Mutation-derived neoantigens represent an important class of tumour-specific, tumour rejection antigens, and are attractive targets for TCR gene therapy of cancer. The majority of such mutations are patient-specific and targeting these requires a fully personalized approach. However, some mutations are found recurrently among cancer patients, and represent potential targets for neoantigen-specific TCR gene therapy that is more widely applicable. Therefore, we have investigated if some cancer mutations found recurrently in hematological malignancies encode immunogenic neoantigens presented by common European Caucasoid HLA class I alleles and can form targets for TCR gene therapy. We initially focused on identifying HLA class I neoepitopes derived from calreticulin (CALR) exon 9 mutations, found in ~ 80% of JAK2wt myeloproliferative neoplasms (MPN). Based on MHC class I peptide predictions, a number of peptides derived from mutant CALR (mCALR) were predicted to bind to HLA-A*03:01 and HLA-B*07:02. However, using mass spectrometry and ex vivo pMHC multimer staining of PBMC from MPN patients with CALR exon 9 mutations, we found no evidence that these peptides were naturally processed and presented on the surface of mCALR-expressing target cells. We next developed a protocol utilizing pMHC multimers to isolate CD8+ T cells from healthy human donor PBMC that are specific for mCALR and additional putative neoepitopes found recurrently in hematological malignancies. Using this approach, CD8+ T cells specific for HLA-A*03:01- and HLA-B*07:02-presented mCALR peptides and an HLA-A*11:01-presented mutant FBXW7 (mFBXW7) peptide were successfully isolated. TCRs isolated from mCALR-specific CD8+ T cell populations were not able to recognize target cells engineered to express mCALR. In contrast, mFBXW7-specific CD8+ T cells were able to recognize target cells engineered to express mFBXW7. In conclusion, while we found no evidence for mCALR derived neoepitope presentation in the context of the HLA class I alleles studied, our data suggests that the recurrent pR465H mutation in FBXW7 may encode an HLA-A*11:01 presented neoepitope, and warrants further investigation as a target for T cell based immunotherapy of cancer.Electronic supplementary materialThe online version of this article (10.1186/s40425-018-0386-y) contains supplementary material, which is available to authorized users.

Immunotherapies targeting truly tumor-specific targets focus on the expansion and activation of T cells against neoantigens or oncogenic viruses. One target is the human papilloma virus type 16 (HPV16), responsible for several anogenital cancers and oropharyngeal carcinomas. Spontaneous and vaccine-induced HPV-specific T cells have been associated with better clinical outcome. However, the epitopes and restriction elements to which these T cells respond remained elusive. To identify CD8+ T cell epitopes in cultures of tumor infiltrating lymphocytes, we here used multimers and/or a functional screening platform exploiting single HLA class I allele-engineered antigen presenting cells. This resulted in the detection of 20 CD8+ T cell responses to 11 different endogenously processed HLA-peptide combinations within 12 HPV16-induced tumors. Specific HLA-peptide combinations dominated the response in patients expressing these HLA alleles. T cell receptors (TCRs) reactive to seven different HLA class I-restricted peptides could be isolated and analysis revealed tumor reactivity for five of the six TCRs analyzed. The tumor reactive TCRs to these dominant HLA class I peptide combinations can potentially be used to engineer tumor-specific T cells for adoptive cell transfer approaches to treat HPV16-induced cancers.