Multiparameter FACS Analysis of G-CSF and IL-6 Signaling through Stat3 and Stat5 In Primary Pediatric AML Samples.

Abstract 1051

Pediatric acute myeloid leukemia (AML) remains a devastating disease with a relapse rate approaching 50%. New targeted therapies, such as kinase inhibitors, are entering clinical use rapidly, yet the signaling abnormalities targeted by these drugs remain poorly understood. Pathways involving Stat3 and Stat5 are known to be aberrantly active in AML. Increased constitutive and/or ligand-activated Stat3/5 activity has been associated with chemoresistant disease, likely due to increased pro-survival gene expression. Our hypothesis is that analysis of Stat3/5 activation, and expression levels of key proteins in the pathways, will provide insight into mechanisms of aberrant signaling, and will guide the development and implementation of targeted therapies. We performed multiparameter FACS analysis of tyrosine-phosphorylated Stat3 (pStat3; n=56) and pStat5 (n=38) in pediatric AML samples from the Children's Oncology Group. We also measured expression levels of Stat3 protein, G-CSF receptor (G-CSFR), and gp130 (IL-6 receptor; n=56). Constitutive activation was common and varied over a wide range (median 36.1% pStat3+, range: 0.7 – 82%; median 33.9% pStat5+, range: 5.9 – 77%). There was non-significant correlation between the percent of pStat3+ cells and the percent of pStat5+ cells in unstimulated cells. As expected, there was a significant correlation between constitutive %pStat3+ and total Stat3 expression (by mean fluorescence intensity (MFI)), with linear correlation coefficient R=0.375 (p=0.004). Most samples were not responsive to G-CSF or IL-6. For Stat3 activation, 33/56 (59%) failed to demonstrate at least a 2-fold increase in the pStat3 MFI after 15 minute treatment with G-CSF (100 ng/ml), and 35/56 (63%) did not respond to IL-6 (50 ng/ml IL-6 + 100 ng/ml soluble IL-6 receptor (sIL6R)). A pStat3 response to these high doses only was seen in 23% of samples after G-CSF and 20% after IL-6. A pStat3 response was seen to 10 ng/ml G-CSF in 18% and to 5 ng/ml IL-6 + 10 ng/ml sIL6R in 16%. No samples responded to 1 ng/ml G-CSF and 1 sample (2%) responded to 0.5 ng/ml IL-6 + 1 ng/ml sIL6R. The pattern of pStat5 responses to G-CSF was similar: 53% did not respond to any dose, 26% responded to the highest dose, 21% responded to the intermediate dose, and none responded to the lowest dose. Only 1 sample (3%) showed an increase in pStat5 after IL-6 stimulation. As in unstimulated cells, there was no correlation between the pStat3 and pStat5 responses to G-CSF, suggesting divergence of the pathways from G-CSFR to each Stat. In contrast, the pStat3 response to G-CSF significantly correlated with the pStat3 response to IL-6 (R=0.423, p=0.001), suggesting convergence of the pathways from each receptor to Stat3. Surface G-CSFR expression positively correlated with the GCSF-induced pStat3 response (R=0.365, p=0.006). Interestingly, surface gp130 expression negatively correlated with the IL6-induced response (R=-0.352, p=0.008), meaning samples with robust IL-6 responses tended to have fewer gp130+ cells. In summary, multiparameter FACS analysis of baseline and ligand-induced Stat3 and Stat5 activity provides biologic insight into signaling aberrancies in AML, and this knowledge will promote rational design and testing of new therapies.