Autophagy, which literally means “eat yourself”, is more than just a lysosomal degradation pathway. It is a well-known regulator of cellular metabolism and a mechanism implicated in tumor initiation/progression and therapeutic resistance in many cancers. However, whether autophagy acts as a tumor suppressor or promoter is still a matter of debate. In acute myeloid leukemia (AML), it is now proven that autophagy supports cell proliferation in vitro and leukemic progression in vivo. Mitophagy, the specific degradation of mitochondria through autophagy, was recently shown to be required for leukemic stem cell functions and survival, highlighting the prominent role of this selective autophagy in leukemia initiation and progression. Moreover, autophagy in AML sustains fatty acid oxidation through lipophagy to support mitochondrial oxidative phosphorylation (OxPHOS), a hallmark of chemotherapy-resistant cells. Nevertheless, in the context of therapy, in AML, as well as in other cancers, autophagy could be either cytoprotective or cytotoxic, depending on the drugs used. This review summarizes the recent findings that mechanistically show how autophagy favors leukemic transformation of normal hematopoietic stem cells, as well as AML progression and also recapitulates its ambivalent role in resistance to chemotherapies and targeted therapies.

The Janus kinase 2 (JAK2)-driven myeloproliferative neoplasms (MPNs) are chronic malignancies associated with high risk of complications and suboptimal responses to JAK inhibitors such as ruxolitinib. Therefore, better understanding cellular changes induced by ruxolitinib is required to develop new combinatory therapies to improve treatment efficacy. Here, we demonstrated that ruxolitinib induced autophagy in JAK2V617F cell lines and primary MPN patient cells through the activation of protein phosphatase 2A (PP2A). Inhibition of autophagy or PP2A activity along with ruxolitinib treatment reduced proliferation and increased death of JAK2V617F cells. Accordingly, proliferation and clonogenic potential of JAK2V617F-driven primary MPN patient cells, but not of normal hematopoietic cells, were markedly impaired by ruxolitinib treatment with autophagy inhibitor or PP2A inhibitor. Finally, preventing ruxolitinib-induced autophagy with a novel potent autophagy inhibitor Lys05 improved leukemia burden reduction and significantly prolonged mice overall survival compared with ruxolitinib alone. This study demonstrated that PP2A-dependent autophagy mediated by JAK2 activity inhibition contributed to the resistance to ruxolitinib. Altogether, our data support that targeting autophagy or its newly identified regulator PP2A could enhance sensitivity to ruxolitinib of JAK2V617F MPN cells and improve MPN patient care.