Distinct Proteomic Signaling Networks Stratify Responses to Ara-C Based Induction Therapy in Patients with Acute Myeloid Leukemia (AML)

Background: AML represents a group of clonal hematopoietic stem cell disorders in which aberrantly regulated signaling pathways lead to oncogenic progression. In a prior study, a panel of 36 cytokine response nodes was measured in single cells by multi-parameter phosphoflow cytometry in newly diagnosed AML patient samples. Unique cancer network profiles were revealed that correlated with response to chemotherapy (Irish et al, Cell (2004) 118, p217).

Objectives: Given the heterogeneity of AML, the current study was designed to expand the number of signaling nodes to 152 per patient sample. The signaling nodes were organized into 4 biological categories:

1. Protein expression of receptors and drug transporters

2. Response to cytokines and growth factors,

3. Phosphatase activity,

4. Apoptotic signaling pathways.

Methods: Multi-parameter flow cytometric analysis was performed on peripheral blasts taken at diagnosis from 33 AML patients who attained a complete response (CR, n=9) or no-remission (NR, n=24) to one cycle of standard 7 + 3 induction therapy (100–200mg/m² cytarabine and 60mg/m² daunorubicin).

Results: The data show that expression of the receptors for c-Kit and FLT3L and the drug transporter ABCG2 were increased in patients who attained an NR versus CR. Readouts from the cytokine-Stat response panels and the growth factor-Map kinase and PI3-Kinase response panels (see Table 1) revealed increased signaling in blasts taken from NR patients versus blasts taken from patients who clinically responded to therapy. To determine the role of phosphatases, a physiologic phosphatase inhibitor, peroxide, (H²O²) revealed increased phosphatase activity in CRs versus NRs. In the absence of treatment with H²O², CRs had lower levels of phosphorylated PLCƒ×2 and SLP-76 versus NRs, and attained higher levels of phosphorylated PLCƒ×2 and SLP-76 upon H²O² treatment. Lastly, interrogation of the apoptotic machinery using agents such as staurosporine and etoposide showed that NR patient blasts failed to undergo cell death, as determined by cleaved PARP and cleaved Caspase 8. Of note, in NR patient blasts, these agents did promote an increase in phosphorylated Chk2 suggesting a communication breakdown between the DNA damage response pathway and the apoptotic machinery. In contrast, blasts from CR patients showed significant populations of cells with cleaved PARP and caspase 8 consistent with their clinical response outcomes.

Conclusions: In this study, 152 signaling nodes per patient sample were measured by multi-parameter flow cytometry and revealed distinct signaling profiles that correlate with patient response to ara-C based induction therapy. Alterations were seen in expression for the c-Kit and Flt-3L receptors, the ABCG2 drug transporter, cytokine and growth factor pathway response, phosphatase activity and apoptotic response, all of which could stratify the NR from the CR patient subsets. Whether there is a combination of hierarchical nodes to best predict response to therapy is currently under investigation.