Analysis of Wilms' tumor protein 1 specific TCR repertoire in AML patients uncovers higher diversity in patients in remission than in relapsed

• Published in: Annals of hematology 2024-09
• Main Authors: Gielis, Sofie, Flumens, Donovan, van der Heijden, Sanne, Versteven, Maarten, De Reu, Hans, Bartholomeus, Esther, Schippers, Jolien, Campillo-Davo, Diana, Berneman, Zwi N, Anguille, Sébastien, Smits, Evelien, Ogunjimi, Benson, Lion, Eva, Laukens, Kris, Meysman, Pieter
• Format: Article
• Language: English
• ISSN: 0939-5555; 1432-0584; 1432-0584
• DOI: 10.1007/s00277-024-05919-1

The Wilms' tumor protein 1 (WT1) is a well-known and prioritized tumor-associated antigen expressed in numerous solid and blood tumors. Its abundance and immunogenicity have led to the development of different WT1-specific immune therapies. The driving player in these therapies, the WT1-specific T-cell receptor (TCR) repertoire, has received much less attention. Importantly, T cells with high affinity against the WT1 self-antigen are normally eliminated after negative selection in the thymus and are thus rare in peripheral blood. Here, we developed computational models for the robust and fast identification of WT1-specific TCRs from TCR repertoire data. To this end, WT1 (WT1-37) and WT1 (WT1-126)-specific T cells were isolated from WT1 peptide-stimulated blood of healthy individuals. The TCR repertoire from these WT1-specific T cells was sequenced and used to train a pattern recognition model for the identification of WT1-specific TCR patterns for the WT1-37 or WT1-126 epitopes. The resulting computational models were applied on an independent published dataset from acute myeloid leukemia (AML) patients, treated with hematopoietic stem cell transplantation, to track WT1-specific TCRs in silico. Several WT1-specific TCRs were found in AML patients. Subsequent clustering analysis of all repertoires indicated the presence of more diverse TCR patterns within the WT1-specific TCR repertoires of AML patients in complete remission in contrast to relapsing patients. We demonstrate the possibility of tracking WT1-37 and WT1-126-specific TCRs directly from TCR repertoire data using computational methods, eliminating the need for additional blood samples and experiments for the two studied WT1 epitopes.