A shared feature of many bone marrow failure syndromes is their propensity to develop myelodysplasia (MDS) or acute myeloid leukemia (AML). The molecular mechanisms that underlie this susceptibility are largely unknown. Severe congenital neutropenia (SCN) is an inherited disorder of granulopoiesis that is associated with a marked increased risk of developing MDS/AML. Somatic mutations of CSF3R, encoding the G-CSF receptor (G-CSFR), that truncate the carboxy-terminal tail are associated with the development of MDS/AML in SCN. Transgenic mice carrying a ‘knock-in’ mutation of their Csf3r (termed d715 G-CSFR) reproducing a mutation found in a patient with SCN have normal basal granulopoiesis but an exaggerated neutrophil response to G-CSF treatment. We previously reported that the d715 G-CSFR is able to cooperate with the PML-RARƒÑ oncogene to induce AML in mice. Herein, we summarize data supporting the hypothesis that alterations in the bone marrow microenvironment induced by G-CSF contribute to oxidative DNA damage in hematopoietic stem/progenitors cells (HSPCs) and possibly leukemic transformation.
We previously showed that G-CSF treatment is associated with a marked loss of osteoblasts in the bone marrow, thereby potentially disrupting the osteoblast stem cell niche (Semerad, Blood 2005). Of note, patients with SCN chronically treated with G-CSF are prone to develop osteopenia, suggesting that osteoblast suppression by G-CSF also may occur in humans. We first asked whether the d715 G-CSFR was able to mediate this response. Wild-type or d715 G-CSFR were treated with G-CSF for 1–7 days and osteoblast activity in the bone marrow measured by expression of CXCL12 and osteocalcin. Consistent with previous reports, a decrease in osteocalcin and CXCL12 was not apparent until after 3 days of G-CSF treatment and reached a maximum after 7 days. Surprisingly, the magnitude of osteoblast suppression was greater in d715 G-CSFR compared with wild-type mice. The fold-decrease in osteocalcin mRNA from baseline in wild-type mice was 147 ± 70.1 versus 1,513 ± 1091 in d715 G-CSFR mice (p < 0.001). Likewise, a greater fold-decrease in CXCL12 mRNA was observed. We next assessed oxidative stress in c-KIT+ Sca+ lineage− (KSL) progenitors after G-CSF treatment. In both wild-type and d715 G-CSFR KSL cells no increase in reactive oxygen species (ROS) was observed at baseline or 12 hours after a single dose of G-CSF. However, after 7 days of G-CSF, a significant increase (3.4 ± 0.1 fold; p = 0.009) in ROS was observed in d715 G-CSFR but not wild-type KSL cells. To determine whether oxidative stress contributed to DNA damage, histone H2AX phosphorylation (pH2AX) was measured by flow cytometry. No increase in pH2AX was observed after short-term (less than 24 hour) G-CSF treatment. However, a modest but significant (1.9 ± 0.1 fold; p = 0.0007) increase in pH2AX was observed in d715 G-CSFR but not wild-type KSL cells after 7 days of G-CSF. To determine whether increased oxidative stress was casually linked to DNA damage, we co-administered the antioxidant N-acetyl cysteine (NAC) during G-CSF treatment. As expected, induction of ROS in KSL cells was markedly suppressed by NAC administration. Importantly, the increase in pH2AX levels in d715 G-CSFR KSL cells induced by G-CSF was completely blocked by NAC administration. Finally, to determine whether alterations in the bone marrow microenvironment, specifically decreased CXCL12 expression, contributed to DNA damage, we treated mice with AMD3100, a specific antagonist of CXCR4 (the major receptor for CXCL12). Treatment of wild-type or d715 G-CSFR mice with a single dose of G-CSF (3 hour time point) or with AMD3100 alone did not induce H2AXp. However, co-administration of AMD3100 with a single dose of G-CSF induced modest but significant H2AXp in d715 G-CSFR KSL cells (5.74 ± 1.06 fold; P<0.001). Collectively, these data suggest a model in which alterations in the bone marrow microenvironment induced by G-CSF may contribute to genetic instability in HSPCs and ultimately leukemic transformation. The mutant CSF3R may contribute to leukemogenesis through both increased ROS production in HSPCs and increased suppression of osteoblasts.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.