Single-Cell Profiling of Aberrant Cytokine Signaling in Patients with Chronic Myeloid Leukemia (CML) at Diagnosis and during Dasatinib Therapy

Aberrant cytokine and growth factor signaling is the hallmark of CML and results from constitutive oligomerization of the oncogenic BCR-ABL tyrosine kinase (TK). Inhibition of BCR-ABL by imatinib mesylate is the current standard of care of CML and results in durable responses in majority of patients. However, a proportion of patients shows primary or secondary resistance to imatinib, which can be attributed either to selection of clones harboring mutations in the kinase domain of BCR-ABL or activation of a BCR-ABL independent pathway. Dasatinib, a potent multikinase inhibitor, can rescue some imatinib-resistant patients, but carries an increased risk of adverse effects due to inhibition of off-target wild-type kinases, particularly in immune effector cells. In concord, recent in vitro data indicate a profound immunosuppressive effect of dasatinib. The aim of this study was to analyze and predict TK inhibitor (TKI) resistance and off-target effects using single-cell profiling of aberrant phosphoprotein networks upon cytokine stimulus by multiparameter flow cytometry. The study cohort consisted of 5 healthy controls, 4 non-treated CML patients at diagnosis and 5 CML patients on dasatinib therapy and in cytogenetic remission. Stimuli included GM-CSF, IL-4+IL-6+IFNγ and IL-2+IL-10+IFNα and they were added to freshly drawn whole blood or bone marrow. The readout phosphoproteins were pERK1/2, pSTAT1, pSTAT3, pSTAT5a and pSTAT6 (with isotype controls), and were analyzed separately from granulocytes, monocytes, CD3+, CD4+ and CD8+ lymphocytes and regulatory T-cells. In unstimulated blood samples from healthy controls the phosphoproteins were essentially unphosphorylated. The responses to cytokines were consistent among individuals resulting in phosphorylation of ERK1/2, STAT3 and STAT5a on GM-CSF stimulus, STAT-1, STAT-3 and STAT-5a on IL-2+IL10+IFNα and STAT-1, STAT-3 and STAT-6 on IL4+IL6+IFN-γ. Compared to healthy controls, increased baseline phosphorylation of STAT-1, STAT-3 and STAT5a, but not ERK1/2 was seen in CML patients at diagnosis, especially in myeloid cell lineages (neutrophils/monocytes), but also in lymphocyte subgroups. The responses to cytokine stimulation were modest overall, in particular the ERK1/2 responses to GM-CSF were absent. This indicated the inactivation of the Ras/MEK/MAPK pathway and saturation of other BCR-ABL downstream pathways. Already at diagnosis, the phosphorylation pattern of a TKI primary resistant patient differed profoundly from the responding patients. Marked activation of STAT-1 and STAT-3 was seen in granulocytes and monocytes stimulated either by GM-CSF or by combination of IL2+IL10+IFN-α, suggesting activation by a pathway circumventing BCR-ABL. In dasatinib treated patients, the baseline activation status was similar in granulocytes and monocytes and slightly diminished in lymphocytes when compared to healthy controls. Similarly, the responses to cytokines resembled those seen in healthy controls, in contrast to published in vitro data. Remarkably, in some of the dasatinib treated patients, STAT1 and STAT3 responses were even more pronounced than in healthy controls. This underlines the importance of studying the in vivo/ex vivo effects of TKIs on off-target kinases, in particular of drugs with a short half-life such as dasatinib. In conclusion, inter-individual differences in TKI response and immunomodulatory effects of pan-TKI dasatinib can readily be discerned by analyzing key intracellular phosphoprotein responses to cytokine and growth factor stimuli ex vivo. The method allows profiling of aberrant signaling pathways in different subsets of leukocytes in CML patients and can be used to predict TKI resistance and spectrum of potential adverse effects due to inhibition of wildtype targets. Similar analyses of signaling pathways at the stem cell level are ongoing and may aid in understanding TKI resistance of CML stem cells.