P634: MHC-BASED LARGE-SCALE SCREENING FOR ANTI-TUMOR T CELLS IN CHRONIC LYMPHOCYTIC LEUKEMIA REVEALS CD8+ T CELLS WITH SPECIFICITY AGAINST THE CLONOTYPIC B-CELL RECEPTOR IMMUNOGLOBULIN

Abstract

Background: Chronic lymphocytic leukemia (CLL) is currently incurable, indicating a need for novel strategies towards disease eradication, including reinvigoration of anti-tumor immune responses. T cells in CLL appear selected by restricted antigens, with evidence suggesting that the selecting epitopes may lie within the clonotypic B-cell receptor immunoglobulins (BcR IG). Aims: We previously performed ad hoc prediction of putative T-cell class I neoepitopes contained within the clonotypic BcR IG of CLL patients. Here, we performed major histocompatibility complex (MHC)-based large-scale screening to identify autologous CD8+ T cells recognizing the predicted neoepitopes. Methods: We evaluated 653 peptides derived from the clonotypic BcR IG of 25 CLL patients (median 21 predicted neoepitopes/patient, across 13 MHC-I alleles). Considering the MHC-I typing of each patient, we constructed patient-specific peptide-MHC multimers labeled with a unique DNA barcode plus a fluorochrome (PE). We generated MHC-specific multimer libraries which we then mixed with respective autologous T-cell enriched peripheral blood mononuclear cells (PBMCs). Duplicate samples/patient were analyzed. In addition, known viral peptide-MHC multimers labeled with a different fluorochrome (APC) as well as three MHC-matched healthy donor PBMCs were used as controls. PE- and APC-positive multimer-binding CD8+ T cells were sorted for each cell sample; subsequently, DNA barcode amplification and sequencing was performed. Sequencing data were processed (Barracoda software) to obtain the number of clonally reduced barcode reads assigned to a given sample and peptide-MHC specificity. The number of clonally reduced reads for a given pMHC specificity was used to estimate the frequency of T cells specific for a given epitope, based on the average number of T cell receptor–MHC multimer interactions detected in the total MHC multimer-binding T cell pool in a given cell sample. Peptides with a log fold change >2 were considered as significant [log fold change: (number of DNA barcode reads associated with a specific pMHC)/3x (number of total barcode reads derived from the same sample)]. Results: Overall, 3 peptide-MHC multimers were recognized by CD8+ T cells: (i) VTVADTAVYY (peptide A) and (ii) INLNPSLKRR (peptide B), both within the context of the A03*01 allele, and (iii) YSFTSYWINW (peptide C) within the context of the A24*02 allele. Peptide A derived from the IGHV4-34 FR3 region of a somatically hypermutated clonotypic BcR IG, containing a single A to V somatic hypermutation (SHM) at position 96. Peptide B derived from the IGHV4-39 CDR2-FR3 junction of a somatically hypermutated clonotypic BcR IG, containing 3 SHMs (T to I at position 65, Y to L at position 67 and S to R at position 74). Peptide C derived from the IGHV5-10-1 CDR1-FR2 junction of a clonotypic BcR IG assigned to stereotyped subset #1, containing the sole SHM of the IG (S to N at position 40). The immunogenicity of these peptides was further corroborated by the fact that they were recognized not only by the autologous T cells of the patient from whom the peptide derived, but also by T cells of other patients as well as a healthy donor sharing the respective MHC allele. Summary/Conclusion: In conclusion, we offer proof of concept that the targeted SHM which shapes the CLL BcR IG repertoire may produce idiotypic targets for T cell-based therapy or peptide vaccine design. Characterization of the epitope-binding T cells is currently underway by our group, aiming to provide further insight on how these cells could be recruited into effective anti-tumor responses.