Rapid Expansion of Acute Myeloid Leukemia-Reactive Cytotoxic T Cells from CD8+CD62L+ Blood Lymphocytes of HLA-Matched Healthy Donors In Vitro

Abstract

Allogeneic cytotoxic T-lymphocyte (CTL) therapy in acute myeloid leukemia (AML) is hampered by the poor efficiency of growing leukemia-reactive CTLs from healthy donors in vitro. We established an allogeneic mini-mixed lymphocyte-leukemia culture (MLLC) approach by stimulating comparably small numbers (104/well) of CD8+ T cells isolated from healthy donors against irradiated primary AML blasts in 96-well plates. Prior to use, CD8+ T cells were immunomagnetically separated into a CD62L(high)+ subset enriched for naive precursors and central memory cells as well as a CD62L(low)+/negative subset containing effector memory cells. The culture medium contained IL-7, IL-12, and IL-15. After 2 weeks, IL-12 was replaced by IL-2. Mini-MLLCs were performed in seven different healthy donor-AML pairs that were matched for HLA class I according to high-resolution molecular typing. Following 2 weekly re-stimulations with primary AML blasts, mini-MLLC responder populations were analyzed for reactivity on day 19 of culture using split-well IFN-gamma ELISPOT assays. AML-reactive CD8+ T-cell responders were obtained from all 7 donor-AML pairs. The majority of reactive cultures originated from the CD62L(high)+ subfractions. In 4 out of 7 pairs MLLC responder populations mainly recognized AML blasts, but not Epstein-Barr virus transformed B-lymphoblastoid cell lines of donor and patient origin. The AML-reactive CD8+ T cells were restricted by single HLA class I alleles as determined by blocking experiments using a panel of HLA allele-specific monoclonal antibodies. Representative mini-MLLC responders demonstrated strong cytotoxicity against primary AML blasts in 51Chromium-release assay. Cross-reactivity testing identified an HLA-A*0201-restricted CTL population that recognized AML blasts much stronger than non-malignant monocytes of the same patient. This CTL neither recognized recipient-derived primary fibroblasts nor other hematopoietic cells suggesting a leukemia-associated rather than a minor histocompatibility antigen as the target structure. Several MLLC-derived CTL populations expressed unique T cell receptor Vbeta chains consistent with clonal origin from AML-reactive precursors. Multiple CTL responders reached a cell yield exceeding 108 after 6 to 10 weekly re-stimulations with AML blasts. Our results suggest that in healthy individuals most AML-reactive CD8+ CTLs originate from the CD62L(high)+ peripheral blood subpopulation containing naive precursor and central memory T cells. This mini-MLLC approach allows the rapid expansion of AML-reactive CD8+ CTLs from HLA-matched healthy donors to cell numbers sufficient for antigen identification strategies or adoptive immunotherapy trials.