Cancer immunotherapy: Identifying cancer testis antigen peptides to enhance antitumor response.

Abstract

e20022 Background: Cancer testis antigens (CTAs) are highly immunogenic, tissue-restricted proteins that may be over expressed in hematological malignancies. In this study, an in silico analysis of CTA-derived peptide presentation on HLA class I molecules is reported with an aim to identify putative CTA peptides with a high estimated binding affinity to known HLA class I molecules. Methods: Eighty unique HLA Class I alleles (HLA-A, B, and C), across 78 hematopoietic cell transplant recipients were computationally studied for their ability to bind peptides derived from three CTAs. The amino acid sequences of all three CTA proteins were obtained from UniProtKB. NetMHCPan-4.1 was utilized for in silico prediction of HLA class I binding affinities of the 9-mer peptides generated from these CTAs. Results: The CTA MAGE-A1 had 1202 unique peptides predicted to bind at least one HLA class I allele and 176 of these peptides were classified as strong binders (Rank <0.5 and IC50<500nM). Conversely, 71 unique HLA class 1 alleles were predicted to bind strongly to at least one MAGE-A1 derived peptide. The ability of different MAGE-A1 peptides to bind HLA varied considerably, e.g., peptide TSYVKVLEY, was predicted to bind strongly to 22 different HLA class I alleles. Similarly, 41 CTAG-1B-derived and 164 MAGE-A2-derived peptides were predicted to generate strong binders with at least one HLA class 1 allele. Conversely, 37 and 71 unique HLA class I alleles were predicted to bind strongly to at least one CTAG1B peptide and one MAGE-A1 peptide, respectively. CTAG1-B derived peptide FATPMEAEL and MAGE-A2-derived peptide FAHPRKLLM were predicted to bind strongly to 17 and 22 HLA class 1 alleles, respectively. In aggregate most patients presented multiple peptides across the 6 HLA class I molecules studied (Table). The number of strongly binding MAGE-A1 and CTAG-1B peptides from unique patients. Strong binding defined by Rank <0.5 and IC50<500nM. Conclusions: Most patients presented several CTA derived peptides, suggesting that the aggregate HLA presentation across the 6 classical HLA class I molecules will yield high odds of presentation of CTA expressed by target tissues. These HLA genotype-informed quantitative analyses raise the possibility that agents such as checkpoint inhibitors and immunomodulatory drugs may synergize with hypomethylating agents and augment CTA-directed T-cell responses.[Table: see text]