HLA-Peptidomics and the Identification of Clinical Relevant Minor Histocompatibility Antigens,

Abstract 4038

T-cell recognition of minor histocompatibility antigens (MiHA) plays an important role in the graft-versus-tumor (GVT) effect of allogeneic stem cell transplantation (allo-SCT). However, MiHA recognition is also associated with graft-versus-host disease (GVHD). It is assumed that the selective infusion of T-cells reactive with hematopoiesis-restricted MiHA may help to separate the GVT and GVHD effects of allo-SCT. However, the number of attractive MiHA identified to date remains limited.

In this study we aimed to determine whether it is feasible to identify MiHA using HLA-peptidomics in a reverse-immunology approach, based on bona fide eluted MiHA epitopes. Successful development of such a technology could allow the rapid identification of new MiHA, required to make antigen-selective adoptive T-cell therapy a realistic option. In addition, when compared to classical forward approaches, this strategy may provide tools to efficiently identify favorable GVT-involved MiHA, rather than random identifying targets of activated T-cells isolated during a GVT-response.

To identify biological relevant MiHA candidates, HLA class I peptides were isolated from lysed EBV-transformed B-cells (EBV-LCL), analyzed by mass spectrometry (MS) and matched with a human protein database (IPI). This effort resulted in a set of fifteen thousand peptides, encoded in the normal reading frame with high probability MS scores. To identify potential MiHA candidates, the total set was matched with our newly developed public available Human Short Peptide Variation Database (http://srs.bioinformatics.nl/hspv), dedicated to polymorphic peptides. The quality of this peptide set was demonstrated by a detection efficiency of fifty percent of known MiHA including various length variants and eluted MiHA counterparts. Subsequently the combined use of gene expression databases, validated single nucleotide polymorphism (SNP) arrays and HLA-peptide binding assays resulted in a further selection of 27 high potential HLA-A*0201 and B*0701 MiHA candidates. This set was used for the generation of pMHC tetramers by UV-mediated exchange technology. Next, pMHC tetramer positive specific T-cell lines were generated from eighteen healthy SNP-typed PBMC donors following MACS isolation. To decrease the incidence of isolating low affinity T-cells, due to self-tolerance induction, pMHC tetramer isolations were only performed using donors homozygous negative for the specific SNP. After repeated pMHC tetramer pull down, in vitro expanded cell samples were analyzed on a multi-color FACS LSRII flow cytometer and clonally expanded following FACS cell sorting.

Using this approach we were able to detect 16 unique pMHC tetramer positive T-cell populations corresponding with 70% of eluted MiHA candidates. Most of these pMHC tetramer positive T-cell populations were detected in multiple individuals, and appeared to be oligoclonal. Although most T-cell clones produced IFN-γ when co-cultured with peptide-pulsed target cells, there appeared to be a wide variety of peptide affinity among the pMHC tetramer positive T-cell clones. High throughput screening of all clones for MiHA specific recognition patterns of SNP-typed EBV-LCL panels revealed a clear correlation between the peptide-affinity of the T-cell clone and its capacity to recognize endogenously processed and presented peptide. Collectively these efforts resulted in the validation of two previously described MiHA and the identification of three new biological relevant MiHA.

In summary, this study resulted in the establishment of an algorithm for the high-throughput identification of MiHA based on the combined use of HLA-peptidomics and reverse-immunology by pMHC tetramers. Our data indicate that the technology developed within this project can be of great value to the efficient identification of novel MiHA with potential clinical value especially when epitope selection criteria are supplemented with gene expression data, allowing pre-selection for those MiHA candidates with a hematopoiesis restricted gene expression patterns that may direct reactivity towards GVT.