CML cells gain a survival advantage over normal cells due to dysregulated tyrosine kinase activity of BCR/ABL protein, which phosphorylates a number of proteins, potentially activating multiple signal transduction pathways. Imatinib specifically targets BCR/ABL protein and results in anti-proliferation and apoptosis. Even in the absence of mutation, the leukaemic clone may not be eradicated by imatinib, suggesting that leukaemic stem cells may not be absolutely reliant on BCR/ABL activity for survival. Reports on GM-CSF production by CML cells suggest a possible autocrine role, therefore we examined if GM-CSF could protect CML CD34+ cells from imatinib. CD34+ cells from peripheral blood of patients with CML in chronic phase (n=9) and from normal bone marrow donors (n=5), were labeled with CFSE (Carboxy-Fluorescein diacetate Succinimidyl Ester) to enable tracking of cell division. Normal and CML samples were cultured with and without GM-CSF (300pg/mL) to assess response to this cytokine as a single agent. CML CD34+ cells were cultured for 3 days in serum deprived medium with imatinib only, GM-CSF (300 pg/mL) + imatinib, GM-CSF (300pg/mL) + E21R (a GM-CSF analogue able to block cytokine binding) (10mg/mL), and GM-CSF (300pg/mL) + E21R (10mg/mL) + imatinib. In each condition, imatinib was titrated over the range of 0 to 10 microM. Cultures were analysed by flow cytometry to evaluate the proliferation index (PI), a ratio of final cultured cell to precursor cell number, where a PI of 1 represents no proliferation, and a PI of 2 corresponds to approximately three division cycles on average. Data are summarised in Table 1. Effect of GM-CSF and imatinib on proliferation Culture Proliferation Index Statistics using Student t test, symbols denote comparisons between culture conditions Normal Control 1.07+/−0.06 * NS Normal+GM-CSF 1.11+/−0.03 * # NS CML Control 1.43+/−0.18 ** p>0.01 #, & NS CML+GM-CSF 2.30+/−0.40 ** CML+10μM imatinib 1.18+/−0.09 *** NS CML+10μM imatinib+GM-CSF 1.54+/−0.19 @, & CML+10μM imatinib+GM-CSF+E21R 1.17+/−0.09 *** @ p>0.01 GM-CSF induced strong proliferation in all CML, but not normal samples. Imatinib reduced proliferation of CML CD34+ cells at 1 microM and above, and the addition of GM-CSF reduced this proliferative effect at concentrations of imatinib up to 10 microM. The protective effect of GM-CSF was clearly blocked using E21R. When the proliferation of CML CD34+ cells cultured within the total mononuclear fraction was compared to purified CD34+ cells, it was found that there was an approximate 40% enhancement of PI at each concentration of imatinib. This enhanced proliferation was inhibited by the addition of E21R, indicating GM-CSF may be produced by non-CD 34+ cells. ELISpot assay confirmed the production of GM-CSF by non-CD34+ CML cells (34±23%), but not CD34+ cells (0.06±0.02%). Fluorescent inhibitor of apoptosis (FLICA) assay showed that GM-CSF could reduce the proportion of cells with activated caspases induced by imatinib by over 40% during a 3 day in vitro culture, promoting increased cell survival. As the GM-CSF receptor and BCR/ABL may share common signaling pathways, during blockade of BCR/ABL activity by imatinib, GM-CSF can compensate to maintain cell viability and proliferation. These findings have implications for optimizing imatinib therapy by manipulating cytokine signaling.