Immunology Letters, 29 (1991) 9-14 Elsevier

IMLET 01588

Minor histocompatibility antigens El izabe th S impson

Transplantation Biology Section, Clinical Research Centre, Harrow, Middlesex, U.K.

(Accepted for publication 31 January 1991)

1. Summary

Immune responses against foreign tissue or or- gans can be directed against alloantigenic differ- ences between donor and host encoded by genes of the major histocompatibility complex (MHC; HLA in man and H-2 in mouse). However, when MHC an- tigens are matched, as in HLA-identical siblings, or between different mouse strains sharing the same H- 2 haplotype, graft rejection still occurs and is then directed against alloantigenic differences termed minor histocompatibility (H) antigens. Their molecular nature is not yet determined but they are recognised by T cells in an MHC-restricted manner, so are assumed to be derived from molecules co- expressed with MHC class I or II glycoproteins, pos- sibly as peptides or as "super-antigens". The genes encoding them are scattered throughout the genome, including the Y chromosome, on which the H-Y anti- gen gene has been mapped in both man and mouse. One striking feature of minor H antigens is their recognition by T cells but not by antibodies. This made work with them, before our ability to generate T cell responses and maintain T cell clones in vitro, very slow but currently the use of MHC-restricted T cell clones has enabled detailed mapping studies and should eventually allow for their molecular charac- terisation.

Key words: Transplantation; Minor histocompatibility antigen

Correspondence to: E. Simpson, Transplantation Biology Sec- tion, Clinical Research Centre, Watford Road, Harrow, Middle- sex, HA1 3U J, U.K.

2. In vivo recognition of minor H antigens

Histocompatibility (H) antigens were defined by Snell as those against which graft rejection responses are directed when tissues are exchanged between ge- netically non-identical members of the same species. He was working with inbred mouse strains and was interested in the genetics and inheritance of histocompatibility antigens [1, 2]. His approach was to take pairs of different mouse strains and exchange grafts between the parental strains, their F1, F2 and backcross progeny. From an analysis of the graft re- jection tissues in these various combinations, the rules of transplant rejection were established and the number of independently segregating histocompati- bility antigens estimated. The rules were that histocompatibility (H) antigens were inherited and expressed in a co-dominant manner, so that F1 hybrids expressed all H antigens of both parental strains and would therefore not reject parental tis- sues; parental strains would reject tissues from F1 hybrids but also from such a high proportion of F2 mice that it was clear that a number of H genes were independently segregating. Initially, this number was considered less than 10 but subsequent work by Bailey [3], using backcross analysis and the genera- tion of strains congenic for putatively single H anti- gens has increased that number to more than 40. As each H antigen was defined in this genetic analysis, it was given a number, hence H-l, H-2, H-3, H-4 etc. They are the products of genes scattered throughout the genome and have been mapped by co- segregation with other already mapped genes [3]. H- 2 was discovered early because of its ability to induce rapid primary graft rejection responses - a charac- teristic not shared by other H antigens [4]. It was thus termed the major histocompatibility antigen

0165-2478 / 91 / $ 3.50 © 1991 Elsevier Science Publishers B.V. 9

gene, which we now know as the major histcompati- bility complex (MHC) - a genetic region that houses many genes whose products have not only histocompatibility functions but whose physiologi- cal function as self-molecules is to serve as guidance molecules to present exogenous and endogenous ligands to T lymphocytes. The H antigens other than H-2 were then termed minor H antigens, although Bailey has also referred to them as H-non-2 to em- phasise their common original definition by histo- genic analysis. Gorer 's antibodies raised by immuni- sation between inbred mouse strains were found to react with Snell's H-2 antigens and Snell and Gorer together laid the foundation of serological analysis of H-2 genetics and polymorphism [2]. It was tacti- cally assumed that antibodies would be generated to the other H antigens but this expectation was not ful- filled, for reasons that are now somewhat clearer. Thus the analysis of minor H antigens had to depend on classic histogenic methods until in the mid- seventies the growth of T cells in vitro provided an- other tool with which to examine these antigens. However, the pr imary definition of a minor H anti- gen is that it is the target of a graft rejection response in vivo. Not all ligands recognised by T cells in vitro will necessarily also induce in vivo graft rejection responses.

3. In vitro recognition of minor H antigens and en- dogenous super-antigens

MHC-restricted T cell responses to virus were reported by Zinkernagel and Doherty in 1974 [5], and within a few months came reports of similar MHC-restricted T cell responses to minor H anti- gens [6, 7]. Bevan reported MHC-restricted cytotox- ic T cell responses following secondary mixed lym- phocyte culture (MLC) between MHC-matched strains which differed at multiple minor H loci and our laboratory showed MHC-restricted cytotoxic T cells from secondary MLC using spleen cells from fe- male mice immunised with syngeneic male tissue and thus assumed to be specific for the male-specific antigen, H-Y, which had been originally defined by Eichwald's skin grafting experiments [8]. H-Y- specific, HLA-restricted cytotoxic T cells were subse- quently isolated from human peripheral blood lym- phocytes grown in vitro [9] and both human and mouse H-Y-specific T cells have been used to define

10

the chromosomal localisation on their respective Y chromosomes [10, 11]. The generation of H-Y- specific T cell responses in mouse was found to be under immune response (Ir) gene control and both M H C and non-MHC Ir genes determined respon- siveness/non-responsiveness in the various inbred mouse strains [12]. M H C Ir genes were both class I - only certain M H C class I alleles allowed for the associative recognition of H-Y - and class lI, of which possession of the H-2-IA b allele conferred strong responsiveness, although others could sub- serve this class II associative recognition function, implying that the generation of class I-restricted, H- Y-specific cytotoxic cells was dependent on help from class II-restricted H-Y-specific helper T cells [13]. Consistent with this has been the isolation of CD4 ÷ H-Y-specific T cell clones restricted by class II molecules from the same MLC as CD8 ÷ class I- restricted, H-Y-specific T cell clones [14]. However, each H-Y-specific T cell clone may not be responding to the same H-Y epitope; the gene may be complex [15], as has been reported by Roopenian for H-3 and H-4 [16, 17]. These results, defining genetically separable epitopes recognised by class I and class II- restricted minor specific T cell clones in vitro, sug- gest that the minor H antigens defined in vivo by their isolation on congenic strains may obligatorily depend on two linked genes, one whose product is processed through the class I and the other through the class II pathway. In man, both class I and class II-restricted T cells have been isolated in vitro from peripheral blood lymphocytes of patients undergo- ing graft-versus-host (gvh) reactions following bone marrow transplantation (BMT). These include H-Y- specific T cells and T cells specific for polymor- phisms segregating in families and present in some but not all non-related individuals expressing the ap- propriate restricting H L A molecules [18]. It is likely that these T cells are specific for the human homo- logues of mouse minor H antigens. The in vivo defi- nition of minor H antigens in man is very much more difficult than in mice since, in an outbred popula- tion, any MHC-matched pair is likely to differ at multiple minor H loci, and it is difficult to separate the effects of each. However, it can be inferred from the existence of gvh following BMT between HLA- matched pairs and from the continued need to give immunosuppression to recipients of HLA-matched organs to prevent graft rejection, that minor H anti-

gens exist in man and are similar to those in mice. In mice it is clear that in vivo as well as in vitro recog- nition of minor H antigens is MHC-restricted [19].

Another class of alloantigen recognised by T cells in vitro is the self-super-antigens exemplified by Mls- 1 a, Mls-2 a and Mls-3 a. These were originally de- fined by Festenstein as those antigens present in some strains which stimulated vigorous primary mixed lymphocyte responses (MLR) by splenic T cells from H-2-matched mice which did not express them [20]. These in vitro findings implied a high precursor frequency of Mls-specific T cells. We now know this to be the case, since all T cells expressing particular Vfl genes have the capacity to react with the Mls antigens which are presented to CD4 + T cells by class II molecules, not within the peptide binding groove but associated with the external face of the molecule [21]. Mice which express Mls-1 a an- tigen (encoded by a gene on chromosome 1) have amongst their CD4 +CD8 + immature thymocytes some which express Vfl6, Vfl7, Vfl8.1 and Vfl9: how- ever, these thymocytes are clonally deleted, so that virtually no single positive CD4 +CD8- or C D 4 - 8 + thymocytes or any peripheral T cells ex- press any of these Vfl chains [22, 23]. This phenome- non is known as negative selection and is one of the ways in which the T cell repertoire is modified during differentiation o f T cells in the thymus (Table 1) [24]. For Mls-2 a (encoded by a gene which may map to

chromosome 4) and Mls-3 a (encoding gene not yet mapped), the ligands cause clonal deletion of all Vfl3-expressing T cells. There are other examples of endogenous ligands whose expression, in associa- tion with MHC class II molecules, causes deletion of T cells expressing particular Vfl genes (see Table 1): Vfl5 deletion is effected by two ligands, the gene for one of which is on chromosome 12 [25]. VBll can also be deleted by two ligands, one of which is encod- ed on chromosome 6 between Tcrb and Ly-2 [26]. VB17a deletion occurs in all H-2E + strains and the ligand is expressed in a tissue-specific way but the gene has not yet been mapped [24]. Whilst the Mls antigens are "super-antigens" in the sense that they stimulate vigorous primary MLR, there is no evi- dence that the other deleting ligands mentioned are super-antigens in this sense, although their structure and binding affinities to class II molecules may be Mls-like rather than peptide in nature: but this is merely speculative.

There is also no evidence as to whether the Mls an- tigens or any of the TCR V/3 deleting ligands are histocompatibility antigens as defined by in vivo graft rejection responses.

4. Alloantigens recognised by T cells and/or B cells

Table 2 identifies a number of alloantigens show- ing either variant polymorphism or polymorphism

TABLE 1

Molecular nature of minor H antigens.

Endogenous ligands of mice affecting T cell repertoire selection.

TCR chains affected Endogenous ligand Chr. Deletion or positive selection (MHC molecule involved)

vfl6, 7, 8.1, 9 Mls-1 a 1 deletion (A and E) VB3 Mls-2 a ?4 deletion (A and E) Vfl3 Mls-3 a ? deletion (A and E) Vfll7a unknown ? deletion (E) Vfll7a unknown ? positive (K) Vfll 1 unknown 6 negative (E) Vfll I unknown ? negative (E) Vfl5 unknown 12 negative (E) V/34 unknown ? positive (K or D) TGc~fl ) unknown Y positive (9) (D b) TGc~fl ) H-Y Y negative (or) (D b) ? Idiotypic cross-reactive ? negative (A and/or E)

with GT

11

TABLE 2

Alloantigens in mice recognised by T cells and/or B cells.

T cell epitopes B cell epitopes

MHC fl2m Minor H antigens: H-l,

H-3, H-4, H-7, H-25, H-Y Endogenous super-antigens:

Mls-1 a, 2 a and 3 a

MHC f12 m

Thy-1 Ly 1 (CD5) Ly 2 (CDS) L3T4 (CD4)

in expression (expressed vs. null alleles). They are listed according to whether they can be recognised by T cells and/or antibody and a dichotomy is ap- parent for all alloantigens other than those of the MHC or, like fl2m, those associated with MHC molecules [27, 28]. Antibodies generally recognise native protein configurations and the different molecular forms of MHC, Thy-1, CD4, CD5 and CD8 represent no problems. T cells are limited to recognition of MHC/l igand complexes, of which al- logeneic MHC molecules are as good an example as MHC modified by association with a variant f12 m [29], a viral peptide [30], or endogenous ligands, minor H antigens or super-antigens. The subtleties of alteration of MHC molecules by binding exo- genous or endogenous peptides or super-antigens are not apparently detectable by antibodies: again /32 m appears to be an exception, although we do not know that the epitope(s) of/32m recognised by T cells and antibody are the same.

There may be evolutionary reasons for the appar- ent separation of the T and B cell repertoires: for ex- ample, antibodies to viral peptides bound to self- MHC molecules could by stearic hindrance block the generation of protective T cell responses to the same viral epitopes [27]. However, one consequence of this dichotomy is that no antibodies to minor H antigens have been made, and this has slowed down the identification of these antigens at the molecular level and has made it more difficult to do expression studies.

12

5. Molecular nature of minor H antigens: super- antigens and peptides

The absence of minor H-specific antibodies re- stricts the investigation to the use of T cells and to antibodies specific for particular TCR Vfl chains. These have been instrumental in establishing the de- letion of T cells expressing particular Vfl families in mice expressing Mls and other endogenous ligands which may be super-antigens. The studies of Mar- rack and Kappler [21] have suggested that super- antigens are relatively large molecules and are bound to an outer face of class II molecules for presentation to T cells and not in the peptide binding groove. However, although a number of the genes encoding super-antigens and super-antigen-like ligands have been chromosomally mapped [25, 26], they have not yet been cloned. Such a route, although difficult, is possibly the best one to identify their molecular na- ture. A biochemical approach would have to contend with substantial heterogeneity of MHC-bound molecules, making purification and sequencing very difficult indeed.

Endogenous peptides are another category of molecules which may serve as minor H antigens. Re- cent work by Rammensee [31] separating peptides from cell extracts by HPLC and using them to sensi- tise target cells for lysis by minor H-specific cytotox- ic T ceils strongly suggests that his H-Y-specific H- 2b-restricted T cell line and H-4-specific H-2 b res- tricted T cell line are peptide-specific. These peptides may not be minor H antigens as defined in vivo and, until the relevant purified peptides can be sequenced and the encoding genes isolated, we will not know whether the genes correspond to those already chro- mosomally mapped for these two minor H antigens. However, this painstaking biochemical approach is an excellent start to the molecular characterisation of endogenous ligands recognised by T cells and pro- vides evidence for their peptide nature.

6. Possible physiological function of minor H anti- gens/endogenous ligands: selection of the T cell repertoire

The role of Mls and other TCR VI3 ligands in nega- tive selection has already been discussed. Although deletions can be extensive (for example, V/36, Vj37, Vfl8.1 and Vfl9 together are expressed on more than

10°70 of peripheral T cells in mice not expressing Mls- la ) , there is no evidence to suggest such mice are immunocompromised. Indeed, given the inher- ent plasticity of the T cell repertoire, it is perhaps un- likely that functionally significant holes in the T cell repertoire would be thus created, except with respect to responses to exogenous super-antigens, such as Staphylococcal enterotoxins, which also engage T cells expressing particular TCR V~ chains.

Another type of selection, controlled by co- recognition of self-MHC molecules and endogenous ligands, also occurs in the thymus. This is positive selection and leads to expansion of those clones which bind the MHC/l igand in such a way that they are triggered to divide. It is unclear how this signal differs at the molecular level from that of other MHC/l igand interactions causing clonal deletion. Positive selection can involve T cells expressing par- ticular TCR V~3 chains, for example V~17a selected on a class I molecule plus an unknown ligand and V/34 on an H-2 d molecule plus an unknown ligand [32] (Table 1). Positive selection could be a way of skewing the T cell repertoire to give it a propensity towards making immune responses to cross-reactive exogenous antigens. If minor antigens were some of the selecting endogenous ligands, they would thus be candidates for non-MHC Ir genes with a dominant pattern of inheritance.

Positive and negative selection involving T cells expressing certain idiotypes, or antigen specificities, can also occur. Three examples are given in Table l: two involve mice transgenic for expression of the TCR specific for H-Y plus D b [32]. In H-2 b males (H-Y positive) negative selection of virtually all T cells occurs in the thymus whereas in H-2 b females (H-Y negative) positive selection of T cells occurs and transgene-expressing T cells are expanded in the thymus and exported to the periphery. Neither posi- tive nor negative selection occurs in non-H-2 b mice expressing this TCR transgene. This negative selec- tion is clearly an example of a minor H antigen, H-Y, expressed in vivo acting as a T cell repertoire- selecting ligand. The third example involves the con- trol of immune responsiveness to the co-polymer G- T [34]: BALB/c mice are non-responders, DBA/2 (like BALB/c, H-2 d) are responders. The F1 hybrid does not respond. BALB/c and F1 mice fail to re- spond because they express a self-ligand cross- reactive with G-T and this causes deletion of T cells

with receptors that would recognise GT presented on self-MHC class II molecules. BALB/c mice there- fore have a very specific hole in the T cell repertoire created by a self-ligand. This sort of negative selec- tion would be a very powerful and selective way of affecting the T cell repertoire and, if minor H anti- gens functioned in this way, such non- responsiveness, which is a non-MHC Ir gene effect, would be inherited in a dominant fashion, i.e. responsiveness would be recessive.

7. Conclusion

Minor H antigens, defined by their in vivo func- tion to serve as targets of graft rejection responses, are still not understood at the molecular level, although their recognition by MHC-restricted T cells in vivo as well as in vitro suggests that they may be endogenous peptides or super-antigens. In each of these cases, they could function physiologically as non-MHC Ir genes by affecting selection of the T cell repertoire. They would thus share with MHC an- tigens a primary function of regulation of the im- mune response. The function of both MHC and minor H antigens as alloantigens in graft rejection would follow as a consequence of this primary function.

References

[1] Snell, G. D. (1984) J. Genet. 149, 87. [2] Gofer, P. A., Lyman, S. and Snell, G. D. (1948) Proc. R. Soc.

Lond (B). 135, 499. [3] Bailey, D. W. (1975) Immunogenetics 2, 249. [4] Klein, J. (1975) Biology of the Mouse Histocompatibility-2

Complex. Springer-Verlag, New York. [5] Zinkernagel, R. M. and Doherty, E C. (1974) Nature 248,

701. [6] Gordon, R. D., Simpson, E. and Samelson, L. E. (1975) J.

Exp. Med. 142, 1108. [7] Bevan, M. J. (1975) J. Exp. Med. 142, 1349-1364. [8] Eichwald, E. J. and Silmser, C. R. (1955) Transplant. Bull.

2, 148. [9] Goulmy, E., Termijtelen, A., Bradley, B. A. and Van Rood,

J. J. (1977) Nature 266, 544. [10] Simpson, E., Chandler, P., Goulmy, E., Disteche, C. M.,

Ferguson-Smith, M. A. and Page, D. C. 0987) Nature 326, 876.

[11] McLaren, A., Simpson, E., Eppler, J. T., Studer, R., Koop- man, E, Evans, E. P. and Burgoyne, P. S. 0988) Proc. Natl. Acad. Sci. USA 85, 6442.

[12] Fierz, W., Brenan, M., Mullbacher, A. and Simpson, E. (1982) Immunogenetics 15,261.

13

[13] Simpson, E. (1982) Immunol. Today 3, 97. [14] Tomonari, K. (1983) J. Immunol. 131, 1641. [15] Scott, D., McLaren, A., Dyson, J. and Simpson, E. (1990)

Immunogenetics, in press. [16] Roopenian, D. C. and Davis, P. A. (1989) lmmunogenetics

30, 335. [17] Davis, P. A. and Roopenian, D. C. (1990) lmmunogenetics

31,7. [18] Goulmy, E. (1988) Transplant. Rev. 2, 29. [19] Matzinger, P., Zamoyska, R. and Waldmann, H. (1984) Na-

ture 308, 738. [20] Festenstein, H. (1973) Transplant. Rev. 15, 62. [21] Pullen, A. M., Wade, T., Marrack, P. and Kappler, J. W.

(1990) Cell 61, 1365. [22] Kappler, J. W., Staerz, U., White, J. and Marrack, P. C.

(1988) Nature 332, 35. [23] MacDonald, H. R., Schneider, R., Lees, R. K., Howe, R. C.,

Acha-Orbea, H., Festenstein, H., Zinkernagel, R. M. and Hengartner, H. (1988) Nature 332, 40.

[24] Kappler, J. W., Pullen, A., Callahan, J., Choi, Y., Herman, A., White, J., Potts, W., Wakeland, E. and Marrack, P. (1985) Cold Spring Harbor Symp. Quant. Biol. 15,401.

[25] Woodland, D., Happ, M. P., Bill, J. and Palmer, E. (1990) Science 247, 964.

[26] Tomonari, K. (1990) Immunogenetics 32, 60. [27] Simpson, E. (1986) Immunol. Lett. 12, 185. [28] Loveland, B. and Simpson, E. (1986) Immunol. Today 7,223. [29] Rammensee, H. G., Robinson, P. J., Crisanti, M. and Bevan,

M. J. (1986) Nature 319, 502. [30] Townsend, A. R., Gotch, E M. and Davey, J. (1985) Cell 42,

457. [31] Rotzschke, O., Falk, K., Wallny, H. J., Faath, S. and Ram-

mensee, H. G. (1990) Science 249, 283. [32] Tomonari, K., Lovering, E. and Spencer, S. (1990) Im-

munogenetics 31, 333. [33] Kisielow, P., Bluthmann, H., Staerz, U. D., Steinmetz, M.

and Von Boehmer, H. (1988) Nature 333, 742. [34] Vidovic, D. and Matzinger, P. (1988) Nature 336, 222.

14