In collaboration with colleagues in Baylor College of Medicine, we explored the feasibility of generating LAA-specific T-cells from stem cell donors of patients with myeloid leukemia. T-cell lines were manufactured from 10 healthy donors by stimulation with 15mer peptide libraries of five LAA: proteinase 3 (Pr3), preferentially expressed antigen in melanoma (PRAME), Wilms tumor gene 1 (WT1), human neutrophil elastase (NE) and melanoma-associated antigen A3,(MAGE) A3. These are known to be expressed in myeloid leukemias and recently shown by us to be well-represented antigens in AML.53 We previously demonstrated that T cells to WT1, PRAME and Proteinase3 (Pr3) increase after HSCT in association with GVL effects.54,55,56 All lines responded to the mix of five LAA and were multi-specific in interferon-γ ELISpot assay. Although donors showed individual patterns of antigen recognition, all responded comparably to the LAA mix. Known immunogenic peptides of WT1 and Pr3 were identified by epitope mapping in T-cell lines. Experiments showed recognition of partially HLA-matched myeloid leukemia blasts.57 Further studies in ALL showed that WT1-, PRAME- and survivin-specific LAST lysing autologous ALL blasts in vitro could be generated from patients with ALL.58 These findings support the development of a single GMP-grade multi LAST product from donors or patients. While most individuals can generate multispecific LAST, the frequencies of these T cells are about one log lower than those observed using viral pepmixes. Furthermore, there is a risk that the pepmix peptides recognized by T cells are not naturally presented by leukemia. To overcome this concern we created full length lentiviral constructs of WT1 and PRAME to transduce DC to stimulate T cells. In comparison with pepmix-generated CTL, lentiviral-DC induced higher frequencies of CD4+ and CD8+ T-cells, which efficiently recognized lentiviral-transduced targets.
Showing the most recent 10 out of 37 publications
