Acceleration of AML Progression By Cigarette Smoke Exposure or Condensate Exposure and Associated DNA Methylation Alterations

Subsets of acute myelogenous leukemia (AML) are characterized by molecular alterations with prognostic significance, however, little is known about how modifiable behaviors, such as cigarette smoking, intersect with genetic factors. Mutations rendering the Fms-like tyrosine kinase 3 (FLT3) constitutively active, such as internal tandem duplication (ITD), are associated with refractory disease and therapy resistance. Inhibition of the FLT3 kinase shows some benefit in this population, as highlighted by the FDA approvals of midostaurin and gilteritinib, but overall outcomes remain poor. Cigarette smoke exposure (CSE) also marks a population of patients with poor prognosis. Current and former smokers who develop AML are known to have worse survival as compared to never smokers (Alfayez M., et al. ASCO 2019), but the impact of FLT3 mutation and subsequent associated treatment response has not been studied. Also, the underlying mechanism of how history of cigarette smoking influences leukemia biology and response to therapy is poorly understood.

In order to model a history of smoking in AML patients, NOD-SCID mice (n=25) were exposed to CSE using a smoking robot for 2 hours, 5 days/week, for 2 weeks. Mice were then inoculated via tail-vein injection with luciferase tagged human FLT3-ITD cells and leukemia burden was monitored through noninvasive imaging. CSE continued through the duration of the experiment, post engraftment. Control "non-smoking" mice (n=15) were only injected with leukemia cells. Within one week post leukemic introduction, a significant increase in leukemic burden as measured by bioluminescence was apparent in mice exposed to CSE versus control mice (P-value<0.0001). To model the impact of smoking cessation upon diagnosis of AML in patients, experiments were modified to halt CSE once leukemic engraftment was detectable by non-invasive imaging. Smoking cessation versus continuous smoke exposure yielded reduced relative leukemic burden. Mice with continuous smoke exposure had higher rates of leukemia compared to mice who ceased smoking (n=10) one week prior (P-value =0.0064). These rapid changes in leukemic burden suggest that CSE may prime the microenvironment to promote leukemia progression or directly affect leukemia cells. To address the latter possibility, human AML cells were exposed to cigarette smoke condensate (CSC), which contains the chemicals present in cigarettes, for two weeks before introducing the cells into mice. A significant increase in leukemic burden was observed in mice injected cells exposed to CSC compared to mice injected with unexposed leukemia cells (P-value <0.001), highlighting a direct role for the chemicals in cigarettes on in vivo leukemia proliferative factors.

Smokers are known to carry altered global DNA methylation signatures that persist decades after quitting. To measure DNA methylation changes in the in vivo models described above, we examined spleens of non-smoking and smoke exposed mice by reduced representative bisulfite sequencing (RRBS). Sequences were mapped to either the human or mouse genome, (enabling identification of leukemia specific versus microenvironment specific alterations) and were compared in the smoking and non-smoking mice. Over 200 genes exhibited significant DNA methylation alterations in their promoter regions. Genes involved in RNA polymerase activity and chromatin remodeling were highly represented amongst those with altered DNA methylation.

The clinical significance of our observations was confirmed in a cohort of 58 treatment naïve FLT3-ITD AML patients at MD Anderson receiving intensive induction therapy: 41 never smokers and 17 ever smokers. Smokers had significantly reduced survival as compared to the never smokers (median overall survival of 18 vs 23 months, P-value 0.0092).

Collectively our findings indicate that short-term CSE is sufficient to alter DNA methylation patterns and accelerate the early progression of FLT3-ITD AML in vivo. Smoking cessation upon diagnosis may slow leukemic growth relative to smoking throughout AML therapy prompting the consideration of behavioral interventions for smokers with AML. Improved understanding of the mechanism of leukemic progression and drug resistance from CSE is expected to lead to improved treatment paradigms designed for patients with a history of cigarette smoking.