HLA matching has played a pivotal role in the success of clinical kidney transplantation, especially in the early years when less effective immunosuppression was available. On the population level, it has clearly been shown that an increasing number of HLA antigen mismatches is associated with inferior graft survival. Recent analyses by the Collaborative Transplant Study still demonstrate the existence of such an HLA matching effect,1 although the differences in graft survival are smaller due to more potent immunosuppressive medication. Another important consequence of a transplant with HLA mismatches is the likelihood to develop donor specific antibodies (DSA), which makes the search for compatible donors difficult in case a second transplant is necessary.
In the meantime, it has become clear that just counting the number of HLA mismatches is not a very accurate way to estimate the risk for immunological graft rejection in an individual patient. Patients transplanted with many HLA mismatches may do well whereas much better HLA matched grafts fail. Apparently, there are also qualitative aspects involved and one of these is the fact that the immunogenicity of individual HLA mismatches differ.2 This differential immunogenicity is reflected by the chance that recipients will produce DSA after transplantation. Recent studies demonstrate a crucial role for certain amino-acid configurations on the donor HLA, often referred to as eplets or epitopes.3,4 Every HLA allele consists of a unique string of eplets, while the individual eplets may be shared by other HLA alleles. By aligning donor and recipient strings, one can determine the number and the position of eplets on donor HLA, which are not present on patient HLA. It appears that the number of foreign eplets on the donor HLA is predictive for the chance that patients will develop de novo DSA and as a consequence have a poorer graft survival. Several articles show that this mismatched eplet load is a better parameter to establish risk to develop DSA than HLA antigen mismatches, which is an important reason to introduce this principle in the clinical transplantation setting as suggested in this issue by René Duquesnoy, the expert in the field.5
However, one should realize that eplet load is just the beginning of a story, which will be very similar to the one on HLA antigen matching in transplantation. Just counting the number of mismatched eplets is useful for calculating the chance to develop DSA on the population level but is certainly not the optimal way to assess risk for an individual patient. Similar to the situation of HLA antigen mismatches, the immunogenicity of individual eplets is different. The international registry of HLA epitopes (www.Epregistry.com.br) has records of theoretical eplets and of antibody verified eplets, which, so far, have been shown to be actual targets for antibodies. It is to be expected that the latter group will increase as more studies are performed. The ensuing challenge is to identify or even predict the most immunogenic eplets for an individual patient, which is not feasible yet. Therefore, the current introduction of a match strategy based on eplet load will only be based on statistics reasoning that a higher eplet load increases the chance that one of these eplets is an immunogenic one in relation to the immune repertoire of the recipient. Hence, using cut-offs for eplet load on the level of the individual patient does not suffice.
Importantly, the immunogenicity of an eplet does not only depend on its molecular characteristics in relation to the B cell repertoire of a patient. The production of IgG DSA requires help of CD4+ T cells to the eplet specific B cells, which is based on recognition of allogeneic peptides derived from the mismatched HLA antigen in context of the HLA class II molecules on the B cell. Also for this aspect a tool has been developed, which predicts the number of peptides derived from the mismatched HLA molecule(s) that theoretically can be presented as T cell epitopes by the HLA class II molecules on the B cells of the recipient. This Predicted Indirectly ReCognizable Epitopes tool,6 measuring the potential peptide load of the mismatched HLA molecule, seems to be somewhat predictive for the chance to develop DSA on the population level but again for an individual patient this tool is hardly useful. Also, here the challenge is to discriminate the real immunogenic peptides from the irrelevant peptides, which do not serve as epitopes.
Once a patient has developed HLA antibodies, proper knowledge of the target epitopes will enable the prediction of acceptable and unacceptable mismatches facilitating virtual cross matching. Although the eplet which was involved in the induction of the antibodies plays a crucial role in the interaction between antibody and HLA molecule, the final reactivity and affinity of the antibody is based on additional interactions between several contact sites on the antibody molecule (complementary determining regions) with amino acid configurations in the neighborhood of the immunizing eplet.5 To be able to predict in a reliable way which HLA alleles will and which will not react with the antibody, the crucial interactions between antibody and antigen defining the actual epitope must be analyzed in detail. Human monoclonal antibodies7 and antibodies purified by absorption and elution8 have proven to be suitable tools for such analyses and are the basis of the detailed descriptions of several antibody verified epitopes included in the HLA epitope registry. Recently, sera from pregnant women have also been applied for this purpose9 but such reagents carry the risk of being a mixture of different antibody specificities leading to a less reliable definition of epitopes, especially when sera from multiparous women are used.
One striking aspect of several antibody verified epitopes is the fact that they consist of a nonself eplet in combination with a crucial self-component. It is suggested that this is due to the fact that the immune repertoire of B cells has a low avidity for self-HLA,5 a situation similar to the T cell repertoire. However, T cells are educated in the thymus to recognize peptides in the context of self-HLA whereas B cells should be able to react with any foreign structure without any involvement of self-HLA.10 A more likely explanation for the inclusion of self-structures in the actual antibody epitope is the fact that different HLA molecules have a similar structure and function and therefore share amino acid sequences. If an antibody is induced by an allogeneic eplet, other potential contact sites for the antibody in the environment of this eplet are therefore, by chance, often self-components.
In conclusion, we agree with René Duquesnoy that epitope-based matching can already be applied in the clinical setting (actually, we are already doing so in the acceptable mismatch program of Eurotransplant) but a lot of fine-tuning will be necessary to obtain an optimal match strategy, which benefits all individual patients.
1. Süsal C, Opelz G. Current role of human leukocyte antigen matching in kidney transplantation. Curr Opin Organ Transplant
2. Claas FH, Dankers MK, Oudshoorn M, et al. Differential immunogenicity of HLA mismatches in clinical transplantation. Transpl Immunol
3. Duquesnoy RJ. HLA epitope based matching for transplantation. Transpl Immunol
4. Kosmoliaptsis V, Mallon DH, Chen Y, et al. Alloantibody responses after renal transplant failure can be better predicted by donor-recipient HLA amino acid sequence and physicochemical disparities than conventional HLA matching. Am J Transplant
5. Duquesnoy RJ. Are we ready for epitope-based HLA matching in clinical organ transplantation? Transplantation
6. Otten HG, Calis JJ, Keşmir C, et al. Predicted indirectly recognizable HLA epitopes presented by HLA-DR correlate with the de novo development of donor-specific HLA IgG antibodies after kidney transplantation. Hum Immunol
7. Marrari M, Mostecki J, Mulder A, et al. Human monoclonal antibody reactivity with human leukocyte antigen class I epitopes defined by pairs of mismatched eplets and self-eplets. Transplantation
8. El-Awar N, Nguyen A, Almeshari K, et al. HLA class II DQA and DQB epitopes: recognition of the likely binding sites of HLA-DQ alloantibodies eluted from recombinant HLA-DQ single antigen cell lines. Hum Immunol
9. Duquesnoy RJ, Hönger G, Hösli I, et al. Identification of epitopes on HLA-DRB alleles reacting with antibodies in sera from women sensitized during pregnancy. Hum Immunol
10. Novotny J, Bruccoleri RE, Kourilsky P. On the molecular nature of “restrictive” antigenic elements present on major histocompatibility complex (MHC) proteins. Res Immunol