Effects of G-CSF On Acute Lymphoblastic Leukemia.

Abstract 2564

G-CSF is commonly used to treat chemotherapy-induced neutropenia and for the mobilization of hematopoietic stem cells for transplantation in patients with leukemia. Administration of G-CSF has profound effects on the bone marrow microenvironment including the cleavage of molecules required for the maintenance of lymphopoiesis, including CXCL12 and VLA-4. We have recently reported that G-CSF results in the dramatic suppression of B-lymphopoiesis. This, together with previous reports by ourselves, and others, showing that disruption of CXCL12 or VLA-4 slow the progression of B-lineage ALL lead us to consider that G-CSF may similarly antagonize the progression of ALL.

To explore this possibility, we examined the impact of G-CSF administration on six human ALL xenografts using a NOD/SCID mouse model. Mice were engrafted without radiation and G-CSF commenced when 1% of the bone marrow consisted of ALL cells. G-CSF was administered twice daily for 10 days, at which time all animals were culled and leukemia assessed in the blood, bone marrow and spleens. Surprisingly G-CSF was found to increase disease progression in two of xenografts investigated (1345 and 0398, referred to as G-CSF responsive xenografts hereafter), while the remainder demonstrated a small reduction in leukemia, with one showing a statistical significant decrease. No evidence for a direct mitogenic effect of G-CSF could be demonstrated in any of the xenografts using exogenous G-CSF in vitro cultures in the presence or absence of human or murine stromal support. Consistent with these findings, and previous reports, little to no G-CSF receptor was detected by flow cytometry or microarray analysis of xenografts.

Microarray analysis of the xenografts revealed significant differences in gene expression between the G-CSF responsive xenografts and the remainder of the samples. A total of 83 genes were expressed at a higher level and 127 genes at a lower level in the G-CSF responsive xenografts. The more highly expressed genes included cell cycle regulators (eg cyclin A1), adhesion molecules (eg ALCAM), extracellular matrix components and surface receptors. Perhaps the most interesting was the exclusive expression of the acetylcholine receptor (cholinergic receptor, nicotinic, beta 4, nAChRb4) in the G-CSF responsive cases. Analysis of a large public dataset of childhood ALL samples revealed significantly higher expression of this gene in ALL samples with rearranged MLL (p<0.03). However, small numbers of cases in all ALL subgroups had greater than an 2 fold higher expression compared to normal B cell progenitors. The role of nAChR in the response of ALL cells to micro-environmental changes induced by G-CSF remains to be determined, however, nAChR has known roles in cell proliferation and inhibition of apoptosis. Furthermore G-CSF is known to induce acetylcholine production in other tissues.

In summary, G-CSF inhibited leukemia progression in the majority of patient xenografts, however, in a subset of samples G-CSF accelerated disease progression. Clinically, G-CSF administration to ALL patients has not been associated with any major adverse outcomes. However our data suggest that a small subset of patients may experience accelerated disease. Identification of features associated with adverse responses to G-CSF will permit the identification of patients for whom G-CSF may present a risk for increased disease progression.