A Shared SF3B1 Neoantigen Is Presented on Primary Malignant Cells and Induced Pluripotent Stem Cell-Derived Hematopoietic Lines

Myeloid malignancies are susceptible to immunologic destruction but antigen-specific T cell immunotherapies for these diseases are currently limited. Immunotherapies targeting neoantigens created from recurrent protein-coding mutations should selectively eradicate malignant cells and spare their normal counterparts to limit the risk of myeloablation. In particular, the paucity of neoplastic cells in MDS specimens impedes the identification and validation of targetable neoantigens for the development of immunotherapies for this disorder. Induced pluripotent stem cells (iPSC), generated from primary patient cells and differentiated into hematopoietic lines, can recapitulate the genotype and phenotype of the original disease (Hsu et al., Blood 2019) and could provide a renewable source of MDS progenitor cells for the development of novel T cell immunotherapy. SF3B1 mutations serve as a neoplastic driver for myeloid neoplasms (Papaemmanuil et al., Blood 2013), occuring in 20-30% of all MDS and >60% of MDS with ringed sideroblasts (Cazzola et al., Blood 2013). We hypothesized that SF3B1 mutations produce HLA-presented neoantigens, which can be targeted to eliminate primary MDS cells and iPSC with SF3B1 mutations.

We interrogated the amino acid sequences from protein-coding mutations in SF3B1 using HLA-binding prediction algorithms (Immune Epitope Database Stabilized Matrix Method and Artificial Neural Network, netMHCpan 4.0) to predict binding to 20 prevalent HLA-A, -B, and -C alleles, and identified a candidate SF3B1^mut epitope with strong predicted binding to HLA-B*40:01. HLA binding is necessary but not sufficient for T cell recognition of an epitope. We thus next evaluated the immunogenicity of the candidate SF3B1^mut neoantigen epitope by stimulating CD8⁺ T cells isolated from healthy volunteer donors with peptide epitope-loaded autologous monocyte-derived dendritic cells, and isolated SF3B1^mut -specific T cell clones, confirming the candidate epitope’s immunogenicity. Two clones showed high functional avidity for the epitope in cytotoxicity assays (Figure 1A). To determine whether the epitope was naturally processed and presented by primary malignant myeloid cells, high-avidity SF3B1^mut-specific clones were co-cultured with hematopoietic cells from patients with active MDS or acute myeloid leukemia (AML), then assessed for antigen recognition in a CD107a degranulation assay. High-avidity clones showed increased degranulation in response to a primary SF3B1^mut HLA-B*40:01⁺ sample compared to samples lacking either the mutation or restricting HLA (Figure 1B), indicating that the SF3B1^mut epitope is naturally processed and presented. We then evaluated whether SF3B1^mut-specific clones could recognize SF3B1^mut iPSC-derived hematopoietic cells. We reprogrammed hematopoietic stem/progenitor cells (HSPC) from SF3B1^mut MDS patient bone marrow and established two iPSC lines: SF3B1^mut and an isogenic control with wild-type SF3B1. Multipotent hematopoietic progenitor lines (MPP-5F) were then generated from iPSC lines by doxycycline-dependent expression of five HSPC transcription factors (Hsu et al., Blood 2019). In a CD107a degranulation assay, high-avidity clones showed ~30% degranulation with SF3B1^mut MPP-5F, compared to ~5% with the isogenic control (Figure 1C), indicating that the iPSC-derived MPP-5F line recapitulated presentation of the neoantigen.

We present a novel neoantigen with promise as an immunotherapy target for MDS, and for other hematologic malignancies with SF3B1 mutations including secondary AML, MDS/myeloproliferative neoplasm overlap syndrome, and advanced chronic lymphocytic leukemia (Cazzola et al., Blood 2013). Future experiments will assess the efficacy of SF3B1^mut-specific T cell immunotherapy in vivo in a murine model and evaluate transfer of SF3B1^mut-specific T cell receptors as a potential therapeutic approach. Our studies also indicate that iPSC-derived MPP-5F lines have potential to substitute for primary patient cells in neoantigen validation studies, a finding with broad utility for immunotherapy development in hematologic malignancies. Ongoing studies will assess MPP-5F presentation of a diverse panel of antigens, including minor histocompatibility antigens, non-mutated cancer antigens, and other neoantigens.