Introduction: Acute myeloid leukemia is an aggressive malignancy characterized by relapse and resistance to therapy. When treated with standard chemotherapy 60-80% of patients will achieve a remission however most will suffer a relapsed and once relapsed the disease is much more resistant to repeated treatment. Despite the central role of resistance little is known about the mechanism of chemotherapy responsiveness in AML. Our laboratory has previous shown a metabolic shift occurs in AML cells following chemotherapy exposure with increased mitochondrial oxygen consumption leading to resistance. Adenosine monophosphate activated protein kinase (AMPK) is a central metabolic regulator and leading candidate to coordinate this metabolic shift. However, the role of AMPK in chemotherapy response in AML is unknown.

Methods: A Cas9 expressing AML model was generated and partially infected with a lentiviral vector coding an sgRNA targeting the alpha 1 subunit of AMPK, PRKAA1. This subunit was chosen as cell line and clinical expression datasets demonstrated that AML cells expressed little or none of the second alpha subunit PRKAA2. PRKAA1 deleted clones (AMPK KO) were isolated and loss of PRKAA1 expression and AMPK activity were confirmed. AMPK KO cells were characterized for responsiveness to chemotherapy, metabolic inhibitors and the DNA damage response.

Results: Consistent with a role in response to chemotherapy AML cells treated with cytarabine demonstrated a significant increase in AMPK activation. When partially AMPK deleted populations were treated with cytarabine or doxorubicin a significant enrichment was seen in the deleted cells suggesting that loss of AMPK promotes therapy resistance. To confirm this finding isogeneic AML cell lines were treated with chemotherapy. AMPK KO cells were significantly more resistant to doxorubicin and cytarabine compared to isogeneic controls expressing an sgRNA targeting the ROSA26 (ROSA KO) safe harbor locus. When the mechanism of this resistance was explored AMPK KO cells were found to have diminished H2AX, p53 and p21 induction following exposure to doxorubicin culminating in a decrease in apoptosis. To assess if this resistant phenotype was present in vivo, AMPK KO or ROSA KO cells were tail vein injected into sub-lethally irradiated mice. Upon confirmation of engraftment mice were treated with a combination of cytarabine and doxorubicin. AMPK KO mice had a significantly decreased survival following treatment then treated ROSA controls. Additionally, AMPK KO and ROSA KO cells were treated with the glycolysis inhibitor 2-Deoxyglucose, the autophagy inhibitor chloroquine and the fatty acid oxidation inhibitor extomoxir. In all cases AMPK KO cells were significantly more resistant to these agents than the ROSA KO cells suggesting a proapoptotic role of AMPK following exposure to metabolic inhibitors. Finally, the clinical relevance of these findings was assessed by examining outcomes of AML patients treated with chemotherapy in relationship to their levels of PRKAA1 expression. Patients with the lowest expression of PRKAA1 had a significantly worse overall survival consistent with our findings.

Conclusions: In AML cells activated AMPK provides pro-apoptotic signaling in response to DNA damaging agents allowing incorporation of metabolic status into cell fate decisions following treatment. Utilization of metabolic inhibitors that activate AMPK in combination with chemotherapy should be explored in the treatment of AML.

Disclosures

Pardee:Karyopharm: Research Funding; Novartis: Speakers Bureau; Rafael Pharmaceuticals: Employment; Celgene: Speakers Bureau; Amgen: Speakers Bureau.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution