ULK1 inhibition as a targeted therapeutic strategy for FLT3-ITD-mutated acute myeloid leukemia

FMS-like tyrosine kinase 3 internal tandem duplication (FLT3-ITD)-mutated acute myeloid leukemia (AML) accounts for up to 30% of adult AML cases, and the prognosis of patients with FLT3-ITD-mutated AML is extremely dismal owing to the high relapse rate after chemotherapy and allogeneic stem cell transplantation [1]. Activating ITD mutations in FLT3 increases cell proliferation and survival, while blocking cellular differentiation through the constitutive activation of canonical pathways such as MAPK/ERK, PI3K/AKT, and STAT5; these mechanisms, together with other recurrent molecular abnormalities, are implicated in AML induction [2].

Several FLT3 tyrosine kinase inhibitors (TKIs) have been developed to target the aberrantly activated FLT3 receptor and to suppress constitutive tyrosine phosphorylation in FLT3-ITD AML [3, 4]. A recent phase III randomized study (RATIFY) demonstrated the survival benefit of a combination of chemotherapy with FLT3 TKIs, leading to the approval of the FLT3 inhibitor midostaurin by the US Food and Drug Administration [5]. However, therapeutic responses to the currently available FLT3 TKIs, if any, are short-lived and followed by early relapse in nearly all cases [4, 6, 7]; accordingly, the development of resistance to these TKIs impedes their therapeutic efficacy. Secondary mutations in the FLT3-TK domain have been demonstrated as one of the mechanisms underlying this resistance [6]. Multiple FLT3-TK domain mutations have been identified in therapy-resistant patients and cell lines [3, 6]. Therefore, the development of inhibitors to block each of these mutations would require a major effort [3, 7]. More recently, mutational analysis of samples from patients who had relapsed after FLT3-TKI treatment, as well as data from preclinical studies suggest that a cellular adaptive mechanism involving the activation of signaling pathways also plays a role in the FLT3-TKI resistance pathway [8], however, these pathways remain poorly elucidated. In addition, the inability of FLT3 TKIs to eliminate leukemia stem cells also contributes to treatment failure. Therefore, novel FLT3-ITD-targeted therapeutic strategies are necessary.

Autophagy is a cell-protective and degradative process that recycles damaged and long-lived cellular components [9]. Autophagy has opposing roles in cancers, depending on the context, oncogenesis, biological features, treatment, and progression [10‐12]. In certain cases, autophagy acts primarily as a tumor-suppressive mechanism by maintaining genomic integrity and preventing proliferation and inflammation [12]. Recent evidence suggests that autophagy contributes to cancer initiation and progression, as well as the development of resistance to treatment [10, 12‐14]. The dependence of the survival of numerous tumors on autophagy suggests that the inhibition of autophagy represents a therapeutic strategy against cancer that is broadly applicable across various tumor types [15‐17]. Autophagy may also have context-specific roles in different types of AML, depending on the clonal origin, oncogenic drivers, and state of leukemia expansion [18]. Autophagy is frequently found to be reduced in human AML blasts, and the loss of key autophagy genes has been shown to lead to leukemia initiation and progression in a mouse model. Disruption of autophagy and downregulation of autophagy-related genes are reportedly involved in the initiation and progression of AML. In contrast, recent studies have revealed that leukemic blasts and leukemia stem cells can utilize autophagy to respond to the specific energy demands during accelerated proliferation, and to counteract chemotherapeutic stress [19, 20]. These mechanisms may be employed to enhance the sensitivity of chemotherapeutic agents targeting leukemia cells. Furthermore, autophagy represents an attractive therapeutic target for various cancers, especially for autophagy-addicted tumors [15‐17, 21]. Accordingly, autophagy inactivation by pharmacological inhibition or knockout of essential autophagy-related genes sensitizes specific subsets of AML cells to chemotherapies and small-molecule inhibitors, reduces the proportion of functional leukemia-initiating cells, and extends patient survival [19]. To date, most clinical efforts have focused on the use of general autophagy inhibitors, including chloroquine, hydroxychloroquine, and bafilomycin A1, either as a single agent or in combination with other anti-cancer drugs [14]. However, the development of potent and selective autophagy inhibitors is difficult, in part because most core autophagic proteins used in small-molecule inhibitor screening techniques function far downstream in the pathway and are not readily druggable [22, 23]. A major challenge related to the use of autophagy-targeting drugs is their lack of potent and selective activity [23].

Among the ATG proteins that mediate all steps of autophagy flux, uncoordinated 51-like kinase 1 (ULK1) has serine/threonine protein kinase activity and is considered a key protein in the regulation of autophagy initiation [24, 25]. ULK1 forms a stable complex with ATG13, FIP200, and ATG101 to regulate the recruitment of the VPS34 complex, which consists of VPS34, beclin-1, and ATG14L [25]. In addition, ULK1 acts in later stages of the autophagy pathway, including autophagosome maturation. ULK1 undergoes complex post-translational modifications, including mammalian target of rapamycin complex 1 (mTORC1) and AMP-activated protein kinase (AMPK) [26], depending on the cellular conditions. ULK1 has been shown to promote cell survival in several cancers [27]. Overexpression of ULK1 is negatively correlated with prognosis in various cancers, including colorectal cancer, breast cancer, human nasopharyngeal carcinoma, and esophageal squamous cell carcinoma [28].

ULK1 is considered a promising target for autophagy inhibition because of its central role in pathway activation, druggable nature, and apparent selectivity for autophagy over other cellular functions [29]. Several small-molecule inhibitors of ULK1 have been recently reported. Although a second mammalian ATG1 ortholog, ULK2, also promotes autophagy, the loss of ULK1 alone is sufficient to abrogate autophagy in many cell types, underscoring its particularly important role. The crystal structures of ULK1 have provided insights into the druggable pockets of kinase [30], leading to the development of several inhibitors, including SBI-0206965, MRT 68921, and recently, ULK-101 [27]. Although they are still in the early stages of development, these agents provide new opportunities for the pharmacologic inhibition of autophagy that can potentially be translated to the clinic.

In this study, we evaluated the effects of the ULK1 inhibitors MRT 68921 and SBI-0206965 on apoptosis, and the potential of therapeutic strategies involving ULK1 targeting in FLT3-ITD AML cells. In particular, we aimed to elucidate the mechanisms underlying the effects of these ULK1 inhibitors on FLT3-ITD AML cell death.

Increasing evidence suggests that dysregulated autophagy is a hallmark of malignant disorders and plays complex roles in the tumorigenesis and progression of cancers and their resistance to treatment, in a context-dependent manner [10, 11, 31]. The diverse functions of numerous autophagy-related proteins in cancer have been extensively evaluated [31, 34]. Among them, ULK1 has drawn the most interest because it plays a key role in autophagy initiation as an only serine/threonine protein kinase [35, 36]. The findings that high levels of ULK1 are associated with poor prognosis in several solid tumors [28, 37] and that ULK1 inhibition can rescue the phenotype induced by caspase-3 deletion in AML1/ETO AML cells [38] suggest that ULK1 is a promising new therapeutic target for cancers.

FLT3-ITD mutations activate the survival pathway in an oncogene-addicted state. Given the critical role of autophagy in the cell biology of FLT3-ITD AML [39], ULK1 is likely to play an important role in this AML subtype. The higher levels of activated ULK1 and its downstream molecules in FLT3-ITD cells observed in our study suggest that ULK1 is indispensable for survival in FLT3-ITD-mutated cells and that its selective inhibition may potentially induce cell death. We demonstrated here for the first time that two ULK1 inhibitors, MRT 68921 and SBI-0206965, induced caspase-dependent apoptosis preferentially in FLT3-ITD-mutated leukemia cell lines and primary leukemic blasts obtained from FLT3-ITD AML patients, while sparing normal CD34 (+) cells.

MRT 68921-mediated apoptosis in FLT3-ITD cells was associated with dose-dependent increases in the cleavage of caspase-3 and -9, and PARP levels. Pretreatment of MV4–11 cells with the pan-caspase inhibitor z-VAD-FMK notably reduced the levels of cleaved caspases and PARP, indicating that ULK1 inhibitors induced apoptosis in a caspase-dependent manner. The interaction between ULK1 and caspase-3 was demonstrated in AML1/ETO leukemia cells [38]. However, the molecular interplay between ULK1 and activation of the caspase cascade in FLT3-ITD AML remains to be elucidated. We found that the ULK1 inhibitor increased apoptosis by MMP loss with FLT3 degradation in FLT3/ITDAML cells. Further research is needed to clarify the mechanism by which FLT3-ITD prevents mitochondrial apoptosis. Ju, H. Q et al. showed that FLT3-ITD causes an increase in aerobic glycolysis through AKT-mediated upregulation of mitochondrial hexokinase, which regulates the mitochondrial permeability transition pore [40]. Recently, studies in various FLT3ITD AML models provided evidence that FLT3 inhibition induces mitochondrial oxidative stress that determines the triggering of an apoptotic response [41, 42]. Our study suggests that the role of ULK1 in mitochondrial membrane potential and mechanisms by which the ULK1 inhibitor induces mitochondrial apoptosis in FLT3/ITD cells requires further analysis.

ULK1 inhibition also induced cell death, albeit to a lesser degree, in FLT3-WT leukemia cells, necessitating a clarification of whether the apoptosis-inducing effects of ULK1 inhibition were primarily due to ULK1 inhibition or off-target effects thereof. Both MRT 68921 and SBI-0206965 inhibited ULK1 activity, as demonstrated by the reduction in phosphorylation of ULK1 downstream molecules ATG13 (Ser318) [43], beclin-1 (Ser14) [25], and VPS34 (Ser249) [44] as well as ULK1 itself (T180) [24]. Phosphorylation of ATG13 at serine 318 is directly correlated with ULK1 activity [45]. We observed that the extent of apoptosis was highly correlated with the reduction in ATG13 phosphorylation in FLT3-ITD cells.

Although MRT68921 and SBI026965 were developed to inhibit autophagy [29, 46], the action of these ULK1 inhibitors remains unclear. Particularly, the action spectrum of these ULK1 inhibitors has not been fully evaluated in cancer cells, including leukemia blasts, which are considered to have abnormal autophagy regulatory mechanism in a context-dependent manner. In addition, most published data described the autophagy inhibitory effects of the ULK1 inhibitors in the settings of induced autophagy, not in the basal conditions. Interestingly, both MRT 68921 and SBI-0206965 activated LC3-II conversion and GFP-LC3 puncta and autophagosome formation, which are indicators of autophagy induction. However, these results did not confirm that autophagic flux was increased. The increase in p62 level after ULK1 inhibitor treatment indicated that autophagic flux was blocked by the ULK1 inhibitor. These results suggest that the ULK1 inhibitors induced autophagic flux blockage in parallel with autophagy activation. Gross autophagy revealed offset results for both actions. Although the ULK1 inhibitor suppressed p62 degradation, the accumulation of LC3 II, autophagosomes, and LC3 puncta were lower in MV4;11 cells than in U937 cells, indicating that autophagic flux is greater in MV4;11 cells. We demonstrated that this activated autophagy in FLT3/ITD AML cells played a prosurvival role by enhancing ULK1 inhibitor-induced apoptosis using the autophagy inhibitor. Reagents conventionally used for autophagy inhibition have different mechanisms and action targets. 3-MA inhibits autophagy by blocking autophagosome formation, whereas bafilomycin A1 and HCQ inhibit the late phase of autophagy. Since the ULK1 inhibitor activated autophagosome formation, which is the early phase of autophagy, the apoptosis-inducing effects of an addition of 3-MA, rather than BafA1 or HCQ, on ULK1 inhibitors were investigated.

The mechanisms by which ULK1 inhibiors result in the activation of prosurvival autophagy in FLT3-ITD cells remain to be investigated. Under stress conditions, AMPK can inactivate mTORC1 by phosphorylating the mTOR upstream regulator TSC2 at the Thr1227 and Ser1354 residues, leading to autophagy activation [47]. The feedback regulatory mechanism between AMPK and ULK1 remains to be further elucidated. Cheong H et al. showed that ammonia production due to amino acid catabolism is responsible for ULK1-independent autophagy [48]. A recent study suggested that the requirement for ULK1 in autophagy induction can be bypassed by an increase in VPS34 kinase activity, caused by inhibition of a newly identified and negative regulatory VPS34 acetylation event [49]. Additional analysis of ULK1-independent pathways operating in autophagy may support our findings.

The direct molecular interactions between ULK1 and the FLT3-ITD-mutated protein are unknown. In this study, treatment of MV4–11 cells with ULK1 inhibitors led to decreases in the levels of phospho-FLT3, as well as p-ERK and p-STAT5, which are associated with FLT3 activation. These findings indicate that ULK1 inhibition led to the suppression of FLT3 signaling. It has been shown that autophagy induces the degradation of oncogenic fusion proteins in AML, leading to cell differentiation and death of leukemic blasts [50, 51]. Furthermore, the FLT3-ITD-mutated protein is post-translationally regulated and degraded by autophagy [39]. However, in this study, we found that FLT3-ITD molecules and autophagosomes are located at different sites in MV4;11 cells treated with ULK1 inhibitors. This indicates that ULK1 inhibition-induced autophagy may not be the mechanism underlying FLT3-ITD degradation. The addition of the autophagy inhibitor 3-MA to MRT 68921 led to decreases in the levels of p-FLT3; therefore, autophagy was speculated that play a protective role against FLT3 degradation in FLT3-ITD AML cells. In addition, proteasome inhibition partially attenuated MRT 68921-induced apoptosis, suggesting that proteasomal degradation of FLT3-ITD contributed to the ULK1 inhibition-mediated reduction in the levels of activated FLT3 and its downstream molecules.

The UPR, which involves three major transducers, PERK, IRE1, and ATF6, occurs in response to intracellular and extracellular events that perturb protein folding in the endoplasmic reticulum (ER) [52]. Autophagy and UPR pathways are interconnected and regulate cellular responses to apoptotic triggers such as environmental and genetic stresses in cancer cells [12, 52]. ER stress (ERS) is thought to activate autophagy via UPR-mediated transcriptional upregulation of the autophagic machinery [53]. We evaluated changes in the activation of the PERK/eIF2α/ATF4 pathway, which is one of the essential branches of the UPR, after treatment of MV4–11 cells with the ULK1 inhibitors. PERK, eIF2α, and ATF4 phosphorylation decreased upon treatment with MRT 68921, indicating that the ERS pathway is suppressed following ULK1 inhibition. However, when AML cells were treated with MRT 68921 plus PERK activator or inhibitor, there were no significant changes in apoptosis This indicates that ERS does not play a critical role in the ULK1 inhibition-mediated regulation of apoptosis and autophagy in FLT3-ITD cells.

To our knowledge, the present study is the first to demonstrate that ULK1 inhibitors effectively induce mitochondria-mediated, caspase-dependent apoptosis in FLT3-ITD-mutated leukemia cell lines and primary leukemia cells, but not in normal CD34 (+) cells. Paradoxically, ULK1 inhibitors also induced prosurvival autophagy, as evidenced by the synergistic enhancement of apoptosis following the addition of autophagy inhibitors. Further studies are necessary to determine whether ULK1 inhibition is an effective strategy for overcoming FLT3-TKI resistance. Owing to its central role in autophagy regulation, druggable nature, and apparent selectivity over other autophagy inhibitors, along with minimal effect on normal CD34 (+) cells, ULK1 represents a promising molecular target for FLT3-ITD AML treatment.