested interest in the Mhc; all other major players in the Mhc drama of the post-1950 era were occupied with other exploits. In 1951, Jack H. Stimpfling was still teaching bacteriology to nursing students at the University of Colorado in Boulder; he would not team up with Snell at Bar Harbor until 6 years later. D. Bernard Amos was completing his medical studies in London and was 1 year away from joining Gorer at the Department of Pathology, Guy's Hospital in London, to search for intra-H2 recombinants. Gustavo Hoecker in Santiago, Chile, was attempting to develop an immunological means of protecting mice against the growth of transplanted syngeneic leukemias; 1 year later, however, he and his wife Olga Pizarro were to travel to London to work with Gorer on the tissue distribution of H2 antigens, and then to Bar Harbor to analyze H2 determinants serologically. Donald C. Shreffler was still a graduate student at the University of Chicago at that time, dreaming of becoming an agriculturist. The "serum serological" or Ss protein, his laissez-passer to the H2 show, did not enter his dreams until a dozen years later. And in Prague, too, the protagonists of the future Czech School of Immunogenetics were in various stages of completing their education. In all fairness, however, in 1950 W. Elwood Briles, W. H. McGibbon, and M. R. Irwin (see OWEN 1989 ) did describe a red blood cell antigen (then designated "D" but later renamed "B") in the domestic fowl ("chicken"; BRILES et al. 1950 ) that was demonstrated by Louis W. Schierman and Arne W. Nordskog 11 years later to elicit the rapid rejection of allogeneic skin grafts and was henceforth recognized as the Mhc of this species (SCHIERMAN and NORDSKOG 1961 ). But in 1950, D (B) appeared to be nothing more than one further member of a long series of chicken blood group antigens.

In 1951, Jean Dausset at the Regional Blood Transfusion Center of the St. Antoine Hospital in Paris was enthusiastically harvesting the first fruits of his success in developing a technique for the detection of antibodies capable of agglutinating leukocytes. He was convinced at the time, however, that the antibodies present in the sera of leukopenic patients were of autoimmune origin; only several years later did he realize that the antibodies were produced in response to donor leukocytes introduced during blood transfusions. Despite her Ph.D., Rose O. Payne at the Department of Medicine, Stanford University School of Medicine, was still working as a technician in the Division of Hematology, but like Dausset she was already intrigued by the possibility that leukopenias and thrombopenias might be of immunological origin. Stimulated by Dausset's paper, she began a search for the presumed autoantibodies, only to realize, again like Dausset, that the antibodies were induced by transfusion. After discovering a better source of antibodies in the sera of multiparous women, she was then to proceed, with Walter Bodmer, in describing the LA series of antigens that were ultimately to contribute the last two letters to the HLA designation of the human Mhc. In 1951, Jon van Rood was still trying to decide what to do with his life, but finding himself momentarily in New York, where he had followed his girlfriend, he took a sabbatical in the Department of Internal Medicine, Columbia University Presbyterian Hospital. The days of computer-assisted analysis of sera from multiparous women for leukoagglutinating antibodies and the definition of the 4A/4B system of HLA-B antigens still lay some 10 years in the future.

The different approaches Gorer and Snell undertook in their studies came to epitomize the two faces of the H2 system—immunological and genetic. Gorer, an M.D., found his way to the Mhc via his attempts, originally suggested to him by J. B. S. Haldane, to explore a possible link between blood group antigens and the mysterious factors responsible for the rejection of allogeneic tumor transplants. Both Haldane and Gorer thought that such a link might exist because the rejection of tissue grafts seemed to resemble the body's reaction to transfused incompatible red blood cells (GORER 1938 ). Since the reaction was known by then to be mediated by antibodies in the blood, it was tempting to seek parallels with the rejection of transplanted solid tissues. Gorer ultimately became a strong advocate of the humoral theory of graft rejection and crossed swords with his colleague, Peter B. Medawar, who was an equally forceful proponent of the cellular theory, the notion that allogeneic grafts are destroyed by cytotoxic cells. The nature of the "allograft (homograft) reaction," as the response of the recipient to transplanted allogeneic tissue came to be known, then assumed a leading part in Gorer's research, along with the serological characterization of the H2 antigens. His interests thus became squarely focused on the immunological aspects of the H2 system, and the genetic aspect was relegated to a back seat. Snell, a Ph.D. trained in genetics (see WEIR 1994 ), was less interested in the mechanism of graft rejection than in the nature of the loci responsible for it. Inspired by Clarence C. Little's review on the genetics of tumor transplantation (LITTLE 1941 ), he became intrigued by the histocompatibility loci themselves—their numbers, properties, and function. He realized, of course, that to study the H loci, a method enabling them to be distinguished from one another would have to be designed. Since there were apparently many H loci and they all manifested themselves by the same effect—rejection of an allogeneic graft—they appeared to be individually indistinguishable. Snell the geneticist saw only one way to differentiate the loci: by creating a situation in which the donor and the recipient of a graft differed at one of the H loci only, a different one in each donor-recipient pair. To create such a situation, Snell hit on the idea of producing congenic (coisogenic) strains—pairs of mouse strains in which the difference between the members in each pair would be restricted to one H locus only (SNELL 1948 ). Here, his genetic background, and specifically his exposure to Drosophila genetics in Hermann J. Muller's laboratory at the University of Texas in Austin, proved to be a tremendous asset, for it was in this intellectual atmosphere that Snell found the kernel of the coisogeny principle.

It was, however, one thing to produce coisogenic lines differing at genes coding for easy-to-follow visible markers in an organism that multiplied like, well, like flies. It was another altogether to attempt a similar feat with genes for which the only indicator of a difference was the growth of a tumor and hence death of the host. Since it is difficult to obtain progeny from a dead mouse and since the generation time of a mouse is significantly longer than that of a fly, there were obviously some logistical problems to be solved, and the patience of an investigator was to be stretched to its limits. If a fire destroys the results of your efforts when you are halfway through a long-term experiment, who could blame you for giving up? Yet, Snell remained undeterred. Sticking to the ingenious strategies of cross-intercross and cross-backcross-intercross systems that he had designed (SNELL 1948 ), he made a new start in the production of the congenic lines. But while the wheels of mouse generations were slowly turning and he was waiting patiently for the lines to attain the desired homogeneity of their backgrounds, Snell used his preliminary results to get the ball rolling with one of the H loci, a locus that appeared to behave differently from all the others and that therefore did not seem to require full coisogenicity to yield itself to analysis. Serendipity, too, helped him to jump the gun.

Since some of the strains Snell used in the production of congenic lines differed at visible markers, he was able to follow the behavior of these markers in the repeated backcrosses and to keep an eye open for a possible linkage of a marker with an H gene. And luck was on his side! He noticed, first of all, that by using well-established tumors (i.e., rapidly growing tumors transplanted from one mouse to another over many generations) and selecting the dose carefully, he could often reduce the requirement for H compatibility to a single locus. He even began to suspect that it was always the same locus that determined the growth of a tumor in such situations. In other words, it appeared that one of the many H loci was more important than the others in determining the fate of a transplanted tumor. Confirmatory evidence emerged from the segregation of one of the marker genes, the mutation Fused (Fu), responsible for the fusion of tail vertebrae and deformations of the tail. It appeared, and was indeed demonstrated by Snell in an experiment deliberately set up for this purpose, that the H locus, with its strong effect on tumor growth, and the Fu locus were genetically linked. Thus, not only did Snell discover an H locus that stood out from the crowd, he also found a visible marker gene with which he could follow the segregation of the former. One further development then established that this particular H locus was indeed unique among all the H loci.

Up to this stage, the studies in Gorer's and Snell's laboratories had run parallel, without any communication between the two researchers. Snell found no reference to Gorer's work in Little's review and so remained unaware of it, and Gorer, of course, could not have known about Snell's progress since the latter had so far published nothing on this subject. In addition, personal contacts across the Atlantic were curtailed during World War II. In 1946, however, Little met Gorer at a conference in Italy, and the two made arrangements for Gorer to visit The Jackson Laboratory, where Little was Director (see RUSSELL 1987 ). When Gorer and Snell finally had the opportunity to compare notes, they were both amazed at the similarity of their experiences with tumor transplantation. There was even the distinct possibility that they had both stumbled upon the same genetic system that strongly affects the outcome of transplantation. To test this possibility, they set up a cross involving the Fused mutation, and Gorer then tested the segregants for the presence of antigen II, which he discovered with his antisera, before the tumor was inoculated into the mice. The outcome of the linkage test was clear-cut: Snell's Fu-linked H locus was either identical or closely linked to Gorer's antigen II-encoding locus. Following Snell's proposal to designate individual histocompatibility loci by the letter H and a serial number (SNELL 1948 ), Gorer and Snell agreed to call this first identified H locus H2 in recognition of the fact that it encoded antigen II (GORER et al. 1948 ).

After their parting of ways, both investigators continued to work on the H2 locus, Gorer focusing on its serological, Snell on its histogenetic characterization. Gorer had first to solve a number of technical problems because, in standard serological assays, mouse antibodies turned out to be highly capricious in their behavior, and reproducibility of results proved to be difficult to achieve (GORER and MIKULSKA 1954 ). It was, therefore, not until 1954 that Gorer and associates described the first series of H2-controlled antigens (GORER and MIKULSKA 1954 ; AMOS et al. 1955 ). In the meantime, Snell had designed and applied a clever strategy for distinguishing alleles at the H2 locus by tumor grafting. The principle of the method was to cross an inbred strain, generally M, with the Fu-bearing stock F, then to outcross the Fu-bearing hybrids with another inbred strain, generally N, and finally to inoculate cells of an M-derived tumor into the segregating progeny to find out whether the M and the N strains carried identical or different alleles. If the alleles were different, the following genotypes and corresponding outcomes of the transplantation could be expected:

Here H2^m, H2ⁿ, and H2^f are the H2 alleles of the M, N, and F strains, respectively, the F-stock being heterozygous for the dominant Fu mutation. The expectation is that, as long as the recipient of the inoculum bears at least one H2^m allele of the tumor donor, it cannot reject the transplant. This expectation follows from the assumption that the nature of the allograft reaction is immunological and hence that the recipient cannot react against antigens that it itself possesses: it is tolerant of such molecules. In this case, therefore, one subset of the outcross progeny would be susceptible and the other subset would be resistant to the inoculated tumor, and the resistance would be Fu-linked. In the opposite case, in which the M and N strains carried the same H2^m allele, all the outcross individuals would be expected to succumb to the tumor because they would all bear the H2^m allele derived from the N strain. And finally, if the rejection of the tumor transplant were effected not via H2- but via non-H2-encoded antigens, then, as in the first case, a subset of the progeny would be susceptible to the tumor and the other subset resistant, but the resistance would not be Fu-linked.

In the first application of this design, Snell tested a group of inbred strains and tumors derived from them in various permutations, allowing him to pit already identified alleles against the H2 alleles of each new strain added to the set. The outcome of the experiment was that some of the tested strains did indeed carry identical H2 alleles, but that the majority (four out of six) bore distinct alleles (SNELL and HIGGINS 1951 ). The inclusion of two additional strains into the set not only disclosed the existence of yet another allele and thus strengthened the indication that the variability of the H2 locus was unusually high, but also revealed another peculiar feature of the locus.

In extending the experiment, Snell observed that a tumor from the inbred strain A (H2^a) grew in A-strain mice, but not in mice bearing H2 alleles, b, d, k, or p. In this regard, the tumor behaved as expected. Surprisingly, however, the tumor also killed F₁ hybrids derived from H2^d and H2^k strains. This phenomenon was restricted to the H2^d/H2^k heterozygotes and to H2^a tumors; all other H2 heterozygotes were resistant to the H2^a tumor. Snell explained this observation by speculating that H2^a was either a compound allele encoding antigens otherwise specified by the H2^d and H2^k alleles or H2 was in reality a complex of at least two closely linked loci, one locus encoding H2^d and the other H2^k antigens, H2^a being an H2^dk recombinant (SNELL 1951 ). Ultimately, the latter interpretation proved to be correct, and the two loci of the H2 complex came to be known as H2D and H2K (KLEIN and SCHREFFLER 1971 ).

In 1951, Snell was thus able to demonstrate, first of all, that the H2 was unique among the histocompatibility loci in its strong effect on the fate of histoincompatible tumors; each of the other H loci in isolation had far less influence on the outcome of the transplantation. This point was later formally made in a study involving Snell's congenic strains and tumor as well as normal tissue grafts (COUNCE et al. 1956 ). But the root of the distinction between "major" and "minor" H loci and hence also the origin of the word "major" in the Mhc designation goes back to the early finding that for the growth of most of the established ("old") and rapidly growing (virulent) tumors, only one of the numerous H loci is of significance—H2. Second, Snell obtained evidence that the H2 "locus" was in reality genetically complex, and it was this observation that later led to the use of the word "complex" in the Mhc designation. And third, Snell provided the first indication that the Mhc might be highly polymorphic.

These three features of the Mhc—its strong effect on graft survival, its genetic complexity, and the extraordinary polymorphism of its loci—became the raison d'être for the surge of interest in the Mhc. First, the surgeons, who by then had perfected their techniques of transplanting tissues and organs of almost any kind, felt obliged to acquaint themselves with the system that was spoiling their efforts to replace diseased body parts by healthy grafts. Then the immunologists realized that by understanding the nature of the allograft reaction they could learn a great deal about the entire one arm of the adaptive immune response. Somewhat later still, they came to grasp the real significance of the Mhc as one of three keys to the understanding of the true nature of adaptive immunity (ZINKERNAGEL and DOHERTY 1974 ). Next, geneticists showed more than just a casual interest in the Mhc because the system seemed to offer them the possibility of carrying out fine structure analysis of mammalian genes attainable before then only in phages, bacteria, fungi, and fruit flies. Clinicians began to take note of the Mhc after the first reports describing an association between the possession of particular HLA alleles and the occurrence of certain diseases were published. As the list of HLA-disease associations grew longer and longer, familiarity with the HLA system became a standard component of the medical curriculum. Ultimately, the wave of excitement about the Mhc swept over evolutionary biology, ecology, behavioral sciences, and other disciplines. All these developments in the present era of Mhc studies trace their beginnings back to the two papers published by Snell in 1951. The year 1951 was indeed the annus mirabilis of Mhc studies, and it deserves to be hallmarked as the first year of the Present Era.

View larger version (138K): In this window In a new window Download PPT slide   Figure . George Snell (courtesy of The Jackson Laboratory, Bar Harbor, Maine).

	ACKNOWLEDGMENTS

I thank Ms. Jane Kraushaar and Ms. Lynne Yakes for their assistance in the preparation of this communication.

	LITERATURE CITED

TOP LITERATURE CITED

AMOS, D. B., P. A. GORER, and Z. B. MIKULSKA, 1955  An analysis of an antigenic system in the mouse (the H-2 system). Proc. R. Soc. Lond. Ser. B 144:369-380[Medline].

BRILES, W. E., W. H. MCGIBBON, and M. R. IRWIN, 1950  On multiple alleles effecting cellular antigens in the chicken. Genetics 35:633-652[Free Full Text].

COUNCE, S., P. SMITH, R. BARTH, and G. D. SNELL, 1956  Strong and weak histocompatibility gene differences in mice and their role in the rejection of homografts of tumors and skin. Ann. Surg. 144:198-204[Medline].

GORER, P. A., 1938  The antigenic basis of tumour transplantation. J. Pathol. Bacteriol. 47:231-252.

GORER, P. A. and Z. B. MIKULSKA, 1954  The antibody response to tumor inoculation: improved methods of antibody detection. Cancer Res. 14:651-655.

GORER, P. A., S. LYMAN, and G. D. SNELL, 1948  Studies on the genetic and antigenic basis of tumour transplantation: linkage between a histocompatibility gene and "fused" in mice. Proc. R. Soc. Lond. Ser. B 135:499-505.

KLEIN, J. and D. C. SCHREFFLER, 1971  The H-2 model for the major histocompatibility systems. Transplant. Rev. 3:3-29.

LITTLE, C. C., 1941 The genetics of tumor transplantation, pp. 279–309 in Biology of the Laboratory Mouse, edited by G. D. SNELL. Dover, New York.

OWEN, R., 1989  M. R. Irwin and the beginnings of immunogenetics. Genetics 123:1-4[Free Full Text].

RUSSELL, E. S., 1987  A mouse Phoenix rose from the ashes. Genetics 117:155-156[Free Full Text].

SCHIERMAN, L. and A. W. NORDSKOG, 1961  Relationship of blood type to histocompatibility in chickens. Science 134:1008-1009[Abstract/Free Full Text].

SNELL, G. D., 1948  Methods for the study of histocompatibility genes. J. Genet. 49:87-108.

SNELL, G. D., 1951  A fifth allele at the histocompatibility-2 locus of the mouse as determined by tumor transplantation. J. Natl. Cancer Inst. 11:1299-1305.

SNELL, G. D. and G. F. HIGGINS, 1951  Alleles at the histocompatibility-2 locus in the mouse as determined by tumor transplantation. Genetics 36:306-310[Free Full Text].

WEIR, J. A., 1994  Harvard, agriculture, and the Bussey Institution. Genetics 136:1227-1232[Medline].

ZINKERNAGEL, R. M. and P. C. DOHERTY, 1974  Restriction of in-vitro T cell-mediated cytotoxicity in lymphocytic choriomeningitis within a syngeneic or semiallogeneic system. Nature 248:701-702[Medline].

This article has been cited by other articles:

				K. Paigen One Hundred Years of Mouse Genetics: An Intellectual History. I. The Classical Period (1902-1980) Genetics, January 1, 2003; 163(1): 1 - 7. [Full Text] [PDF]

				J. F. Crow C. C. Little, Cancer and Inbred Mice Genetics, August 1, 2002; 161(4): 1357 - 1361. [Full Text] [PDF]