Metformin, one of the most widely prescribed anti-diabetic drugs, has recently received significant attention as an antineoplastic agent due to its cytotoxic effects in a variety of solid tumor cancer cell lines including prostate, colorectal, lung, pancreatic, and breast [12
]. Its anti-cancer effect is supported by epidemiological studies that show a decrease in cancer incidence in metformin-treated patients [14
]. Although its mechanism of action in hepatic glucose metabolism is well documented, little is known about metformin’s mechanism of action in cancer cells, particularly in leukemia. It has also been proposed that most of metformin’s cytotoxic effects are mediated through activation of the AMPK pathway, the main regulator of cellular energy homeostasis [3
]. Here, we show for the first time that metformin induced cell growth arrest and apoptosis in ALL cell models and primary cells occurs through induction of the ER stress/UPR-mediated cell death pathway, and that this effect is AMPK-dependent. In this context through AMPK’s effect as a negative regulator of the UPR, metformin prevents ALL cells from effectively engaging the UPR to overcome ER and proteotoxic stress-induced irreversible cellular damage leading to apoptotic death.
The notion that metformin is capable of inhibiting the UPR was first reported by Saito et al. [39
], who found that antidiabetic biguanides could inhibit GRP78 activity during glucose deprivation. Our data not only confirm that metformin down-regulates GRP78 but also demonstrate that metformin induces stress in the ER lumen evidenced by activation of ATF6, IRE1α, and CHOP in metformin-treated ALL cells. It has been reported that metformin reduces the ATP/AMP ratio by targeting complex 1 of the respiratory chain [11
]. Therefore, metformin-induced ER stress is likely triggered by accumulation of unfolded/misfolded proteins in the ER as a consequence of ATP depletion in a manner akin to glucose deprivation [48
]. In various normal cell types including cardiomyocytes [17
] and bovine aortic endothelial cells [18
], the activation of AMPK by AICAR or metformin were found to be tissue protective via AMPK-dependent suppression of the UPR. Others have reported that the tissue protective effects of metformin in renal tubular epithelial cells were AMPK-independent [38
]. In contrast, we previously demonstrated that AMPK activation by AICAR, methotrexate, or 2-DG led to inhibition of the UPR and cytotoxicity in ALL cells [15
]. We now report that metformin’s induction of apoptosis in ALL cells is AMPK-dependent and occurs via a UPR-mediated mechanism. Therefore, tissue specificity appears to exist in the mechanism by which certain AMPK activators suppress the UPR and in addition, whereas AMPK suppression of the UPR may be beneficial to normal tissues, it induces cell death in several cancer phenotypes.
Herein, we hypothesized that concomitant persistence of metformin-induced ER stress via ATP depletion and inhibition of GRP78 following treatment with metformin leads to ALL cell death demonstrated by the increased expression of the UPR-mediated apoptotic factors IRE1α, and CHOP [25
]. It has been shown that down-regulation of GRP78 expression is sufficient to induce apoptosis and plays a critical role in physiologic and pathologic stress coping mechanisms used by a variety of cell types [24
]. This report is the first to elucidate the role of UPR in metformin-induced cell death in ALL and in that context it identifies AMPK as an essential regulator of this mechanism. Indeed, down-regulation of AMPK using shRNA completely abrogated metformin-induced apoptosis in ALL, and more important correlated with increased GRP78 expression and down-regulation of the UPR-mediated apoptotic factors IRE1α and CHOP. Mechanistically, this report provides further evidence and confirms our previous findings showing that AMPK acts as a physiological suppressor of the UPR whereas Akt up-regulates the UPR in ALL cells [15
]. Therefore, our recent body of work demonstrates that in ALL cells undergoing metabolic and energetic stress, the UPR is regulated via crosstalk between AMPK and Akt, and these interactions determine the fate (i.e. death vs.
survival) of ALL cells under conditions of metabolic and/or proteotoxic stress.
In our ALL cell models, the mechanisms responsible for the regulation of protein synthesis also appear altered following exposure to metformin. Our data showed that down-regulation of AMPK led to increased phosphorylation of p-eIF2α and dephosphorylation of p-4EBP1, both events being associated with inhibition of protein synthesis. Therefore, shRNA knockdown of AMPK rescued ALL cells through inhibition of protein synthesis coupled with up-regulation of the UPR (increased GRP78 expression). In support for this mechanism is the rescue we observed following co-treatment with the mTOR inhibitor rapamycin, a known inhibitor of CAP-dependent protein synthesis, which resulted in decreased IRE1α and CHOP expression. The relationship between protein translation and UPR activation was recently highlighted by Matsuo et al. [49
], which showed that hyperactivation of 4EBP1 prevented UPR activation. Our findings do contrast with the recent report by Grimaldi et al. [50
] proposing that metformin induces cell death by blocking the catalytic activity of mTORC1 and repressing mRNA translation, although these investigators did not assess the role of the UPR in their model. Based on our data, we propose that it is the balance between proteotoxic stress and UPR activity that results in either death or survival from metformin’s cytotoxicity in ALL cells.
To our knowledge, the induction of PIM kinase activity by metformin had not been reported. Herein, we interpreted that metformin induced PIM-2 kinase expression is a compensatory survival mechanism due to its anti-apoptotic role via down-regulation of BAD [32
]. Further, it has been demonstrated that PIM kinases can negatively regulate AMPK activity [31
] and promote hematopoietic cell growth and survival [45
]. The synergistic effects observed between metformin and PIM kinase inhibition support our postulate that the increase in PIM-2 expression represents a buffering response to metformin. More important, this combination led to synergistic induction of cell death. Based on our findings, we propose a new role for PIM-2 kinase as capable of modulating UPR via down-regulation of AMPK activity. We also found that targeting Akt, another survival kinase capable of modulating AMPK activity through phosphorylation of Ser485 [51
], resulted in significant synergistic cell death when combined with metformin. Our data confirm our report demonstrating antagonistic roles for Akt and AMPK in modulating the UPR [19
]. Based on data presented herein, we propose a model for metformin-induced cell death in ALL cells via ER stress/UPR-mediated apoptosis in which metformin, by decreasing the ATP/AMP ratio promotes i) ATP depletion and unfolded protein accumulation in the ER lumen resulting in proteotoxic and ER stress, and ii) AMPK activation which leads to suppression of the UPR (down-regulation of GRP78), leading to UPR-mediated apoptotic cell death (). In support of this proposed model we present evidence that down-regulation of AMPK rescues ALL cells from metformin-induced apoptosis by interrupting protein synthesis and increasing UPR activity, which would reduce the load of unfolded proteins in the ER and restore the ability of the cells to cope with the ER/proteotoxic stress.
Proposed mechanism of action for metformin in ALL cells.
In summary, our studies demonstrate that the crosstalk between AMPK, Akt, PIM-2 kinase, and UPR signaling pathways determines metformin-induced cell death in ALL. We demonstrate for the first time that metformin induced apoptosis in ALL lymphoblasts occurs via UPR-mediated mechanisms which are entirely AMPK-dependent. We also provide further evidence for the roles of AMPK and Akt as regulators of the UPR in ALL. Finally, our data not only demonstrate the ability of metformin to induce significant cell death in ALL cell lines and primary cells supporting future translation into clinical trials, but also uncover strategies exploiting synthetic lethality by combining metformin and selective inhibitors of these pathways that may also be suitable for clinical translation in patients with ALL.