The simultaneous administration of G-CSF and chemotherapy as a priming strategy has resulted in a clinical benefit in determined subsets of patients diagnosed with acute myeloid leukemia (AML) (Löwenberg et al, NEJM 2003; Pabst T, et al, Blood 2012). However, the mechanism responsible for this anti-leukemic effect is not fully characterized. We hypothesize that the clinical benefit may occur at least partially by the effect of G-CSF on leukemic stem cells (LSC).


The main goal of this project was to determine the effect of G-CSF on primary AML samples in vitro, especially on LSCs.

Methods and patients

Peripheral blood mononuclear cells (PBMC) from 10 AML patients were treated with G-CSF at increasing doses, alone or in co-culture with HS-5 stroma cells. Cell viability (7-AAD -eBioscience- cell death exclusion and volumetric cell counting) and surface phenotype was determined by flow cytometry (FACSVerse, BD) 72 hours after treatment. Data were analyzed using the FlowJo (Trastar) software. For clonogenicity assays, AML primary samples were treated for 18 hours with G-CSF at increasing concentrations and cultured in H4034 Optimum MethoCult (StemCell Technologies) for 14 days. Colonies were counted based on cellularity and morphology criteria.


G-CSF treatment showed no effect on cell viability of the bulk leukemic population or on the CD34 + immature subpopulation. A dose-dependent increase in CXCR4 surface expression was observed, reaching a 1.4-fold of change at the highest concentration of G-CSF (100 μg/mL). In contrast, treatment of leukemia cells with G-CSF in the presence of stroma cells reduced the overall cell viability. Thus, a 32% decrease of cell viability was measured at the highest concentration used (p = 0.0006), while no significant changes in the frequency of each leukemic subpopulations were observed. Clonogenic capacity was significantly reduced in a dose-dependent manner upon treatment with G-CSF, achieving a 41% reduction at the highest G-CSF concentration (100 μg/mL).


G-CSF reduces the viability of leukemic cells when these cells are in co-culture with the HS-5 stroma cell line, suggesting that the presence of stroma cells is required for the cytotoxical effect of G-CSF on the blast population. Interestingly, G-CSF treatment decreased the clonogenic capacity of AML samples, therefore suggesting that G-CSF exerts its effect at least partially on LSCs. Our findings support the design of studies to explore new strategies of chemotherapy priming in AML patients.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.