Background Mutations in splicing factor (SF) genes can cause alternative splicing/mis-splicing of pre-mRNA,leading to aberrant transcripts. These alternative/mis-splicing events include: a. constitutive splicing, b.mutually exclusive exons, c. cassette alternative exons, d. alternative 3’ splice site, e. alternative 5’ splice site,and f. intron retention. Recurrent SF mutations are frequently found in splicing factor genes includingSF3B1, SRSF2, U2AF1, and ZRSR2. These mutations are more common in hematologic malignanciescompared to solid cancers. Mutations in the SF3B1 gene happen to be found in up to 80% of patients witha subtype of Myelodysplastic Syndromes (MDS) characterized by ring sideroblasts. Other cancers withrecurrent mutations in SF genes include uveal melanoma (15-29% in SF3B1), CLL (6-26% in SF3B1),CMML (28-47% in SRSF2; 8-17% in U2AF1), lung cancer (3% in U2AF1), and breast cancer (1-4% inSF3B1). As exemplified in our prior work, disease-causing mutations in SF3B1 (the most commonmutation) alter splicing by disrupting interaction with a protein called SUGP1. Essentially, mutations in theSF3B1 splicing factor gene are the most common/important in MDS, leading to cryptic 3’ ss and upstreamBP, which leads to an RNA gain of function, ~200-1000 mis-spliced transcripts, and intron retention of~18-20 extra nucleotides. From here, the intron retention results in 2 aberrant transcript events, which are1. in-frame transcripts (e.g. those without a stop codon) that are translated/some proteins are degraded. Inthe second event, out-of-frame transcripts are degraded via NMD, but a fraction are translated. Somaticmutations in splicing factor genes (e.g. SF3B1 K700E) from these two events result in degraded proteinsthat may be presented as neoepitopes by the MHC Class I antigen presentation complex to T-cells. Fromhere, one might imagine it possible to establish an immunotherapy target by isolating and identifyingpotential neoepitopes presented by the HLA complex in SF3B1-mutated AML cells. Next, one couldisolate, engineer, and validate a TCR therapy.

Materials and Methods Therefore, we designed a novel approach in which we purifyHLA-presented peptides using a tandem affinity purification method from cells that are engineered toexpress the mutant or wildtype SF protein. To identify the peptide sequence, we ultimately searched themass spectra against a custom de novo assembled transcriptome generated from the RNAseq datasetobtained from the same cells expressing mutant or wildtype SF protein.

Results Ultimately, we found a total of1596 peptides with 615 unique to the mutant.

Conclusions These peptides may prove to be novel immunotherapeutic targets for

AML, and could potentially be used for T-cell therapy or peptide vaccine development.

T. Mitchell: None. T. Yao: None. S. Mukherjee: None. A. Ali: None. A. Yamamoto: None.