JKST6, a novel multikinase modulator of the BCR-ABL1/STAT5 signaling pathway that potentiates direct BCR-ABL1 inhibition and overcomes imatinib resistance in chronic myelogenous leukemia

Chronic myelogenous leukemia (CML) is a hematological malignancy that highly depends on the BCR-ABL1/ STAT5 signaling pathway for cell survival. First-line treatments for CML consist of tyrosine kinase inhibitors that efficiently target BCR-ABL1 activity. However, drug resistance and intolerance are still therapeutic limitations in Ph+ cells. Therefore, the development of new anti-CML drugs that exhibit alternative mechanisms to overcome these limitations is a desirable goal. In this work, the antitumoral activity of JKST6, a naphthoquinonepyrone hybrid, was assessed in imatinib-sensitive and imatinib-resistant human CML cells. Live-cell imaging analysis revealed JKST6 potent antiproliferative activity in 2D and 3D CML cultures. JKST6 provoked cell increase in the subG1 phase along with a reduction in the G0/G1 phase and altered the expression of key proteins involved in the control of mitosis and DNA damage. Rapid increases in Annexin V staining and activation/ cleavage of caspases 8, 9 and 3 were observed after JKST6 treatment in CML cells. Of interest, JKST6 inhibited BCR-ABL1/STAT5 signaling through oncokinase downregulation that was preceded by rapid polyubiquitination. In addition, JKST6 caused a transient increase in JNK and AKT phosphorylation, whereas the phosphorylation of P38-MAPK and Src was reduced. Combinatory treatment unveiled synergistic effects between imatinib and JKST6. Notably, JKST6 maintained its antitumor efficacy in BCR-ABL1-T315I-positive cells and CML cells that overexpress BCR-ABL and even restored imatinib efficacy after a short exposure time. These findings, together with the observed low toxicity of JKST6, reveal a novel multikinase modulator that might overcome the limitations of BCR-ABL1 inhibitors in CML therapy.


Introduction
Chronic myelogenous leukemia (CML) is a hematopoietic stem cell disease associated with abnormal proliferation of clonal cells of the myelogenous lineage [1]. This malignant myeloproliferative disorder is characterized by a chromosomal translocation t(9;22)(q34;q11) that provokes Philadelphia (Ph) chromosome emergence, a hallmark present in 95% of CML cases but also at lower frequencies in acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) [2]. The resulting fused genes of the Ph chromosome encode the BCR-ABL1 oncogenic protein, whose constitutively active tyrosine kinase represents the keystone of CML pathogenesis [3]. BCR-ABL1 activates different intracellular signaling pathways, including RAS/RAF/MAPK, PI3K/AKT and JAK/STAT, which are involved in the control of cell survival and proliferation, among other key cellular responses [4]. Dysregulation of JAK/STAT signaling is indeed a common factor in hematological cancers, particularly in CML, where BCR-ABL1 can directly and indirectly activate STAT5, resulting in the overexpression of proto-oncogenic target genes (e.g., MYC and PIM1) [4,5]. Thus, BCR-ABL1/STAT5 signaling triggers enhanced proliferation of CML clones and genomic instability, sustaining the conditions for a tumoral environment [6,7]. Given the above findings, it is not surprising that targeting BCR-ABL1 has marked a milestone in the treatment of CML.
In 2001, the Food and Drug Administration (FDA) approved the first tyrosine kinase inhibitor (TKI), imatinib mesylate (IM), which binds the BCR-ABL1 catalytic domain, blocking its activation [8]. Twenty years later, a broader spectrum of TKIs is available as first-line treatment for CML [9], allowing most early diagnosed CML patients to reach a near-normal life expectancy. However, TKIs exhibit key limitations in CML therapy, including some serious side effects mainly associated with second-generation TKIs, the emergence of TKI resistance, and the requirement for long-term medication, which triggers high economic costs [9,10]. Approximately 30-50% of CML patients discontinue IM therapy after 5 years, mainly due to the emergence of treatment resistance (15-20%) and intolerance (5-7%) [11]. Hence, mechanisms involved in the generation of IM resistance have been extensively investigated. Interestingly, some are defined as BCR-ABL1-dependent mechanisms (e.g., augmented expression of the BCR-ABL1 gene or mutations affecting the kinase domain of BCR-ABL1 that dodge TKI binding, such as T315I) [12], whereas others are categorized as BCR-ABL1-independent processes (e.g., alterations of efflux influx pumps, quiescent CML stem cell persistence, additional chromosomal abnormalities emergence or SRC family tyrosine kinases overexpression) [8,12,13].
Regarding BCR-ABL1-dependent TKI resistance, it has been demonstrated that the expression levels of the oncoprotein are unaltered with the inhibition of its kinase activity. Consequently, CML patients receiving TKI therapy normally need continuous treatment. Moreover, BCR-ABL1 can act as a scaffold protein, establishing signaling networks that are modified but not blocked by TKIs [14,15]. For these reasons, compounds able to reduce not only kinase activity but also the levels of the oncoprotein itself may represent an effective alternative therapy for CML. Importantly, the potential relevance of this alternative pharmacological strategy has recently been reinforced by the advent of PRO-TACs (proteolysis targeting chimeras), which employ the ubiquitin-mediated degradation system to target oncogenic proteins [16], or the use of CRISPR (clustered regularly interspaced short palindromic repeats) technology to provide permanent oncogene knockout [17] in CML models.
BCR-ABL1-independent IM resistance is associated with increased alternative survival signaling pathways (e.g., RAF/MEK/ERK) that confer apoptosis protection [18]. In this sense, preclinical and clinical studies have suggested that the combination of TKIs with multikinase inhibitors might be an attractive alternative to selective TKIs alone. Considering all these insights, identifying alternative drugs that overcome the main limitations of current CML treatments is an important goal to be achieved by the cancer research community.
Chemical structures based on natural products are widely utilized in drug discovery. In this study, we focused on naphthoquinones (NPQ), a chemical core with numerous medicinal properties (e.g., antibacterial, antifungal, antiviral, anti-inflammatory, antiparasitic or antitumoral) [19]. At present, quinone derivatives are commonly used in the clinic to treat solid tumors (e.g., doxorubicin) or hematological disorders (e.g., daunorubicin). Additionally, new NPQ-based compounds are being investigated in preclinical contexts with promising results in several types of malignancies (e.g., hepatocarcinoma, melanoma, leukemia, lung cancer, breast cancer or prostate cancer) [20]. In this regard, we have previously described original NPQ derivatives [21] and NPQ-hybrid compounds [22] that exhibited effective in vitro and in vivo anti-CML properties and that have contributed to the chemical design of the compound of this study, a novel NPQ-pyrone hybrid. Pyrones have also been described to exert strong cytotoxic activity against leukemic and ovarian cancer cells [23], reduce the viability of prostate cancer cells [24] or even sensitize tamoxifen-resistant breast cancer cells [25]. Therefore, the chemical structure of NPQ-pyrone represents a promising candidate to investigate its associated antitumor mechanisms.
Herein, the present work investigates the pharmacological efficacy and molecular mechanisms involved in the action of a new NPQ-pyrone hybrid compound, 7-(3,4-dimethoxyphenyl)− 10-methylbenzo[h]pyrano [4,3-b] chromene-5,6,8(7 H)-trione, named JKST6, in the context of CML. This study pays special attention to combinatorial therapy with IM and overcoming TKI resistance. Taken together, these findings provide a novel opportunity to find a drug strategy with a wider therapeutic window.

Wash-out assays
Cells were seeded at exponential growth (10,000 cells per well) in a 96-well plate (BD Falcon, France) and then treated with vehicle (0.05% DMSO), IM (0.03-5 µM), doxorubicin (0.1-10 µM) or JKST6 (0.1-10 µM) for 3 h. Then, drugs were removed after thorough washing with medium at 37 • C, and cells were left in drug-free medium for 48 h until cell viability was evaluated using the MTT assay [28].

Real-time monitoring of tumor growth (2D and 3D assays)
First, for 2D assays, CML cell lines were seeded at 5000 cells per well in a poly-L-lysine-coated 96-well plate (BD Falcon, France). Once the cell monolayer was stable after 16 h, it was treated with vehicle (0.05% DMSO) or test compound (0.03-10 µM). YOYO-1 (25 nM) (Invitrogen) was used to monitor cell membrane integrity [22]. Cell proliferation and viability were analyzed using microphotographs taken by an IncuCyte™ HD real-time imaging system (Essen BioScience, UK). The effects of each compound dose on cell growth and membrane integrity were monitored for 5 days. Then, area under the curve (AUC) values were plotted to analyze both the kinetics and dose effects on cell proliferation (IC50) and cytotoxicity (EC50). To support 2D assays, the effect of drugs on cell spheroids was also studied by seeding 3000 cells per well in round-bottomed, ultralow attachment, 96-well plates (ULA, Corning) [29]. Next, 24 h after spheroid formation, cells were treated with vehicle (0.05% DMSO) or drugs. The volume size and cytotoxicity of spheroids were recorded for 5 days. To calculate spheroid volume, confluence measurements were extracted using two different metrics: average phase object area (µm 2 ) and object count (1/image). These metrics were multiplied to obtain the total object area (µm 2 ), which was further used to calculate the spheroid radius (r = √Total Object Area/π) and the spheroid volume (µm 3 ) (V = 4/3 πr 3 ) [29].

Luciferase reporter gene assay
The potential effects of drugs on STAT3 and STAT5-mediated transcriptional activity were evaluated using the Renilla Luciferase Reporter cell lines (Leeporter™ Abeomics) HEK293 and Ba/F3, respectively. Cells were seeded at 500,000 cells per well in a 12-well culture plate and 600,000 cells per well in a 24-well culture plate each. Vehicle (0.05% DMSO) or compounds (0.3-10 µM) were added before stimulation of Ba/F3 and HEK293 cells with 30 ng/ml murine IL3 (mIL3) or 10 ng/ml human IL6 (hIL6), respectively. Cells were lysed in Passive Lysis Buffer (Promega, USA), and luciferase activity was determined using a Renilla-Firefly Luciferase Assay kit (Thermo Scientific, USA) in a microplate reader (Clarity™ R2, BioTek). The results are expressed as relative luciferase units (RLUs) per mg of protein and were normalized to the results obtained in vehicle-treated cells. In parallel, MTT assays were carried out to investigate the effects of the compounds on the viability of the reporter cell lines.

Cell cycle and apoptosis analysis
For cell cycle studies, unsynchronized K562 cells were seeded at 300,000 cells/well in 6-well plates and treated with vehicle (0.05% DMSO) or JKST6 for 12, 24 and 48 h. After treatment, cells were harvested, washed in PBS, and fixed overnight in cold 70% ethanol. Then, the cells were resuspended in PBS containing DNAse-free RNAse (500 μg/ml) for 2 h at 37 • C, incubated with propidium iodide (PI) (5 μg/ml), and analyzed in an Accuri C6 flow cytometer (BD). A FITC-Annexin V Apoptosis Detection Kit I (BD Biosciences) was used for cell death analysis. Briefly, cells were resuspended in 100 μl of ice-cold binding buffer (10 mM HEPES pH 7.4, 140 mM NaCl, 2.5 mM CaCl 2 ), containing 5 μl of Annexin V FITC and 5 μl of PI (50 μg/ml), incubated for 15 min at room temperature, and finally, binding buffer (400 μl) was added. In both experiments, 50,000 events were collected and analyzed by fluorescence-activated cell sorting (FACS). Early apoptosis was also assessed using bisBenzimide Hoechst 33258 (Sigma-Aldrich, USA), and microphotographs were used to count the proportion of apoptotic nuclei per total nuclei observed.

Real-time PCR
Cells were seeded at 300,000 cells/ml in 100-mm dishes and treated with vehicle (0.05% DMSO), JKST6 (3 µM) or IM (1 µM) for 1, 3, 6 and 12 h. Then, the cells were harvested, and total RNA was isolated using PureZOL RNA Isolation Reagent (Bio-Rad, CA, USA). RNA concentration and acid nucleic purity were measured with a NanoDrop 1000 spectrophotometer (Wilmington, USA). Then, 1 µg of total RNA was used for reverse transcription to cDNA with the iScript™ kit (Bio-Rad, CA, USA) following the manufacturer´s protocol. Real-time quantitative PCR (qPCR) was performed using SYBR Green PCR Master mix (Applied Biosystems, Massachusetts, USA). Exon-specific primers (Table 1) [30] were purchased from STAB VIDA (Caparica, Portugal). Gene expression was analyzed with an Mx3005P qPCR System (Agilent, CA, USA). The cycle threshold values were determined using MxPro qPCR Software (Agilent, CA, USA), and target gene expression was normalized to the GAPDH housekeeping gene. All amplifications were performed in duplicate, and Ct scores were averaged for subsequent calculations of relative expression values according to the comparative CT method for qPCR [31].

Drug combination assays
K562-GFP+ cells were seeded at 5000 cells per well in a 96-well plate (BD Falcon, France), and cell proliferation was analyzed using microphotographs in an IncuCyte™ HD real-time imaging system (Essen BioScience, UK). Briefly, cells were exposed to constant ratio combinations of JKST6 with IM by doubling dilutions of the individual drugs over a wide range of concentrations. Inhibition of cell proliferation, relative to untreated controls, was assigned as the effect and ranged from 0 (no cell proliferation inhibition) to 1 (100% cell proliferation inhibition). Dose-effect curves of single or combined drug treatments were plotted and analyzed by the median effect method of Chou and Talalay [32] using Calcusyn software 2.0 (Biosoft, Cambridge, UK). The analysis generated the combination index (CI) values. CI values less than 1, equal to 1, and greater than 1 indicated synergism, addition, and antagonism, respectively.

Acute and subchronic toxicity in vivo
Acute and subchronic studies were approved by the Bioethics Committee of University of La Laguna (CEIBA 2018-0291) and the University of Las Palmas de Gran Canaria (OEBA-44/2019R1), respectively. The experiments were performed according to OECD guidelines for the testing of chemicals (N • 407; 452). The Irwin test [33] was performed in FVB male mice (Charles River Code: 207) to investigate intraperitoneal (i.p.) and oral JKST6 acute toxicity. Briefly, 12 mice were randomly distributed into 4 groups of treatments: 1) i.p. treated with vehicle (dimethilsulfoxide (DMSO): PEG400: saline; 10:40:50); 2) i.p. treated with JKST6 (10 mg/kg); 3) orally administered vehicle (DMSO: methylcellulose 0.5% (MC); 10:90); and 4) orally treated with JKST6 (50 mg/kg). Animals were treated each day for 5 consecutive days. Every day after treatment, clinical signs were gathered by two independent researchers. After 5 days, animals were euthanized by carbon dioxide overdose, necropsy allowed us to explore morphological changes in organs (liver, spleen, kidney), and every organ weight was normalized by animal body weight. The subchronic toxicity assay was performed in BALB/c (Charles River code: 028) male (n = 10) and female (n = 10) mice for 30 days. Vehicle (DMSO: PEG400: saline; 10:40:50) or JKST6 (10 mg/kg) was i.p. administered every two days. Animal welfare was assessed daily to detect any clinical signs. Body weight and food and water consumption were tracked every day of treatment. Body composition (lean, fat and fluid) was determined using an MRI scan (Bruker Minispec, Massachusetts, USA). At Day 30, mice were euthanized by isoflurane overdose followed by necropsy analysis. Blood cell counting (MS4-5, Melet Schloesing, Osny, France) and biochemistry analyses (Pointcare v2, Mano Médical, France) were also performed.

Statistical analysis
The data shown are the results of two to six independent experiments with at least three replicates per experimental condition. The results are expressed as the mean ± standard error of the mean (SEM) or as the mean ± standard deviation (SD) in the in vitro and in vivo assays, respectively. Differences between means of two groups were analyzed with two-tailed Student's t-test. One-way ANOVA followed by Bonferroni post hoc test was used to compare means from more than two groups using GraphPad Prism 8.3.0 Software (CA, USA). Statistical significance was considered when p < 0.05 with a 95% confidence interval. The concentration required to reduce cell viability/proliferation by 50% (IC50) or to induce cytotoxicity by 50% (EC50) was graphically determined using the curve fitting algorithm of GraphPad Software. For the in vivo experiments, equal numbers of animals were selected and randomly distributed into each treatment group. The minimal number of animals to be used was determined with statistical power and sample size analysis using G Power 3.1 Software (Düsseldorf, Germany) and followed the OECD recommendation (N • 407).

Antitumoral efficacy of JKST6 on human leukemia cells
JKST6 was identified within a naphthoquinone-enriched library by high-throughput cell-based phenotypic screening (Fig. 1A). First, the antitumoral effects of JKST6 were evaluated by MTT assay. Interestingly, JKST6 showed higher efficacy against cell lines derived from hematological malignancies (i.e., K562, MOLM-13, HL-60, HEL, MV 4.11, JURKAT cell lines) than against breast cancer-derived cells (i.e., BT-549, MDA-MB-231, HS-578 T) or nonmalignant cells (i.e., RAW 264.7, Vero, hMSCs, and PBMCs) ( Table 2). The cytotoxicity of JKST6 on quiescent human PBMCs was 4-fold lower than that in K562 cells (Table 2). Remarkably, low concentrations (1 and 3 µM) of JKST6 showed no significant effect on PBMC viability compared to K562, whereas doxorubicin, a chemotherapeutic agent used in a broad spectrum of tumor types, reduced PBMC and K562 viability with similar potency (Fig. 1B). The low toxicity displayed by JKST6 in vitro was also validated in vivo. JKST6 (10 mg/kg) was administered to mice in acute (5 days, data not shown) and subchronic (30 days) toxicity assays. The results confirmed that the clinical signs, blood cell populations or biochemical parameters of JKST6-treated animals were mostly unaffected when compared with those of the vehicle-treated group (Supplementary Figure 1, Supplementary Tables 1-2).
As shown in Table 2, JKST6 exhibits potent antitumoral effects in K562 cells, a model of human CML. Thus, we further investigated the irreversible effects of a short exposure (3 h) of K562 cells to JKST6. IM and doxorubicin were also analyzed in this set of experiments. The data revealed that JKST6 (IC50 = 1.57 ± 0.28 µM) and doxorubicin (IC50 = 0.74 ± 0.12 µM) maintained their antitumoral potency after short exposure (3 h) of K562 to the drugs followed by 48 h of wash-out, whereas IM reduced its antitumoral effect (IC50 > 5 µM) (Fig. 1C). These results suggest that short and transient exposure to JKST6 is enough to maintain, at least for 48 h, the inhibitory effects induced by Furthermore, 2D real-time monitoring of K562 cell growth permitted to discern cytostatic (lower than 1 µM) and cytotoxic (higher than 3 µM) JKST6 doses in a long-term experiment (5 days) (Fig. 1D). Instead, IM showed a cytostatic effect around 0.25 µM and cytotoxic effects at doses higher than 1 µM (data not shown). As a representative image, brightfield microphotographs of K562 cells exposed to 3 µM JKST6 or vehicle are shown (Fig. 1D). In addition, kinetic studies showed the rapid antiproliferative effects of JKST6 (IC50 = 0.91 µM at 12 h) compared to IM (IC50 > 5 µM at 12 h) (Supplementary Figure 2). To better mimic a real tumor in vivo, 3D spheroid experiments with K562 cells were performed. Notably, JKST6 significantly reduced spheroid growth compared to vehicle-treated K562 cells (Fig. 1E). Moreover, JKST6 cytotoxicity especially affected the K562 spheroid perimeter at short times (48 h) when compared to IM cytotoxicity (Fig. 1F). Taken together, these results show that JKST6 has a rapid and long-term antitumor effect in CML cells and that its low toxicity in healthy cells might be clinically relevant.

JKST6 inhibits cell cycle entry and induces apoptosis in human CML cells
To further evaluate the manner in which JKST6 suppresses K562 cell growth, we performed cell cycle profiles and apoptosis studies by flow cytometry. As shown in Fig. 2A-B, all doses of JKST6 promoted a significant increase in the subG1 phase, accompanied by a corresponding decrease in the G0/G1 phase, especially after 24 and 48 h of drug exposure. Consistently, changes in proteins involved in cell cycle regulation were also observed after JKST6 treatment (Fig. 2C). In particular, the compound markedly reduced the phosphorylation of pRB, a protein that controls cell cycle entry, at only 6 h posttreatment, whereas Cyclin E, which is associated with G1-S transition, showed a peak at 24 h followed by a reduction at 48 h. Proteins that have key roles in the S and M phases were also altered by JKST6. Thus, JKST6 increased Cyclin A and B levels after 48 h of treatment. We also observed that Wee1, a checkpoint protein that negatively controls mitosis entry, was phosphorylated at short times of compound exposure. A reduction in the activation of H3, a histone that is phosphorylated in different patterns in mitosis, was also evident from 6 to 48 h posttreatment. Moreover, certain negative regulators of mitosis progression (pCHK1, pCHK2, p21) were activated after JKST6 treatment (Fig. 2C).
The increase in subG1 phase implies an enhanced number of hypodiploid cells, which in turn represent dead cells. To identify the mechanism of death induced by JKST6, we performed apoptosis assays that measured Annexin V and propidium iodide (PI) staining using doses and time points used for cell cycle analysis. Significant time-and dosedependent increases in Annexin V-positive cells were detected (Fig. 3A), along with an increase in the number of apoptotic nuclei 24 h after JKST6 treatment (Fig. 3B). Apoptotic death can be dependent or independent of the activity of certain caspases. Therefore, to better understand the death mechanisms activated by JKST6, different initiator and effector procaspases, as well as poly (ADP-ribose) polymerase (PARP), were studied. As shown in Fig. 3C-D, JKST6 treatment induced the cleavage and activity of procaspases 8 and 9 at 12 h, whereas the cleavage of procaspase-3 and PARP was detected from 12 to 48 h after treatment. Moreover, phosphorylation of the DNA double strand breaks marker H2AX was detected just 6 h after treatment, which, together with Chk1 and Chk2 activation, indicated the presence of genetic damage. Finally, Mcl-1, a pro-survival Bcl2 family protein, was induced after short-term JKST6 treatment (Fig. 3C). Taken together, these data revealed that JKST6 reduces K562 cell growth, inhibits cell cycle entry, causes DNA damage, and induces apoptosis.

JKST6 inhibits BCR-ABL1/STAT5 and modulates associated survival pathways in human chronic myelogenous leukemia
To determine whether JKST6 effectively targets the BCR-ABL1/ STAT5 signaling pathway, protein and mRNA expression analyses were performed in K562 cells. Interestingly, hyperphosphorylation of BCR-ABL1 (p Y412 and p Y177 ) was efficiently inhibited by JKST6 (3 µM at 3 h) (Fig. 4A). However, this finding was concomitant with a reduction in the chimeric BCR-ABL1 protein measured in ABL and BCR domains, whereas BCR and c-ABL wild-type proteins remained practically unaltered after JKST6 treatment (Fig. 4A). Importantly, constitutive phosphorylation of pSTAT5 Y694 was inhibited after JKST6 treatment without altering total STAT5 levels (Fig. 4A). To explore the mechanism that causes a reduction in BCR-ABL1 after JKST6 treatment, the levels of polyubiquitinated proteins were studied. The results showed evident increases in ubiquitinated proteins after short-term treatments (0.5, 1 and 3 h) compared to vehicle-treated cells (Fig. 4A). The extent of the posttranslational effect caused by JKST6 was studied using a proteasome inhibitor (MG132) to block proteolytic activity [34] prior to compound addition, as shown in Fig. 4B. Cells were pretreated with MG132 for 2 h before JKST6 treatment for another 1, 3 or 6 h. First, JKST6 alone caused a peak in ubiquitinated proteins at 1 h after treatment, as previously observed. MG132 induced an evident increase in ubiquitinated proteins compared with the vehicle at all time points. Moreover, MG132 (2 h) followed by JKST6 resulted in the potentiation of ubiquitinated proteins after 1 and 3 h of treatment and, importantly, partially prevented the JKST6-induced downregulation of BCR-ABL1 protein at 1 h. Surprisingly, the mixture enhanced the reduction in total oncoprotein levels after 3 and 6 h of treatment (Fig. 4B).
The expression of other key components involved in BCR-ABL1/ STAT5 signaling was analyzed, including CrkL [35], c-MYC and PIM1 [4]. As shown in Fig. 4C, all these proteins were also downregulated by JKST6 treatment, thus indicating a blockage of the BCR-ABL1/STAT5 downstream cascade by the compound. To understand whether STAT transcription factor activity may be downregulated by JKST6, Renilla luciferase assays were performed. Notably, in the BaF3 reporter cell line, JKST6 inhibited IL3-induced STAT5 luciferase activity (IC50 = 4.26 ± 0.05 µM) without disturbing cell viability (IC50 = 9.58 ± 0.58 µM) (Fig.  5A). Although less potent (IC50 = 7.18 ± 1.48 µM), JKST6 treatment also decreased IL6-induced STAT3 transcriptional activity in the HEK293 reporter cell line without causing toxicity (IC50 > 10 µM) (Fig. 5B). RNA analysis was performed to assess whether JKST6 provokes gene expression inhibition alongside protein content reduction of BCR-ABL1 using two different primer pairs that amplify exons 12 and 13 of BCR (BCR) and the junction region of  the oncoprotein (b3a2). In addition, STAT5 target genes (MYC and PIM1) were also studied (Fig. 5C). Notably, the relative gene expression profile showed a different pattern of suppression than the one detected in the immunoblot assays, highlighting the earlier decrease in BCR-ABL1 protein levels (3 h posttreatment) followed by later reductions in BCR and b3a2 mRNA levels (12 h posttreatment). Sustained MYC mRNA downregulation was produced by JKST6, whereas PIM1 downregulation was observed at short times followed by enhanced expression at 12 h. Finally, alternative survival pathways also modulated by BCR-ABL1 in CML were analyzed [7]. Accordingly, JKST6 inhibited the phosphorylation of Src family kinases at 12 h, as well as P38-MAPK and S6, and transiently augmented AKT and JNK phosphorylation when compared to nontreated K562 cells (Fig. 6A-B). ERK1/2 phosphorylation seemed to be increased at 30 min and remained augmented until 12 h after treatment (Fig. 6A). Therefore, these results suggest that JKST6 negatively modulates the BCR-ABL1/STAT5 pathway in K562 cells.

JKST6 potentiates imatinib-induced antiproliferative effects and STAT5 inhibitory activity
Drug combinations allow the use of lower doses that hinder drug resistance development or promote more effective target inhibition [21]. Hence, to evaluate whether JKST6 enhances the antiproliferative effect of IM, dose-response synergistic assays combining both compounds were performed. For these studies, IM-sensitive K562-GFP+ cells were treated with different doses of JKST6 and IM, following a constant ratio combination design [7.5(JKST6):1(IM)] at different time points. Very interestingly, we found a significant increase in IM potency compared with IM alone when combined with JKST6 at all time points (Fig. 7A). Furthermore, isobologram and Chou-Talalay analyses of the combination index (CI) [32] showed that JKST6 and IM acted synergistically in inducing antiproliferative effects (CI (24 h) for ED50 = 0.42 ± 0.10 and CI (24 h) for ED75 = 0.38 ± 0.11)/(CI (48 h) for ED50 = 0.55 ± 0.16 and CI (48 h) for ED75 = 0.46 ± 0.13) (Fig. 7B). Importantly, synergistic effects of the JKST6-IM combination on pSTAT5 Y694 levels were also observed. Thus, when K562 cells were treated with different combination doses of IM (0.03 or 0.05 µM) and JKST6 (0.5 or 1 µM) for 3 h, the phosphorylation of STAT5 at Y 694 was significantly more inhibited than when cells were treated with individual compounds (Fig. 7C).
To examine the irreversible effects of JKST6 on IM-resistant cells, wash-out experiments were performed by short-term exposure (3 h) of K562-R cells to JKST6, doxorubicin or IM followed by drug removal from the media for 48 h. As expected, IM was ineffective (IC50 >5 µM), whereas both JKST6 (IC50 = 2.86 ± 0.17 µM) and doxorubicin (IC50 = 2.12 ± 1.12 µM) were able to significantly reduce the viability of K562-R cells (Fig. 8B). Predictably, live cell imaging of K562-R proliferation in long-term experiments (5 days) showed that IM was less potent on K562-R cells (IC50 = 4.55 ± 1.31 µM) than on K562 cells (IC50 = 0.24 ± 0.2 µM), whereas JKST6 remained equally effective in both cases (IC50 = 1.07 ± 0.19 µM versus IC50 = 0.75 ± 0.10 µM) (Fig. 8C). For a pathophysiological approach, 3D spheroid live-cell imaging assays were performed in AR230-R cells and showed a JKST6 dose-time-dependent volume shrinkage, whereas IM was unable to exert a significant effect on tumor volume (Fig. 8D). Moreover, JKST6 caused rapid (24 h) damage to peripheral cells, while IM cytotoxicity was visible at 24 and 48 h but seemed to be ineffective at 72 h (Supplementary Figure 3A). Finally, due to the importance that BCR-ABL1 overexpression has in IM resistance, we investigated whether JKST6 maintained its molecular mechanism of action in AR230-R cells using AR230 cells as a control (Fig. 9A). Notably, 3 µM JKST6 effectively reduced the levels of total BCR-ABL1 in both clones and thus BCR-ABL1 phosphorylation (pY 177 , pY 412 ) at just 3 h after treatment (Fig. 9A). Likewise, hyperphosphorylation of STAT5 (higher in AR230-R cells) and PIM1 and c-MYC protein expression levels were significantly inhibited at 3 h of JKST6 exposure. In contrast, IM was considerably less efficient in inhibiting BCR-ABL1 and STAT5 phosphorylation in IM-resistant AR230-R cells (Fig. 9A-right panel) than in their IM-sensitive counterparts (Fig. 9A-left panel). Moreover, JKST6 induced γH2AX (at 3, 6 and 12 h after treatment) and reduced CrkL phosphorylation (at 6 and 12 h after treatment) to a similar extent in both cell clones. However, IM was not able to induce γH2AX phosphorylation in AR230-R cells (Fig. 9A-right panel). To investigate whether the reduction in BCR-ABL1 protein levels by JKST6 could contribute to restoring the (caption on next column)   , ) for the indicated times. Data are presented as the mean ± SEM of three independent experiments, each performed in duplicate. Statistical significance was assessed using one-way ANOVA with Bonferroni post hoc test. *P < 0.05; **P < 0.01; **** P < 0.0001 versus VEH-treated cells.
sensitivity of AR230-R cells to IM, cells were pre-exposed to vehicle or JKST6 for 3 h, followed by wash-out, and the dose effect on IM was evaluated at 48 h (Fig. 9B). Of interest, IM potentiated its effects by 14 times when cells were pretreated with JKST6 (IC50 = 0.175 ± 0.003 µM) compared to vehicle pre-exposed cells (IC50 = 2.17 ± 0.17 µM). JKST6 inhibited the same key proteins involved in the BCR-ABL1/STAT5 pathway in the AR230 and K562 CML models; thus, the effect of JKST6 on mRNA expression was analyzed in K562-R cells (Fig. 9C). JKST6 provoked a different modulatory pattern of mRNA expression when compared to the protein levels, according to previous studies in K562 cells. The gene expression of SOCS2, a negative regulator of the JAK/STAT pathway associated with CML progression to the aggressive blast phase [37], was also assessed in both K562 and K562-R cells (Fig. 9C). IM treatment reduced the expression of SOCS2, whereas JKST6 provoked a peak of induction at 6 h in both cell lines. In summary, JKST6 displays the same potential for inhibiting the BCR-ABL1/STAT5 signaling pathway and its downstream effectors in both IM-resistant and IM-sensitive cells. Importantly, JKST6 might rescue the sensitivity of IM-resistant CML cells to IM.

Discussion
NPQ derivatives have emerged as promising therapeutic agents in cancer in the last decade. The present work reveals the pharmacological efficacy of a novel NPQ-pyrone hybrid, JKST6, in human CML cells and shows the molecular mechanisms responsible for its potent antitumoral effects. Interestingly, we show how the JKST6 compound is effective in overcoming imatinib resistance in a wide range of resistant cells with known resistance mechanisms present in CML patients; thus, it is a promising new pharmacological tool that could improve current therapeutic opportunities.
Antiproliferative capacity and low toxicity are two crucial aspects of a clinically relevant antitumoral drug. Therefore, the antiproliferative effects of JKST6 in hematological and nonhematological cancer and normal cells were investigated. Our data show that JKST6 effectively inhibits the growth of human leukemia cells with high potency (i.e., sub µM IC50 values in K562 and MOLM-13) compared to solid tumor cells (e. g., BT-549, MDA-MB-231, MCF7). Remarkably, the viability of nonmalignant cells was hardly affected by JKST6 treatment, which agrees with the low toxicity observed by compound administration in vivo. These results suggest that JKST6 may find specific molecular targets in myeloid-derived cancer cells, which makes them more sensitive to the compound than other leukemias or nonhematological cancer cells. JKST6 displays rapid antiproliferative activity and longer-lasting antitumoral effects on CML cells after short drug exposure when compared to the BCR-ABL1 inhibitor IM. These findings outline a JKST6 durable action that is normally associated with ligand-target covalent bonds, long drug-target residence, or posttranslational effects after drug-target union [38,39]. Interestingly, JKST6 shows a wide window between cytostatic and cytotoxic doses, coinciding with reported studies of pyrones that display elevated antiproliferative over cytotoxic activity [40]. Moreover, the antitumor efficacy of JKST6 was validated in CML cell spheroids, a culture system that faithfully reproduces tumor biology in vivo. In fact, differences between 2D monolayer and 3D cultures have been widely reported, and some of them include signal transduction, gene and protein expression, cell differentiation or drug sensitivity [41]. Furthermore, recent studies have found that leukemic cells are metabolically more stable in 3D than in 2D standard cultures, where they are prone to undergo phenotypic changes with passage [42]. Thus, the growth inhibition of CML spheroids induced by JKST6 in vitro may represent a valid approach for its potential antitumor efficacy in vivo [41].
It has been reported that NPQ-based derivatives and hybrids are able to induce antiproliferative effects on CML cells through diverse mechanisms, including oxidative stress, cell cycle arrest, apoptosis, inhibition of BCR-ABL1/STAT5 signaling and modulation of other survival pathways [21,22,43]. Accordingly, we have previously shown that quinone-containing compounds inhibit CML cell growth through cell cycle arrest [21,22]. Examples include the NPQ-based derivative CM363, which induces cell cycle arrest through augmented S phase [21], and the NPQ-coumarin hybrid NPQ-C6, which increases subG1 and reduces G0/G1 cell cycle phases [22]. In the present study, we demonstrate that JKST6 alters the CML cell cycle by augmenting subG1 and reducing G0/G1 phases in a time-and dose-dependent manner with higher potency than its precursor NPQ-C6 [22]. Importantly, these effects are associated with changes in the phosphorylation and/or expression of several proteins involved in cell cycle progression. Thus, JKST6 reduced retinoblastoma (RB) phosphorylation, which is necessary for cell cycle progression and is normally associated with cyclin D/Cdk4 inhibition [44]. Downregulation of RB phosphorylation has also been related to c-ABL kinase inhibition, which in turn plays a role in S phase progression [45]. Moreover, complexes formed by cyclin E-CDK2, cyclin A-CDK2 and cyclin B-CDK1 regulate the G1/S transition, S phase progression, and G2/M transition, respectively [46]. Interestingly, JKST6 modulates these cyclins together with a significant increase in pCDK1 Y15 . This phosphorylation leads to CDK1 inactivation, which is controlled by the Wee1 and Cdc25 balance [46]. NPQ-containing compounds have been previously identified as Cdc25 inhibitors [21,47], which block phosphatase activity and therefore CDK1 activation. Moreover, the phosphorylation of Wee1 at S 642 has been reported to be AKT-dependent and to cause Wee1 inactivation [48]. In fact, JKST6 transiently induces pWee1 S642 , which coincides with CDK1 active/inactive states. Similarly, JKST6 reduces mitotic histone 3 (H3) phosphorylation at Ser-10 (pH 3 S10 ), which has been reported to be required for proper execution of mitosis [49]. Furthermore, JKST6 induces the expression of p21 protein, a negative regulator of G1 and G2 progression, which has been related to the antiproliferative effects of other NPQs (e.g., shikonin and plumbagin) on leukemic cells [43,50]. In addition, JKST6 strongly stimulates phosphorylation of the downstream effectors of the ATM/ATR-dependent DNA damage checkpoint (pChk1 S296 and pChk2 T68 ) at short times after compound exposure, suggesting an early and potent induction of genotoxicity by the compound [51], which is a reported feature of NPQs and pyrones [52]. All these findings support the hypothesis that JKST6 prevents cell cycle progression by inducing cell arrest in the subG1 phase and triggering mitosis alterations, effects that might be caused, at least in part, by DNA damage.
Loss of apoptotic control allows cancer cells to survive longer; therefore, regulation of apoptosis constitutes a hotspot for any antitumoral drug. In this sense, JKST6 induces apoptosis in K562 cells, as assessed by increased Annexin V activity, apoptotic nuclei, activation of caspase-8, − 9 and − 3 and PARP cleavage. Previous studies have shown that NPQ derivatives and hybrids are able to induce apoptosis in CML cells through the induction of the intrinsic mitochondrial-dependent pathway [21,22,43], which involves the activation of caspases 9 and 3. However, the activation of caspase-8 (apoptosis extrinsic pathway) has also been implicated in NPQ-induced apoptosis in human leukemia cells [53]. Paradoxically, JKST6 upregulates the antiapoptotic protein Mcl-1, an effect that has also been reported in myeloid leukemia cells in response to certain chemotherapeutics [54]. Notably, JKST6 promotes an early increase in the double-strand DNA break marker γH2AX, further reinforcing the hypothesis that the compound triggers DNA damage related to the apoptotic process that cannot be repaired by CML cells.
Cell cycle arrest and activation of apoptosis induced by NPQ-based derivatives and hybrids in CML cells are generally associated with potent downregulation of BCR-ABL1/STAT5 signaling [21,22]. Accordingly, JKST6 leads to a rapid decrease in BCR-ABL1 phosphorylation (p Y412 and p Y177 ) in CML cells, which can be explained by a concomitant reduction in its total protein levels. These fast responses are   often associated with posttranslational events, predominantly regulated by ubiquitin conjugation and proteasome activation [55]. In fact, a rapid decrease in BCR-ABL1 protein levels has been described for drugs that alter the ubiquitin cycle [14,56]. Accordingly, JKST6-induced downregulation of BCR-ABL1 protein was preceded by an increased level of polyubiquitinated proteins. Importantly, previous inhibition of the proteasome by MG132 [34] followed by JKST6 treatment partially prevented JKST6-induced BCR-ABL1 protein downregulation at 1 h after treatment and provoked an increased level of polyubiquitinated proteins. These findings suggest that proteasome activation might contribute, at least in part, to the effects of JKST6 on CML. Moreover, protein polyubiquitination and decay were followed by a reduction in BCR-ABL1 mRNA levels, which supports the hypothesis that protein degradation precedes the inhibition of gene transcription. Although the exact posttranslational mechanism implicated in these effects deserves future research, compounds that cause ubiquitin cycle disturbances have been described as a successful approach to reduce BCR-ABL1 levels and subsequently activate cell death even in IM-resistant CML cells [14].
Interestingly, the rapid inhibition of BCR-ABL1 is effective downstream. In particular, the JKST6-induced inhibition of pSTAT5 Y694 predictably plays a key role in the antiproliferative effects of this compound due to its crucial roles in cell transformation, survival and growth in this neoplasm [4]. JKST6 also shows important functional effects inferred from the inhibition of pCrkL Y207 and STAT5 target genes, such as PIM1 and c-MYC. These findings are particularly relevant, as c-MYC and BCR-ABL1 can upregulate each other; MYC participates in cell transformation, genomic instability and suppression of cell differentiation, thus favoring a poor response to IM and therefore a higher aggressiveness of disease [22,27]. Moreover, c-MYC inhibition caused by NPQ-based derivatives and hybrids has directly been linked with apoptosis induction in CML [21,22,57]. PIM1 is also upregulated by BCR-ABL1/STAT5 signaling and has roles in cell survival, transformation, and apoptosis protection against DNA damage [58]. Furthermore, PIM1 inhibition has been seen to play a central role in the anti-CML pharmacological actions of NPQ-based derivatives and hybrids [21,22]. Interestingly, we observed that JKST6 also had a regulatory role in STAT-mediated transcriptional activity, evidenced by the inhibition of IL3/JAK2/STAT5-dependent transcription. Taken together, these results support the hypothesis that JKST6 downregulates the BCR-ABL1/STAT5 signaling pathway as part of its anti-CML properties.
BCR-ABL1 exerts proleukemic actions by modulating additional signaling pathways, such as PI3K/AKT/mTOR and RAS/MAPK [4]. Notably, JKST6 inhibits the activation of the Src family of kinases, in contrast to IM, which is important due to the involvement of Src deregulated signaling in CML progression and TKI resistance [12]. Remarkably, JKST6 modulates MAPK signaling in CML cells, inducing transient activation of JNK, moderate inhibition of p38MAPK activation, and sustained phosphorylation of ERK1/2. Conversely, other NPQ-based derivatives and hybrids have been shown to downregulate ERK1/2 phosphorylation concomitantly with their antiproliferative actions in CML cells [21,22,43]. However, it has also been described that pERK1/2 T202/Y204 induction is related to activated apoptosis in human leukemia cells [59], which may explain the results observed with JKST6. Alternatively, JNK transient stimulation and pP38MAPK T180/Y182 downregulation have already been reported to be associated with the antiproliferative effects of NPQ derivatives and hybrids in CML [21,22]. Interestingly, the BCR-ABL1/GBR-2 complex recruits and stimulates PI3K in CML cells, and subsequently, AKT and mTOR are activated and promote cell growth and survival [60]. mTOR signaling controls mRNA translation via its direct effector, the ribosomal protein S6 [60]. Furthermore, it has been reported that the constitutive activation of mTOR observed in several tumors can downregulate receptor tyrosine kinase signaling (e.g., IGF-IR) and, eventually, AKT activation as a negative feedback mechanism [61]. Interestingly, JKST6 potently and  , ) for the indicated times. Data are presented as the mean ± SEM of three independent experiments, each performed in duplicate. Statistical significance was assessed using one-way ANOVA with Bonferroni post hoc test. *P < 0.05; **P < 0.01; *** P < 0.001; ****P < 0.0001 versus VEH-treated cells.
rapidly inhibits S6 and increases pAKT T308/S473 , effects that suggest that the compound might inhibit mTOR. However, whether JKST6 exerts rapamycin-like effects in CML and the relevance of these potential properties on its molecular mechanism of action require further research.
TKI intolerance and IM resistance are current key challenges in CML therapy [8]. Therefore, one of the crucial aspects within the studied JKST6 properties has been its ability to potentiate the TKI effect and overcome IM resistance. JKST6 acts synergistically with IM, inhibiting cell proliferation and STAT5 activation in K562 cells, which is a promising finding because it opens an opportunity to reduce the effective dose of IM and, subsequently, the risk of long-term associated toxicity. In addition, JKST6 is highly effective in both AR230-R and K562-R cells, where IM resistance is mainly associated with increased BCR-ABL1 expression [36] or overexpression of Src family kinases [62], respectively. Similarly, JKST6 is highly effective on K562mycT315I cells, whose IM resistance is linked to a point mutation at the so-called gatekeeper residue of BCR-ABL1 [11]. Importantly, this mutation confers resistance to not only IM but also second-generation BCR-ABL1 inhibitors such as dasatinib and nilotinib [63]. JKST6 inhibition of the BCR-ABL1/STAT5 signaling pathway in AR230-R cells supports the hypothesis that the compound maintains its mechanistic modulation against the hyperactivation of the oncogenic pathway, suggesting an important utility of JKST6 in an already established IM-resistance context. The efficacy of JKST6 on CML-resistant cells was further corroborated by the inhibition of MYC and PIM1 mRNA levels [4]. SOCS2, a ubiquitin ligase that negatively regulates the JAK-STAT pathway [64], has been described to be overexpressed in BCR-ABL1 + cells, and this expression could be downregulated by IM treatment. Moreover, refractory and relapse responses to IM are associated with enhanced SOCS2 expression [65]. Accordingly, our results showed higher basal SOCS2 mRNA levels in IM-resistant cells than in IM-sensitive cells. Additionally, significant IM-induced downregulations of SOCS2 expression were detected in both cell lines, contrasting with the described IM inability to modulate SOCS2 in resistant cells. This difference could be a result of the different mechanism of resistance present in K562-R cells, which is related to Lyn overexpression. Conversely, JKST6 stimulated SOCS2 expression after 6 h of treatment in both K562-R and K562 cells, which might contribute to the negative regulation of BCR-ABL1/STAT5 signaling in these cells. In fact, SOCS2 overexpression may play a role in the ubiquitination of BCR-ABL1 induced by JKST6 treatment since the SOCS family participates in target protein degradation [66]. Interestingly, JKST6 is able to circumvent IM resistance in CML cells overexpressing BCR-ABL1, probably largely motivated by a reduction in oncokinase levels, which seems effective in restoring sensitivity to TKI treatment.

Conclusions
In summary, JKST6 possesses a combination of interesting properties that make it into a promising compound to further study in order to overcome the main CML therapy limitations. First, JKST6 efficiently reduces BCR-ABL1, preventing not only its constitutive activation but also its function as a "scaffold protein" that in turn modulates alternative survival pathways. Second, this effective feature of JKST6 activity is maintained in cells that are resistant to IM by BCR-ABL1-dependent or BCR-ABL1-independent mechanisms. Third and clinically relevant, JKST6 potentiates IM-induced antitumoral effects, suggesting that it could be considered in combinatorial therapy with TKIs in CML.