Interplay Between the Immune System and Colorectal Carcinoma - Towards Tumor-Specific Peptide-Based Vaccination for Any HLA-Type

Introduction: Peptide-basedcancer vaccines targeting tumor-associated antigens comprised of synthetic peptides have shown promising clinical results and have progressed to Phase III clinical testing in several tumor entities, including colorectal carcinoma (CRC). As of yet, these vaccines have typically been limited to HLA-A*02 positive patients, which means that more than half of the patient pool will not be eligible for those treatment strategies. The development of novel strategies for vaccine design, which are applicable to a larger fraction of potential patients, is therefore required. CRC seems to be particularly suitable for antigen-specific immunotherapy as it is among those cancers with highest rates of genetic mutations leading to alterations in protein metabolism. In this study, we aimed to identify HLA class I and II tumor-associated antigens from CRC covering the most frequent HLA-allotypes, in order to develop a peptide warehouse containing HLA-specific peptide panels that can be compiled in a patient-individualized manner, according to the respective patient's HLA-typing.

Methods: The HLA presented immunopeptidome of 30 primary CRC samples and matched autologous non-malignant colon tissue was analyzed after HLA-immunoprecipitation by uHPLC tandem mass spectrometry. Identified source proteins and peptides were cross-evaluated with an in-house database of ligandome data derived from different benign tissues to preclude off-tumor presentation of HLA-ligands in order to to prevent any possible induction of autoimmunity by antigen-specific T cells against these ligands. Furthermore, we evaluated pre-existing antigen-specific T cell responses against these novel CRC-associated peptides in a cohort of 25 CRC patients.

Results: About 17,000 MHC class I presented peptides could be identified on CRC and were cross-evaluated with our in-house database containing 39,000 MHC class I peptides presented on non-malignant tissue. The ligands identified were derived from 7,500 unique tumor-associated proteins and from 11,500 tumor-exclusive source proteins, respectively. Moreover, about 2,000 different MHC class II presented peptides from more than 1,600 source proteins were identified on CRC tissue. Interestingly, MHC class II peptides could contain shorter amino-acid sequences which - after further proteasomal degradation - might also bind to MHC class I molecules. For clinical applicability, we prioritized the identified ligands according to their frequencies of presentation and detectability on CRC. For representation in the HLA-specific peptide sets, a cutoff expression on at least 25% of allotype-matched tumors was required. This guided the assembly of a peptide vaccine warehouse consisting of 40 highly specific CRC-associated HLA ligands. Therefore, this selected set of HLA-specific peptides allows vaccination against CRC for about 95% of Caucasian patients. So far, we detected specific T cell reactivity against two peptides (detection of 10-fold and 17-fold increase of IFNγ-producing T cells representing spots as compared to background in enzyme linked immunospot assays - ELISPOT) in CRC patients.

Conclusions: We here provide for the first time a comprehensive evaluation of HLA-ligandomes in CRC across the most frequent Caucasian HLA-alleles. We developed a HLA-allotype specific warehouse of newly identified CRC associated HLA-ligands, which can be individually assembled according to patient specific HLA-alleles. This approach expands the applicability of peptide-based cancer vaccines to literally every CRC patient.