View abstract data
Abstract title SINGLE CELL NETWORK PROFILING (SCNP) OFFERS A NOVEL APPROACH TO IDENTIFY AT DIAGNOSIS AML CHEMOTHERAPY RESISTANT CELL PHENOTYPES
Author MD, PhD Cesano, Alessandra, Nodality, South San Francisco, United States of America (Presenting author)
Co-author(s) Cohen, Aileen
Covey, Todd
Putta, Santosh
Gayko, Urte
Xiaohua, Xin
Fantl, Wendy
Kornblau, Steven
Topic 3. Acute myeloid leukemia - Biology
Keywords AML, Chemotherapy, Flow cytometry, Signal transduction
Abstract text
Background: Despite improvements in AML response to initial chemotherapy the risk of relapse, especially in older patients, remains high. While chemotherapy regimens can eliminate the majority of the AML blasts, the remaining cells, ultimately responsible for disease relapse, have unique characteristics that render them resistant to the chemotherapy. Current tools to predict disease relapse at diagnosis are inexact. In fact, while many genomic (e.g. gene expression arrays) or proteomic (e.g. reverse phase protein arrays) technologies are capable of revealing the biology of “bulk” AML, biologic tools that detect small chemo-resistant cell populations in AML samples are needed. Single cell network profiling (SCNP) is a biologic tool using multiparameter flow cytometry that allows a comprehensive functional assessment of intracellular signaling pathways in heterogeneous tissues at the single cell level. Aims: Our aims were to use SCNP to determine whether: a) in longitudinally collected AML samples, cell surface phenotypes and/or intracellular signaling profiles dominant at relapse could be identified in subpopulations of cells present at diagnosis and b) whether the presence of (rare) blasts with these intracellular signaling profiles could predict for disease relapse in an independent set of AML diagnostic samples. Methods: Three paired diagnostic and relapse AML bone marrow samples were examined using SCNP after sample incubation with cytokines, growth factors, chemotherapeutic agents, and other modulators. The use of fluorochrome-conjugated antibodies that recognized leukemic blasts and intracellular phospho-epitopes allowed signaling to be measured in specific cell types at the single cell level. In addition, drug transporters and surface receptor levels were also measured. Results: Analysis of the myeloblast subpopulations as defined by surface markers revealed heterogeneity between samples (both at diagnosis and relapse) which was not informative in term of relapse risk. By contrast, the intrapatient characterization of intracellular signaling profiles between relapsed and diagnostic samples revealed in all the three relapsed samples the presence of a subpopulation of leukemic cells characterized by simultaneous phosphorylation of Akt and S6 in response to SCF (SCF:p-Akt/p-S6). This functionally defined leukemic subset, although dominant in the relapse samples, was detectable at a much lower frequency in the diagnostic samples. We hypothesized that the presence of this SCF:p-Akt/p-S6 subpopulation at diagnosis could be predictive for early disease relapse and applied the SCF:p-Akt/p-S6 gate to an independent SCNP data set containing 52 diagnostic AML samples from patients who achieved complete remission (CR) after standard induction therapy. Seven of those patients showed a detectable SCF:p-Akt/p-S6 subpopulation in their diagnostic samples and six of the latter patients experienced disease relapse within 2 years. Of note, the presence of a SCF:p-Akt/p-S6 subpopulation was shown to be independent from the blasts c-Kit (SCF receptor) expression levels. In addition, the SCF:p-Akt/p-S6 profile was independent from patient age, AML cytogenetics and Flt-3 mutational status. Conclusions: Our study showed that longitudinal SCNP analysis of AML samples could provide unique insights into the nature of AML chemo-resistance allowing for identification of subpopulations of cells present at diagnosis with unique signaling characteristics predictive of higher rates of relapse.
Images