washed cells were dissolved in NCS solubilizer and counted in a Packard Tricarb Scintillation Counter. Two-way stimulation

Transcription

1 Proc. Nat. Acad. Sci. USA Vol. 68, No. 12, pp , December 1971 Three Closely Linked Genetic Systems Relevant to Transplantation (allotransplantation/cell-mediated immunity/mixed leukocyte reaction/delayed hypersensitivity) E. J. YUNIS AND D. B. AMOS Department of Laboratory Medicine, School of Medicine, University of Minnesota, Minneapolis, Minn ; and Department of Microbiology, Division of Immunology, Duke University Medical Center, Durham, North Carolina Communicated by Maurice B. Visscher, September 7, 1971 ABSTRACT The concept that antigens of a major histocompatibility locus (HL-A), the mixed leukocyte reaction, and the rejection of a graft are expressions of the same genetic region has been generally accepted. We have presented experiments that challenge the above concept and suggest that the rejection time of skin and organ transplants is dependent on immunization against the products of two, and possibly three, separate, but closely linked, genetic systems: HL-A, mixed leukocyte reaction (MLR-S), and hypersensitivity delayed reaction (HDR). The findings presented and discussed here suggest that the HL-A antigens are not the primary factors involved in the cell-mediated component of allotransplantation, and that these antigens and phenotypes of the mixed leukocyte reaction can only be used as a guide in predicting the survival of allografts that are obtained from unrelated donors. The relationship between lymphocyte reactivity in mixed leukocyte culture reactions (MLR) and histocompatibility locus-a (HL-A) has been solidly established in family studies [1, 2]. Cells from HL-A-identical siblings, unlike all other combinations, are usually nonstimulatory in mixed culture. Skin grafts exchanged between HL-A-identical individuals survive longer, on the average, than do grafts between haploidentical or HL-A-different pairs [3, 4]. Kidneys from HL-Aidentical donors require less dosage with immunosuppressive agents to be accepted and exhibit better function than do unselected grafts [5]. The original conclusion drawn from these observations was that the MLR and the response to transplants were attributes of the serologically defined HL-A locus [2]. This article documents some illustrative exceptions, which suggest that the HL-A locus and the MLR locus, although closely linked, are genetically separable entities. MATERIALS AND METHODS HL-A typing was done by a two stage cytotoxicity test [6]. The usual panel contained 166 alloantisera; cells from some subjects, however, were tested with over 400 sera. Reagents included sera used to characterize the HI-A specificities in the Fourth Histocompatibility Workshop (Munksgaard, Copenhagen, 1970) as well as selected sera from the NIH bank and from Drs. Payne, Walford, Tiilikainen, Bashir, Biegel, and Ceppellini. Defined HL-A and additional specific antigens were characterized with two or more- sera that were capable of recognizing each specificity. Mixed leukocyte reactions were performed by a mixture of 106 responder lymphocytes and 2 X 106 stimulatory cells Abbreviations: HL-A, antigens of a major histocompatibility locus-a; MLR, mixed leukocyte reaction; HDR, hypersensitivity delayed reaction ( ite blood cells) in 10% autologous plasma and 90% minimum e~ntlal medium (MEM). The cells were labeled at 7 days by the addition of 2 ACi of [3H]thymidine followed 5 hr later by a flush with cold thymidine [71. The radioactive thymidine used has a specific activity of 1.9 Ci/mol. The packed washed cells were dissolved in NCS solubilizer and counted in a Packard Tricarb Scintillation Counter. Two-way stimulation tests were performed by a similar procedure without mytomycin treatment and by a mixture of 106 lymphocytes from each individual. RESULTS Table 1 illustrates one of three experiments performed with the Sch family. The leukocytes of HL-A-identical siblings A and B are mutually nonstimulatory (of MLR), but the leukocytes of a second pair of HL-A-identical siblings, C and D, who differ by one HL-A allele from A and B, consistently stimulate each other. All four siblings showed stimulation with an unrelated subject, who was phenotypically identical for the HL-A antigens (1,8,2,12) with C and D. This family is also unusual in that the leukocytes from siblings A and B failed to stimulate MLR or to be stimulated by leukocytes from sibling C on two occasions and, as shown (Table 1), stimulated the MLR only feebly on a third occasion. Cells from these siblings were subsequently tested by Dr. Hilaire Meuwissen and also by Dr. Janet Plate with essentially similar results. Mixed leukocyte reaction of an HL-A recombinant sibling of Duke family 0189 is shown in Table 2. The genotype of the recombinant sibling is 9-7/13-LD (A/D-C); the genotype of the AD sibling is 9-7/3-Ao4l and that of the AC siblings is 9-7 Tha-LND. The recombinant did not show mutual MLR stimulation with the AC sibling and showed mutual stimulation with the AD sibling. In family H (Table 3a), the son (B) inherited the haplotypes 2-12 from the mother and 3-7 from the father. He is, therefore, of identical HL-A phenotype (2-12/3-7) to his mother, inheriting one haplotype directly from her and a haplotype serologically indistinguishable from her second allele (haploidentical) from his father. Mutual lymphocyte transformation was produced between these two HL-A haplo-identical, phenotypically identical individuals. A second example of the same phenomenon (Table 3b) is the positive mixed leukocyte reaction between a son in another family, phenotypically identical and haplo-identical (HI-A 2-5/1-7), and his father. The son in this family inherited the 1-7 allele from the father and the 2-5 allele from the mother. A third example (Table 3c) is of stimulation between a male individual and his HL-Aidentical grandfather (Ao77-12/1-8). Further details of these families will be published elsewhere.

2 3032 Medical Sciences: Yunis and Amos TABLE 1. imlr of four siblings from family Sch Sibling (responders) Am* Bm Cm Dm Xm A (1-8/3-Ao7O)t (AC) 282t ,930 B (1-8/3-Ao7O) (AC) ,638 34,482 32,428 C (1-8/2-12) (AD) 660 1, ,430 26,302 D (1-8/2-12) (AD) , , ,403 X (1,8,2,12 unrelated) ,022 20,362 24, * m indicates mytomycin treatment of stimulating cells. t Patient received a kidney allograft from Sibling B, 20 months before MLR. In parentheses, the HL-A genotype or phenotype is given. Siblings A and B are genotype AC, and C and D are genotype AD. t The numbers indicate the mean of triplicate cultures in cpm. Table 4 presents data from two experiments in which three individuals of the phenotype HL-A 1,8,3, and 7; three of the phenotype HL-A 1,8,2, and 12 show mutual MLR stimulation. Other positive mixed leukocyte reactions of identical unrelated HL-A phenotypes obtained in our laboratories include 10 examples in which marked MLR stimulation was obtained between subjects phenotypically: 1,8,2,12 and two each of the phenotypes 1,8,3,7; 1,8,2,5; 2, LND, 9,7; 2,12,3,7; 2, LND, 3,7. Three patients were studied before immunosuppression and 72 hr, 90 days, and 24 months after immunosuppressive cover for kidney transplantation. The immunosuppressive drugs used included: Minnesota globulin from antisera to human lymphoblasts, mg per kg of bodyweight per day for 14 days; Prednisone, 1 mg per kg of bodyweight per day for 3 days, and maintenance of mg per kg of bodyweight per day; Immuran, 5 mg per kg of bodyweight per day for 3 days and maintenance of 2-3 mg per kg per day. The same control donor was used for reference throughout. Lymphocytes of the three patients retained their ability to stimulate the control. but their capacity to respond decreased after transplantation, Autologous plasma was used throughout the investigation. DISCUSSION The HL-A antigens detected serologically behave as if they belonged to two segregant series known as the LA or first and the four or second series [8-12]. From an analysis of the serological reactions with lymphocytes from members of a family, it is possible to deduce the inheritance of the HL-A alleles or haplotypes. Because of the polymorphism of the HL-A region, the paternal haplotypes are usually designated A and B, the maternal haplotypes C and D, and those of the children AC, AD, BC, and BD. If two children inherit the same two haplotypes, e.g., AC and AC, they are said to be HL-A identical [13]. If they share one haplotype only, e.g., AC and AD, they are called haploidentical; if they have neither haplotype, e.g., AC and BD, they are said to be HL-A different [12]. Numer- TABLE 2. MLR of Duke family 0189 Sibling (responders) Am Bm Cm Dm A (3-5/3-Ao41) (BD) 1,250-6,000 5,500 B (9-7/3-LND) (AD/C) - 1,700 1,800 17,000 C (9-7/THaLND) (AC) 13,000 3,100 2,200 25,000 D (9-7/3-Ao4l) (AD) 8,000 16,000 15,000 3,900 Footnotes are as in Table 1. ous studies have confirmed the findings that skin grafts exchanged between HL-A-identical pairs survive longer on the average than grafts between haploidentical and HL-A different pairs [14-16], that MLR between HL-A identical pairs shows no stimulation and that HL-A different pails stimulate MLR more than haploidentical pairs [17, 18]. Other studies still being continued show that kidneys from HL-A identical donors require less immunosuppression and function better than unrelated grafts (5) and that ordinarily, a bone marrow graft from an HL-A-identical donor offers the best possibilities for a replacement therapy [19-21]. These associations are all firmly based and the original conclusion was that HL-A (or Hu 1 as it was first called) was the major histocompatibility locus [1, 12, 22, 23]. The description as given at that time, was of a locus of 15 or more alleles that control MLR, serologic specificities, and transplantation antigens [1, 2]. Evidence has subsequently been reported for the existence of two closely linked loci as determinants for the two series of serologic specificities [10, 11, 24]. It is now necessary to augment these original descriptions by postulating an active region that includes a number of genes of diverse function [25, 26]. Based on the data presented in this article and on data from families described elsewhere, we propose that the two HL-A loci and the MLR locus, although closely linked, are genetically separable entities. From the MLR of the recombinants and from the Sch family, a tentative threepoint mapping of the HL-A-MLR region can be given as first series-second series-mlr loci [25]. Symbols to indicate the alleles at each site are la, and lb (paternal), JC, and id (maternal) for the first HL-A locus; 2A, 2B (paternal), 2C, and 2D (maternal) for the second; and W, X (paternal), Y, Z (maternal) for the MLR-S-locus alleles. The four HL-A-MILR haplotypes are: (A), la 2A W; (B) lb 2B X; (C) IC 2C Y; (D) 1D2DZ. Recombination between the 1st and 2nd HL-A sites would give the new haplotypes la 2B X and lb 2A W for crossingover the paternal gametes and JC 2D Z and JD 2C Y for crossingover in the maternal gametes. These would be serologically distinguishable from the normal haplotypes of that family; stimulation patterns would appear to be controlled by the specificity of the 2nd locus. In Duke family 0189, the C/D recombinant allele carrying HL-A 3 of the 1st series and LND of the second series would be of the recombinational type 1 C 2D Z; thus, this sibling would stimulate all C-carrying siblings but would not stimulate D-haplotype siblings, since they would have the Z allele at the S locus. Recombinants of this type would stimulate both parents. The family reported by Meuwissen et al. [27, 28] and an Amish family studied by

3 Bias and Ward [unpublished studies] appear to be additional examples of this type of recombination. The recombinational frequency within HL-A is somewhat less than 1% [29]. Recombination between HL-A and MLR-S would also give four new haplotyes: JA 2A X and 1B 2B W would be the new paternal gametes and 1C 2C Z and 1D 2D Y the maternal types. This type of recombination also appears to be relatively common. When it occurs, leukocytes from HL-A-identical siblings of the types 1A 2A X and la 2A W would stimulate MLR, whereas leukocytes from haploidentical pairs of the appropriate MLR-S phenotype would now fail to stimulate MLR. In the Sch family, the recombination would have occurred between HL-A and MLR-S loci of child C. Leukocytes from siblings A and B, being of the normal genotype, IA 2A W/1C 2C Y, would not stimulate MLR in each other; sibling C would be 1A 2A W/1D 2D Y. Leukocytes from this child would not stimulate those from siblings A and B since they share the alleles W and Y of the MLR-S, but they would stimulate MLR in sibling D of genotype JA 2A W/JD 2D Z, since the MLR-S genotypes of the two are WY and WZ, respectively. Duke family 0035, previously described, contains 2 HL-A-identical siblings with leukocytes that stimulate MLR (BD genotype) and 2 whose leukocytes do not (BC genotype). Leukocytes from the 2 BD siblings stimulate MLR because one of them carries the recombinant haplotype lb 2B W, thus stimulating the leukocytes of his HL-A-identical sibling, who carries the haplotype lb 2B X. Unfortunately, there is no child of genotype JA 2A W/1D 2D Z to test this hypothesis [30]. Other recorded examples that can best be explained by crossingover, include a Madison family in which leukocytes from HL-A nonidentical (probably haploidentical) fraternal twins fail to stimulate MLR and leukocytes from two other pairs of HL-A-identical siblings who do stimulate TABLE 3. One-way MLR of HL-A identical mother and son (a = Minnesota family H), two-way MLR of HL-A-identical father and son (b = Duke family Q-H), and grandfather and grandson (c = Duke family 0035) Responding cells Am Bm Xm (a) A (2-12/3-7) mother 205 3,421 4,321 B (2-12/3-7) son 3, ,858 X (1,8,2,12) unrelated 17,428 20, Father Son Unrelated (b) Father (1-7/2-5) 1,205 20,809 54,545 Son (1-7/2-5) - 2, ,061 Unrelated (11-W18/Ao54- ) - 4,569 Grandfather Grandson Unrelated (c) Grandfather (Ao77-12/1-8) 420 7,438 34,825 Grandson (Ao77-12/1-8) ,501 Unrelated (9-5/10- ) The HL-A genotypes are given in parentheses. Footnotes are as in Table 1. Genetic Systems Relevant to Transplantation 3033 TABLE 4. MLR of unrelated phenotypically identical 1,3,7,8 and 1,2,8,12 individuals Responding cells Am Bm Cm Dm Xm (a) A (1,8,3,7) 235 6,085 9,201 14, ,808 10,738 29,831 87,094 B 1,8,3,7 19, ,761 50,952 70, ,329 51,999 66,491 C (1,8,3,7) 40,052 8, ,127 40,662 D (1,8,2,12) 33,557 8, ,442 9,249 52,327 28, ,162 X (2,W28,8) 15,203 31,570 29, ,888 83,424 61,127 42, Am Bm Cm Xm (b) A (1,8,2,12) ,472 48,628 90,320 B (1,8,2,12) 113,981 1, ,893 64,879 C (1,8,2,12) 79,350 53, ,319 Z (1,3,7,Ao7O) 110,791 66,993 67, The HL-A 1,8,2,12 phenotypes were determined by identity of 300 antisera, identical strength of cytotoxic reaction and crossabsorption experiments. The experiment summarized in the top table was performed twice. The cells D of the top, and B of the bottom table were obtained from one individual (HN). The rest of the footnotes are the same as in Table 1. MLR [1]. Anomalous reactions that appear not to involve recombination have also been reported. These are examples in which stimulation of MLR only occurs in one direction. Five such examples in otherwise normal subjects have been reported by Seigler et al. [18], three of these are in Duke family Leukocytes from HL-A-identical sisters failed to stimulate MLR in their father and leukocytes from another child failed to stimulate MLR in the mother. In another report, leukocytes from a child failed to stimulate MLR in his father, in another family, leukocytes from a girl did not stimulate MLR in her mother [1]. In all of these families leukocytes of the other parent was stimulated. The simplest explanation is that the parents shared a common MLR-S allele and that the children were phenotypically identical for MLR-S with one of the parents. The HL-A-MLR haplotypes would be of the type la 2A W/1B 2B Z (paternal), 1C 2C Y/JD 2D Z (maternal), and la 2A W/1D 2D Z (child). Leukocytes from such a child would not stimulate MLR in the father but would stimulate the mother. An MLR-S homozygote should fail to stimulate MLR in either parent of 3/4 of his siblings, but should be stimulated by them. If our hypothesis that MLR-S is a separate locus is correct, mutation or deletion of MLR-S should also be possible. Further evidence for the separation of MLR-S from HL-A comes from families in which two members are haploidentical and phenotypically identical, (Table 3). For example, in one of these families, the father and mother had a phenotypically similar haplotype 3-7. A son inherited the 3-7 haplotype from the father and the 2-12 from the mother. Careful typing and absorption of positively reacting sera failed to show any serotypic distinction between mother and son. If the father's HL-A-S haplotype as 3-7-X, then the mother would be 2-12-Y/S-7-Z, and the son would be 2-12-Y/S-7-X, thus accounting for MLR-stimulation between mother and son. By the use of the same model, the MLR of

4 3034 Medical Sciences: Yunis and Amos the reported Minneapolis family can also be explained [27, 28]. The mother carrying the 3-7 haplotype was stimulated in MLR by cells of two of her children that are homozygous 3-7/3-7, because the MLR-S of one of the 3-7 haplotypes is different. The MLR stimulation between phenotypically identical pairs, in the families, or between unrelated individuals, is often attributed to deficiencies in typing. We have tried to rule out this possibility by using the best available typing reagents and reciprocal absorptions. While there are undoubtedly many new subdivisions of HL-A specificities still to be made, the more likely explanation seems to be that MLR-S and HL-A are separate loci and that at the population level, the degree of equilibrium reached between the loci are at present unknown. However, the finding of Eijsvogel et al. [31], Bach et al. [32], and Meuwissen et al. [33] that occasional mixed leukocyte cultures of HL-A-identical, unrelated subjects fail to produce blast transformation suggests a degree of linkage disequilibrium, similar to that found between the LA and four loci. The disequilibrium cannot be of a high order, as evidenced by our failure to find a single example of nonstimulation among unrelated pairs. The preceding discussion has centered around the evidence for the separation of HL-A from MLR-S but the mixed leukocyte reaction appears to depend upon at least two separate factors, one of which we have called MLR-S and the other MLR-R [25]. The evidence for MLR-S has been presented above. By the nature of the specificity of the reactions to the product, it appears to be a membrane-associated material. The nature of MLR-R is less clear. MLR-R (R) appears to be a gene that controls the level of responsiveness. Tentative evidence for MLR-R comes from a consideration of the asymmetrical response of HL-A-identical siblings to cells from the same (allogenic) donor. In Duke family 0099, leukocytes from subject 09 (AD) stimulated MLR in sibling 08 (BD) by 11-fold and MLR in sibling 10 (also BD) only 3.5-fold over background; in family 0013, leukocytes from subject 04 (BD) stimulated MLR in sibling 06 (AC) 40-fold, but leukocytes of subject 07 (AC), stimulatedmlr in sibling 06 (AC) only 10-fold. In other instances studied, the responses were very close (e.g., family 0207 with values of 6.5 and 5.0, etc. ref., 17). It also seems that immunosuppression specifically inactivates the functioning of the MLR-R system, since immunosuppressed successful recipients of kidney grafts may fail to respond, while retaining their ability to stimulate. The possibility that MLR-R is the major gene that regulates lymphocyte responsiveness to alloantigen should be intensively explored since the MLR might give a measure of the ability of man to respond to antigen in the same way that Ir loci regulate the responsiveness of mice [34]. It is important not to confuse MLR reaction with delayed hypersensitivity reaction (HDR). Delayed reactions only develop after adequate exposure to antigen and are a reflection of an aggressive activity of lymphocytes. MLR appears to be quite different. The capacity for MLR-R responsiveness is present in fetal life in humans [35]. MLR-R is even present before birth in the pig, a species that has greater insulation from the outside environment than the human [Plate, Kim, and Amos, unpublished experiments]. Unlike delayed responses, MLR-R is not greatly augmented by presensitization; it appears to be a recognition response rather than a true immunological reaction, whereas the killer cells in a lymphocyte population that responds to an antigenic stimulus may be a minority of the transforming cells. The proportion of cells responding in MLR appears to be considerably greater. The evidence for the existence of yet another specific locus outside the HL-A and MLR-S that controls the production of an antigen recognized only in delayed type responses is mainly circumstantial, but one compelling, clear-cut experiment appears to directly relate to this point. Koch et al. [36] selected five unrelated subjects of the same HL-A phenotype. One of these subjects received skin from the other four and from a fifth donor who was phenotypically distinct. Lymphocytes from two of the phenotypically identical donors were known not to stimulate the recipient in culture, the other three donors stimulated the recipient briskly. The skin from the nonidentical donor was rejected at 11 days, and that from the four HL-A-identical donors was rejected at 12 days. In this series, graft rejection was largely independent of HL-A phenotype and MLR reactivity. It might be argued that factors other than HL-A obscured any difference, but it is hard to accept this as a complete explanation, since differences other than HL-A do not obscure the distinction between HL-A identity and two allele differences in families. The antigens responsible for first-set responsiveness appear to be under the control of independent but linked genes. Once this assumption is made, the distinction between family and unrelated graft relationships becomes apparent. Within a family, HL-A or H-2 gives a very clear indicator of the inheritance of alleles; thus it is possible to select pairs who have the same two haplotypes. Within a family, an HL-A-identical pair will also be MLR-S-identical and HDR-identical. Their cells will type identically and will not stimulate MLR, nor will they exhibit intensive first-set immunity. Unrelated pairs, whether they be mice or men, may be phenotypically identical for H-2 or HL-A but different at MLR-S and HDR. MLR in mice shows that some histocompatibility-2 (H-2)-identical combinations stimulate MLR and others do not. This has generally been taken as evidence of stimulation by factors other than H-2. Most of the comparisons have been between strains, rather than from hybrid crosses. In other species where crosses have been made, minor loci antigens do not stimulate MLR. The mouse data is consistent with the hypothesis that the H-2 alleles are in equilibrium with the HDR and the MLR-S loci [25]. Other studies of the mouse indicate that HDR might be associated with both ends of the H-2 region. Shreffler et al. [37] have postulated that H-2 arose as an inversion and duplication of a segment of a chromosome; thus, the K and D ends of H-2 would be mirror images, distorted somewhat by mutational events within D and K. The MLR-S gene for the mouse lies at the K end but could lie outside H-2 since the H-2 reduplication did not include MLR-S, [38] yet mice differing only by a single D-end specificity can reject grafts [McKenzie, personal communication]. If the reduplication in man also involved an inversion, then we could postulate the map order as HL-A, HDR, and MLR-S, with an overall frequency of recombination of somewhat less than 1%. The hypothesis that three genes are involved in the response to a tissue or bone marrow graft is testable. It will not be easy however, to characterize the HDR locus until tests as specific as those used for the detection of specifically defined antigens have been worked out. The findings described in this paper are important from several points of view in that they: (a) Provide a testable

5 hypothesis for mapping the HL-A region and MLR-S locus as first series-second series-mlr-s, (b) Question the concept that either the MLR or the HL-A phenotype can be used to predict the survival of allografts obtained from unrelated donors, and (c) Suggest that HL-A antigens are not the primary factors involved in the first-set phenomenon of allotransplantation and therefore, attempts to correlate the outcome of an allograft with the degree of HL-A matching will be difficult. The MLR-S products would appear to be of particular relevance in bone marrow or lymphoid grafting, and HDR product would then be primarily responsible for the first-set phenomenon. Of course, HL-A, MLR-S, and HDR, as well as antigens of minor locus systems would all play a role in the rejection of a graft, especially when the recipient has been immunized by blood transfusion or through crossreactivity after exposure to an appropriate antigen. We can predict that HL-A typing will continue to be important in allotransplantation involving related individuals and in attempts to prevent alloimmunization by transfusion of blood, blood products, or other crossreactive antigens. Allotransplantation across the HL-A immunization barrier will always be difficult since it constitutes a mechanism similar to that of the second-set phenomenon. However, in the prospective recipient who has not been immunized, the search for transplantation antigens eliciting the first-set phenomenon must continue. We thank Drs. R. Ceppellini, F. Bach, G. Haughton, J. van Rood, M. Dorf, J. Plate, R. Gatti, H. Meuwissen, H. Seigler, and F. Ward. for their constructive suggestions and stimulating discussions. We also acknowledge the valuable technical assistance of the personnel of our laboratories, especially Mrs. Helen Hallgren, Mrs. Nancy Wood, and Mrs. Donna Kittleson, and the cooperation of all subjects studied, especially the family of Mr. Donald Schlagel. E.J.Y. is supported in part by funds from the Department of Laboratory Medicine of the University of Minnesota. DBA is supported by USPHS grant AI Bach, F. H., and D. B. Amos, Science, 156, 1506 (1967). 2. Amos, D. B., and F. H. Bach, J. Exp. Med., 128, 623 (1969). 3. Ceppellini, R., P. L. Mattiuz, G. Scudeller, and M. Visetti, Transplant. Proc., I, 385 (1969). 4. Stickel, D. L., D. B. Amos, C. M. Zmijewski, J. F. Glenn, and R. R. Robinson, Transplantation, 5, 1024 (1967). 5. Stickel, D. L., H. F. Siegler, D. B. Amos, F. E. Ward, J. C. Gunnels, A. R. Price, and E. E. Anderson, Ann. Surg., 172, 160 (1970). 6. Amos, D. B., H. Bashir, W. Boyle, M. MacQueen, and A. Tiilikainen, Transplantation, 7, 220 (1970). 7. Bach, G. H., and N. K. Voynow, Science, 153, 545 (1966). 8. Payne, R., M. Tripp, J. Weigle, W. Bodmer, and J. Bodmer, Cold Spring Harbor Symp. Quant. Biol., 29, 285 (1964). 9. van Rood, J. J., A. van Leeuwen, A. M. J. Schippers, W. H. Vooys, E. Frederiks, H. Balner, and J. G. Eernisse, Histocompatibility Testing (Munksgaard, Copenhagen, 1965), p Dausset, J., J. Colombani, L. Legrand, and N. Feingold, Nouv. Rev. Fr. Hematol., 8, 841 (1968). 11. Kissmeyer-Nielsen, F., and E. Thorsby, Transplant. Rev. 4, 11 (1970). 12. Ceppellini, R., E. S. Curtoni, P. L. Mattiuz, V. Miggiano, G. Scudeller, and A. Serra, Histocompatibility Testing (Munksgaard, Copenhagen, 1967), p Genetic Systems Relevant to Transplantation Amos, D. B., Transplantation, 5, 1015 (1967). 14. Amos, D. B., B. G Hattler, M. MacQueen, I. Cohen, and H. F. Seigler, in Advances in Transplantation (Munksgaard, Copenhagen, 1967), p van Rood, J. J., A. van Leeuwen, A. M. J. Schippers, R. Ceppellini, P. L. Mattiuz, and E. S. Curtoni, Ann. N.Y. Acad. Sci., 129, 467 (1966). 16. Dausset, J., F. T. Rapaport, P. Ivanyi, and J. Colombani, in Histocompatibility Testing (Munksgaard, Copenhagen, 1965), p Albertini, R. J., and F. H. Bach, J. Exp. Med., 129, 639 (1968). 18. Seigler, H. F., F. E. Ward, D. B. Amos, M. B. Phanp, and D. L. Stickel, J. Exp. Med., 133, 411 (1971). 19. Graw, R. G., Jr., B. G. Levanthal, R. A. Yankee, G. N. Rogentine, J. Whang-Peng, M. H. Ginnig, G. P. Herzig, R. H. Halterman, and E. S. Henderson, Transplant. Proc., III, 405 (1971). 20. Meuwissen, H. J., J. Kersey, H. Pabst, R. Gatti, R. Chilgren, and R. A. Good, Transplant. Proc., III, 41 (1971). 21. Speck, B., L. J. Dooren, J. de Koning, D. W. van Beekum, J. G. Eernisse, F. Elkerbout, J. M. Vossen, and J. J. van Rood, Transplant. Proc., III, 409 (1971). 22. Ceppellini, R., E. S. Curtoni, G. Leigheb, P. L. Mattiuz, B. C. Miggiano, and M. Visetti, in Histocompatibility Testing (Munksgaard, Copenhagen, 1965), p Dausset, J., P. Ivanyi, and D. Ivanyi, Histocompatibility Testing (Munksgaard, Copenhagen, 1965), p Joint report of the Fourth International Histocompatibility Workshop, in Histocompatibility Testing (Munksgaard, Copenhagen, 1970), p Yunis, E. J., J. M. Plate, F. E. Ward, H. F. Seigler, and D. B. Amos, Transplant. Proc., III, 118 (1971). 26. Amos, D. B., and E. J. Yunis, Cell. Immunol., in press (1971). 27. Meuwissen, H. J., R. A. Gatti, P. I. Terasaki, R. Hong, and R. A. Good, N. Engl. J. Med., 281, 691 (1969). 28. Gatti, R. A., H. J. Meuwissen, P. I. Terasaki, and R. A. Good, "Recombination within the HL-A Locus", Tissue Antigens, in press. 29. Svejdgaard, A., A. Bratley, H. L. Heading, C. Jos, F. Kissmeyer-Nielsen, B. Lindblom, A. Lindholm, B. Low, L. Messita, E. Moller, L. Sandberg, L. Staub-Nielsen, and E. Thorsby, Tissue Antigens, k, 81 (1971). 30. Plate, J. M., F. E. Ward, and D. B. Amos, in Histocompatibility Testing, ed. P. I. Terasaki (Munksgaard, Copenhagen, 1970). 31. Eijsvogel, B. P., P. T. A. Schellekens, B. Breur-Briesendorp, L. Koning, C. Koch, A. van Leeuwen, and J. J. van Rood, Transplantation Proc., III, 85 (1971). 32. Bach, F. H., and M. L. Bach, Transplant. Proc., III, 942 (1971). 33. Meuwissen, H. J., P. I. Terasaki, G. Rodey, J. McArthur, H. Pabst, R. Gatti, R. Coifman, and R. A. Good, Histocompatibility Testing (Munksgaard, Copenhagen, 1970), p McDevitt, H. D., and B. Benacerraf, Advan. Immunol., 11, 31 (1969). 35. Ceppellini, R., G. D. Bonnard, F. Coppo, B. C. Miggiano, M. Pospisil, E. S. Curtoni, and M. Pellegrino, Transplant. Proc., III, 58 (1971). 36. Koch, C. T., E. Fredericks, A. van Leeuwen, Progress in Immunology, Cited by J. J. van Rood (Academic Press, 1971). 37. Schreffler, C. S., H. C. David, H. C. Passmore, and J. Klein, Transplant. Proc., III, 176 (1971). 38. Rychlikova, M., P. Demant, and P. Ivanyi, Nature New Biol., 230, 271 (1971).