Proteomic Exploration of the Cell Surface Landscape Reveals New Leukemia Associated Features.

Abstract 2506

Comprehensive identification of leukemia-associated cell surface molecules is required to improve the specificity and sensitivity of flow cytometric analysis of minimal residual disease. A more comprehensive map of leukemia-associated cell surface features could facilitate the selection and subsequent evaluation of new minimal residual disease (MRD) markers as a resource and provide direct cues for new therapeutic targets. Here, we used the chemoproteomic Cell Surface Capturing technology (CSC) to establish a surfaceome map consisting of 807 cell surface detectable proteins detected in an xenograft model derived from diagnostic samples from 19 pediatric patients with acute lymphoblastic leukemia (ALL). Because CSC is based on direct chemical tagging of protein and glycoprotein residues on the surface of living cells, highly viable samples are required to avoid intracellular contaminants. Direct processing of xenograft samples provided optimal conditions for CSC. We included 8 cases with resistant disease based on persistence of minimal residual disease (MRD) during chemotherapy, which represents a group of patients at need for innovative approaches. Comparative analysis of this proteomic dataset showed that CSC recapitulated and expanded the diagnostic immunophenotype of each patient.

To select and test for new proteins with potential value for MRD detection within the large set of identified leukemia proteins, the dataset was filtered against gene expression data from sorted populations of the normal human hematopoietic tree according to their relative RNA expression levels in normal hematopoiesis (DMAP, Novershtern et al, 2011, Cell, 144, 296–309). Based on expected low levels of mRNA expression in normal early and late B-cell precursors in the bone marrow, a subset of markers was identified. These included cell surface features that were previously implicated in leukemogenesis, such as IL7R or FLT3, or shown to serve as diagnostic markers for flow cytometry such as CD58, CD99 and CD300A. For a first clinical validation phase, we selected 38 markers based on their frequency in the CSC dataset to evaluate whether they could contribute to a better distinction of leukemic blasts from their normal counterparts. We tested monoclonal antibodies that were available for 15 candidate markers on diagnostic and remission samples from ALL patients for relevant expression on leukemia cells. Data is currently available for 9 markers that were evaluated prospectively on 51 patients enrolled in current clinical treatment protocols. All markers detected leukemia-associated features in at least a subset of the patients. As anticipated, differences in antigen abundance and antigenic shifts under treatment varied from case to case, underscoring the advantage of increasing the number of available markers for MRD detection. Taken together, our surfaceome data provides an unprecedented view at the cell surface landscape of ALL cells including new prequalified candidate MRD markers, which will accelerate the introduction and subsequent evaluation of multiple parameters for leukemia diagnostics. This also constitutes a valuable resource for functional studies and evaluation of new options for therapeutic targeting.