Neonatal Tolerance

Clinical organ transplantation has been a dream for a century or more, but it became a reality only in 1954, when the Boston group successfully transplanted the first kidney into a monozygotic twin. With the development of ever more benign and powerful immunosuppressive drugs and monoclonal antibodies directed against different elements of the immune system, it has become a routine procedure that bestows a new lease of life on the recipient. The heart is no exception. It is, however, hardly necessary for me to add that the astonishingly good results require continuous treatment with immunosuppressive drugs and that the problem of chronic rejection after many years of good function is yet to be resolved. There are, however, well-documented cases where organs have continued to survive without further immunosuppression or where a state of immunologic tolerance had been induced by one of a variety of strategies, most of them of considerable complexity that makes them unsuitable for routine application. One of the mantra of transplantation is that it is highly undesirable to transplant an organ, such as the heart or kidney into ABO-incompatible recipients, in whom the naturally occurring anti-ABO antibodies are likely to bring about the immediate shutdown of the transplant. This catastrophic process was first described for kidneys in the 1960s (1) and is known as hyperacute rejection. However, incompatible kidneys can survive in adult recipients if vigorous steps are taken to reduce the titre of the recipient’s ABO antibodies before transplantation and intensified postoperative immunosuppressive regimens are applied (2). Recently, ABO-incompatible hearts have also been successfully transplanted (3). Naturally occurring antibodies also stand in the way of xenotransplants. Organs, such as kidneys transplanted from one species to another, are rapidly destroyed (hyperacute rejection), and such antibodies continue to be a formidable barrier to the transplantation of, say pig, organs into patients. It therefore came like a bolt out of the blue when L. West’s group announced in 2001 (4) that they succeeded in transplanting hearts into ABO-incompatible infants, and that such hearts had induced a state of B-cell tolerance (2). After plasma exchange, the recipients were given standard immunosuppression. Not only were most of the hearts accepted but there was also a striking absence of antibodies or a significant reduction in their titre. West and her colleagues therefore postulated that the allogeneic hearts had induced B-cell tolerance in the recipients; that is to say, the recipients’ B lymphocytes, which are responsible for the synthesis of antibodies, had been rendered specifically unresponsive (5, 6). This is an extraordinary finding. I will presently tell you about the studies that led to the concept of immunologic tolerance, but I will for the moment simply say that its induction is strictly dependent on the immunologic immaturity of the experimental animals, such as the mouse, in which tolerance can be induced. In the mouse, the period in which tolerance can be induced ends within a few days after birth (7), but the newborn mouse is morphologically extremely undeveloped. My former post-doctoral fellow, Lee Rayfield (now the Bishop of Swindon!), and I set out to determine when immunologic maturity comes about in the human fetus, and we published our data in 1980 in Clinical and Experimental Immunology (8). We were aided and abetted by the obstetrician Charles Rodeck, who provided us with blood samples from fetuses taken by fetoscopy during the third trimester as well as cord blood samples. The assay used by us was the in vitro cell-mediated cytotoxicity assay, which quantitatively measures the reactivity of T-lymphocytes by the release of a 51chromium label from target cells. Our findings were as follows: first, cord blood lymphocytes were almost as reactive as adult blood lymphocytes. Second, fetuses that were 15 to 22 weeks of gestation showed variable responses: half, spread throughout the age range studied by us, were totally negative. A further 23% gave marginal responses. The remaining 37% gave positive responses, but never before the 18th week of gestation. We concluded that although the human fetus is immunologically still pretty immature up to the 18th week, the ability to give a full-fledged T-cell response had developed by the time of birth. In other words, the immune system of the newborn infant is already pretty well developed and unlikely to be susceptible to tolerance induction. I should add that our experiments were entirely concerned with T lymphocyte function and not with antibody production. However, antibody production generally depends on the cooperation of T lymphocytes, and it is therefore not unreasonable to think that our conclusions were true also for antibodies. Indeed, several groups had shown in the 1960s that the human infant is capable of forming antibodies and giving delayed-type hypersensitivity responses (9–11). Our data were certainly in accord with one or two attempts at inducing tolerance to skin grafts in human neonates. Perhaps, the most telling involved the 1960 study of Fowler et al. (12) who transplanted skin grafts from the blood donors used in exchange transfusions in 12 newborn infants who were being treated for jaundice. Although some of the grafts survived somewhat longer than expected, it is quite possible that the clinical condition of the infants was responsible for the slight effect, and the authors concluded that, “The newborn infant already displays a highly developed homograft reaction mechanism.” The data of West and her team are therefore all the more astonishing, but were based on her work in Oxford a few years earlier when she demonstrated neonatal tolerance to cardiac allografts in the mouse (13). I will now give you a brief outline of the tolerance story, in which I have been intimately involved and to show that the induction of immunologic tolerance depends strictly on the immaturity of the immune system. So what is immunologic tolerance all about? It began with some observations by an American immunogeneticist, Ray Owen (Fig. 1), who in 1945 totally out of the blue published data showing that cattle dizygotic twins possess not only their own red blood cells but also some that were patently derived from their twin partners (14). Having discovered that an embryologist—also American—had shown much earlier in the 20th century that dizygotic cattle twins almost invariably develop an anastomosis of their placentae early in gestation, Owen concluded that the twin foetuses must have exchanged red cell precursors, which then had established themselves in their new hosts and churned out red cells of their own lineage in adult life.FIGURE 1: Ray D. Owen, c. 1940.This vital discovery was duly ignored by the immunologic community until in 1948, Burnet (Fig. 2) and Fenner (15), two Australian virologists, built it into their speculations about the existence of a phenomenon of immunologic tolerance. According to them, cellular self-replicating antigens, such as those on living cells, would induce unresponsiveness when introduced into the immunologically immature fetus. All that was missing was proof that such a phenomenon actually existed.FIGURE 2: Macfarlane Burnet.Strangely, Medawar (Fig. 3) and Billingham (Fig. 4), working at the University of Birmingham at that time, quite independently made a relevant discovery—namely, that cattle dizygotic twin calves, which by rights should reject each others’ skin grafts, usually fully accepted them (16). They were at a loss to explain this unexpected finding until Medawar came across the monograph by Burnet and Fenner. The penny dropped and Medawar and Billingham now interpreted their results in terms of tolerance: evidently, the exchange of red blood cell precursors early in fetal life had induced a state of unresponsiveness to the foreign cells that extended also to the skin. They decided that this needed to be experimentally proven, and this was the moment in which Medawar, in whose department I had just completed my Bachelor of Science degree in zoology, recruited me as a member of his small team. The initiation of the experiments, carried out in the mouse as well as in chickens, coincided with Medawar’s 1951 move to University College, London, where he had been appointed the Chair of Zoology.FIGURE 3: Peter B. Medawar seen on the right with Macfarlane Burnet.FIGURE 4: Rupert E. Billingham, c. 1980.Our approach was to inject living nucleated cells, such as spleen cells, into the late mouse foetus, and then to transplant a skin graft from the cell donor to the mice that had received the cells, once they had reached adulthood (17). This was rather hit-and-miss, and we lost a lot of embryos. Later, Billingham and I developed a technique for the intravenous inoculation of cells into newborn mice (18), and because the newborn mouse is generally still immunologically immature, this gave a great boost to our investigations and enabled us to analyze our data. It also led to the discovery of graft-versus-host disease (Fig. S1, SDC, https://links.lww.com/TP/B91), brought about by a response of the transplanted cells against the host’s histocompatibility antigens. Our experiments were made possible by the use of highly inbred mouse strains, so that the cell and the skin donors could be taken from different animals of the same strain. Our early data were published in Nature (17) in 1953, and our magnum opus—a massive paper illustrated by a number of color plates so unusual for the Philosophical Transaction of the Royal Society (19) that publication was delayed by about 6 months—saw the light of day in 1956. (The Philosophical Transactions are about to celebrate a major anniversary, and I have just been told that our article, together with others from other disciplines, has been selected for special commentary in this connection.) I will briefly summarize our main findings. First, the presentation of donor cells late in fetal life or at the time of birth led in many instances to the development of a full-blown tolerance, as evidenced by the long-term acceptance of the donor skin grafts (Fig. 5). This was in stark contrast to the rapid rejection of such grafts in normal mice (Fig. S2, SDC, https://links.lww.com/TP/B91).FIGURE 5: CBA mouse carrying an A strain graft after injection of A strain spleen cells on the 17th day of gestation.Second, tolerance was highly donor-specific, in that it extended to the skin from the cell donor strain but not to the skin taken from a third-party animal (Fig. S3, SDC, https://links.lww.com/TP/B91). Third, the ability of tolerant mice to make antibodies against the antigens of the donor strain was likewise subverted. Fourth, tolerance was not an all-or-none phenomenon, in that some animals tolerated their grafts for limited periods. Fifth, tolerance was usually accompanied by a low level of cellular chimerism, ranging from 1% to 10%, depending on the strain combination. There are those, like the pioneer transplant surgeon Tom Starzl (Fig. S4, SDC, https://links.lww.com/TP/B91), who believe that the development of tolerance in patients is dependent on this microchimerism although it is not easy to distinguish between cause and effect, and I have worked with a model of tolerance induction in adult mice, using cellular extracts, in which chimerism cannot play a role. However, the chimerism of our tolerant mice mirrored the red cell chimerism of Owen’s cattle twins. Sixth, the tolerant state in our mice could be completely undermined by the inoculation of either normal or presensitized lymphoid cells that were syngeneic with the tolerant animals (Figs. S5 and S6, SDC, https://links.lww.com/TP/B91). This told us that the tolerant state was not mediated by a suppressor mechanism but by the clonal elimination of cells with the potential to recognize the donor’s antigens. We happened to be fortunate in the choice of our strain combinations, for in other combinations it was later shown by my group and by others that immunoregulatory mechanisms could be operative, the cells that we called suppressor T lymphocytes but that are nowadays given the more respectable name of immunoregulatory T lymphocytes. Seventh, the capacity for tolerance induction in mice soon disappeared within a few days after birth (18–20) (Fig. S7, SDC, https://links.lww.com/TP/B91), unless the dose of cells was greatly increased, in which case the period of tolerance induction could be extended (21) (Fig. S8, SDC, https://links.lww.com/TP/B91). In other words, antigen load can be an important factor. Eighth, turning to the induction of tolerance in chick embryos, we found that the inoculation of small amounts of donor blood into the venous circulation of 10-day-old to 11-day-old fetuses using a vein of the chorioallantoic membrane, could likewise induce tolerance to skin grafts transplanted later in life (Fig. S9, SDC, https://links.lww.com/TP/B91). What is more, using the ingenious technique of fetal parabiosis developed by a Czech biologist, Milan Hašek (22) (Fig. 6), we were able to reproduce the cattle data almost exactly: the parabionts not only became red cell chimeras but displayed a marked tolerance to each others’ skin (Fig. 7 and Fig. S10, SDC, https://links.lww.com/TP/B91).FIGURE 6: Milan Hašek, c. 1955.FIGURE 7: White Leghorn skin graft transplanted to a Rhode Island Red chicken that had been in parabiosis in ovo from day 10 of incubation with the White Leghorn skin donor. The chicken was not only fully tolerant of its parabiont’s skin but also proved to be a red blood cell chimera.A word about Milan Hašek is appropriate here. He published an article in 1953 (22) in which he described the parabiosis of chick embryos and the effect this had on the ability of the parabionts to produce anti–red cell antibodies later in life. It was significantly reduced. Unfortunately for Hašek, who was at that time a faithful communist subscribing to the wrong-headed ideas of Lysenko on how the environment can directly change the genome, interpreted his data as a form of vegetative hybridisation. It was only when Medawar and I met Hašek at an embryologic congress in Helsinki that Hašek realised that he had misinterpreted his data. Hašek later renounced his communism and paid the price by being deprived of the directorship of the Institute of Biology in Prague. Subsequent to our observation tolerance was observed in many species and for many antigens, including soluble proteins. Not only did it hold out the hope that the barrier to organ transplantation could eventually be overcome, but it also explained why it is that adult animals do not normally respond to their own tissue antigens: they develop tolerance to them in foetal life, and autoimmune responses occur only after mutations or if the antigens had not been present or available to the developing immune system during foetal development. Medawar (Fig. S11, SDC, https://links.lww.com/TP/B91) and Burnet received the Nobel Prize in Medicine and Physiology in 1960, unfortunately excluding Ray Owen (Fig. S12, SDC, https://links.lww.com/TP/B91). Owen was still alive at the age of 98 years but not in good shape at the time I presented this paper in January, when I visited him in his nursing home in January. Anyone interested in the history of immunologic tolerance could do worse than to consult my “A History of Transplantation Immunology” published in 1997 (23) (Fig. S13, SDC, https://links.lww.com/TP/B91) and, more informally, my autobiography “Sunday’s Child. A Memoir,” published in 2009 (24). A rather fine historic overview was published as a Special Feature in this journal to mark the 50th anniversary of tolerance (25). Tolerance in adult animals or in patients depends on the use of quite complex methods that reduce the immune system to a primitive level, encouraging the development of immunoregulatory mechanisms. A major multicenter attempt to compare the efficacy of six different immunoregulatory cells—mainly T lymphocytes and also monocytes—is in progress at the present time (The one study funded by the European Economic Community). Although the remarkable data of Lori West and her team are not easy to explain, they seem to think that full or incomplete B-lymphocyte tolerance was induced by the ABO antigens. This could well be a function of the carbohydrate nature of the ABO antigens, as postulated by them (5, 6). ACKNOWLEDGMENTS The author thanks Prof. Lori West for her invitation to present this article at this year’s ISHLT annual meeting, Dr. Bruce Motyka for his wonderful help in tracking down some references, and Prof. E. Simpson for dragging the author and his ancient slides into the digital age.