Abstract
Internal tandem duplication (ITD) or tyrosine kinase domain mutations of FLT3 is the most frequent genetic alteration in acute myelogenous leukemia (AML) and are associated with poor disease outcome. Despite considerable efforts to develop single-target FLT3 drugs, so far, the most promising clinical response has been achieved using the multikinase inhibitor midostaurin. Here, we explore the activity of the indolocarbazole EC-70124, from the same chemical space as midostaurin, in preclinical models of AML, focusing on those bearing FLT3-ITD mutations. EC-70124 potently inhibits wild-type and mutant FLT3, and also other important kinases such as PIM kinases. EC-70124 inhibits proliferation of AML cell lines, inducing cell-cycle arrest and apoptosis. EC-70124 is orally bioavailable and displays higher metabolic stability and lower human protein plasma binding compared with midostaurin. Both in vitro and in vivo pharmacodynamic analyses demonstrate inhibition of FLT3-STAT5, Akt-mTOR-S6, and PIM-BAD pathways. Oral administration of EC-70124 in FLT3-ITD xenograft models demonstrates high efficacy, reaching complete tumor regression. Ex vivo, EC-70124 impaired cell viability in leukemic blasts, especially from FLT3-ITD patients. Our results demonstrate the ability of EC-70124 to reduce proliferation and induce cell death in AML cell lines, patient-derived leukemic blast and xenograft animal models, reaching best results in FLT3 mutants that carry other molecular pathways' alterations. Thus, its unique inhibition profile warrants EC-70124 as a promising agent for AML treatment based on its ability to interfere the complex oncogenic events activated in AML at several levels. Mol Cancer Ther; 17(3); 614–24. ©2018 AACR.
Introduction
Acute myelogenous leukemia (AML) is the most common myeloid malignancy in adults (1, 2) and internal tandem duplication (ITD) or kinase domain mutations of FMS-like tyrosine kinase 3 (FLT3) represent the most frequent genetic alterations accounting for nearly 30% of cases and being associated with poor disease outcome (3). FLT3 activation results in the induction of downstream prosurvival pathways, including MAPK/extracellular signal–regulated kinase, PI3K/Akt and STAT5, and causes increased cell proliferation with suppression of apoptosis (4). FLT3 has been a target of choice for many years and, as a result, several drugs are currently in advanced clinical trials (3). However, despite promising preclinical results, none of FLT3 inhibitors reported so far sustained its activity as single-agent therapy for. prolonged time, often due to the emergence of resistance in FLT3-ITD (5). In fact, best clinical data, including significant survival benefit, have been reported in the RATIFY trial with the multikinase inhibitor midostaurin, a semisynthetic analogue of the promiscuous kinase inhibitor staurosporine (6). This clinically validated success story signals that intensive research is needed to discover novel multikinase inhibitors targeting not only FLT3 but also alternative pathways that cover the complex oncogenic events activated in AML and the potential resistance mechanisms that can emerge after clinical treatments. For instance, resistance to FLT3 inhibition has been associated with upregulation of PIM kinases (7, 8).
In this sense, PIM proteins (PIM1, PIM2, and PIM3) are serine-threonine kinases with increased expression in a variety of malignancies, including AML (9–11). PIM kinases play an important role in enhancing cell survival and suppressing apoptosis in hematopoietic cells (12, 13), particularly PIM1 and PIM2 whose overexpression has been reported in AML blasts (14). Their expression can be induced by activation of STAT5 (15), one of the most important targets in constitutively activated FLT3, establishing a relationship between both pathways. In fact, constitutively activated FLT3 signaling upregulates PIM1 expression in leukemia cells and this upregulation contributes to the proliferative and antiapoptotic pathways induced by FLT3 signaling (16). Moreover, PIM kinase inhibition has been described to enhance FLT3 inhibitors activity (17). The feedback loop between the two proteins makes the combination of FLT3 and PIM inhibition a rational strategy for AML treatment (18).
EC-70124 is a hybrid indolocarbazole obtained by combination of genes from rebeccamycin and staurosporine biosynthesis pathways (19, compound #8). This compound displays a potent multikinase inhibitor spectrum affecting key intracellular kinases implicated in prosurvival and proliferative pathways. Thus, EC-70124 induces senescence of glioblastoma-initiating cells by inhibition of the NF-kB pathway (20) and exerts antitumoral activity in triple-negative breast cancer by the inhibition of both PI3K/mTOR and JAK/STAT pathways (21). Similar antitumoral activity has been also described in colorectal cancer mediated by PI3K/Akt pathway inhibition (22). Interestingly, EC-70124 also acts as a dual STAT3/NF-kB inhibitor, reverting both tumorigenic and stem cell properties in prostate cancer (23).
In this study, we characterize the pharmacologic and activity profile of this unique multikinase inhibitor in preclinical AML models. EC-70124 inhibits AML cell proliferation and induced cell-cycle arrest and apoptosis. In vivo oral treatment causes complete regression in AML tumor xenograft models, and ex vivo inhibits proliferation in AML patient blasts at biochemically relevant doses. Collectively, these results make EC-70124 a promising new candidate for the treatment of AML.
Materials and Methods
Cell culture and reagents
MV4-11 (catalog no. ACC-102), MOLM-13 (catalog. no. ACC-554), MOLM-16 (catalog no. ACC-555), KG1 (catalog no. ACC-14), OCI-M1 (catalog no. ACC-529), NB4 (catalog no. ACC-207), THP-1 (catalog. no. ACC-16), and HL-60 (catalog no. ACC-3) cell lines were purchased from DSMZ German collection and maintained in RPMI supplemented with 10% FBS and 1% antibiotic–antimycotic mixture containing 5,000 U/mL penicillin and 5,000 U/mL streptomycin except for MOLM-16, which was maintained in 20% FBS. Cells were maintained at 37°C in a humidified atmosphere of 5% CO2. Culture cells have been monitored to ensure that they are Mycoplasma-free using the LookOut Mycoplasma qPCR Detection Kit (Sigma Chemical Co.). All experiments were performed between passage 5 and 20. Cell culture reagents were purchased from Sigma (Sigma Chemical Co.) except for FBS, which was obtained from Gibco (Invitrogen Life Technologies). Culture flasks and dishes were acquired from Thermo Fisher Scientific.
EC-70124 was synthesized by a proprietary process by EntreChem S.L. Midostaurin was synthesized from staurosporine (Biomar Microbial Technologies) and the identity of the isolated product verified by comparison with an authentic sample (HPLC, NMR).
Isolation of leukemic cells from patient samples
AML patient bone marrow aspirates, collected during routine examination, were obtained from Hospital Universitario Central de Asturias (HUCA). Informed consent was obtained from each patient, and the protocol was approved by the Institution and the Clinic Research Ethics Committee of the Health Institute of the Principality of Asturias, as instructed by the Declaration of Helsinki. Mononuclear cells were isolated by Ficoll density-gradient centrifugation and cultured as described for AML cell lines.
Metabolic stability
In vitro incubation with human liver microsomes was performed to evaluate the disappearance of the parent compound (EC-70124) by LC-MS. Half-life is calculated by the equations: T1/2 = 0.693/k, where k is the rate constant. The assay included a control without cofactors at 60 minutes to account for the disappearance of the drug by reasons unrelated to cytochrome activity.
For the in vivo assessment of metabolic stability, CD1 mice were dosed at 24 mg/kg intravenously and their plasma analyzed after 3 hours by HPLC-UV to detect indolocarbazole-related metabolites.
Protein plasma binding
Unbound fraction of EC-70124 and midostaurin upon binding to plasma proteins were determined measured by commercially available TRANSILXL High Sensitivity Binding Kit (Sovicell GmbH) per the manufacturer's instructions. Membrane affinity (MA) of each molecule was previously assessed using the also commercially available TRANSILXL Membrane Affinity Kit to estimate the optimal plasma dilutions and kit suitability. Analysis of drug concentrations was carried out by LC/MS-MS in MRM mode.
Oral bioavailability by Caco-2 assay
Oral bioavailability was measured by bidirectional permeability through Caco-2 cell monolayers (Absorption Systems). The assay determines the permeability through Caco-2 cell monolayers in both the apical-to-basolateral and basolateral-to-apical direction. EC-70124 concentration was 5 μmol/L. Samples were taken from the donor and receiver chambers at 120 minutes and assayed by LC/MS-MS using electrospray ionization to calculate the apparent permeability, Papp, and percentage of recovery.
Kinase-binding assays
The binding affinity of EC-70124 and midostaurin were determined by quantitative-binding affinity of compound–kinase interactions (KdELECT, Discoverex) using a 10-point dose–response curve.
Cell viability assay
For dose–response curves, cells were platted on 96-well dish at a density of 5,000 cells/well and viability was determined by the MTT assay as previously described (24). Data were analyzed using CalcuSyn software (Biosoft) that calculates the IC50 value for each drug.
Annexin V binding
Cells were seeded in 6-well plates at a density of 2 × 105 cells/well. Once the treatments were completed, apoptosis was evaluated by the Annexin V-FICT Apoptosis Detection Kit (Sigma Chemical Co.) as per the manufacturer's protocol. Apoptosis level (Annexin V–positive cells) was determined in 10,000 cells per group using a Beckman Coulter FC500 flow cytometer (Beckton Dickinson).
Flow-cytometry analysis of cell-cycle distribution
After treatment, harvested cells were incubated with a ribonuclease A (100 μg/mL) solution for 10 minutes. Then, samples were stained with propidium iodide (0.005%) for 10 minutes in the dark. A Beckman Coulter FC500 flow cytometer (Beckton Dickinson) was used for counting 2 × 104 cells per sample. Analysis of cell-cycle distribution was carried out using the software Modfit Ver 5.2 (Verity Software House).
Real-time quantitative PCR
Quantitative analysis of CYCLIN D1 (CCND1), PIM1, and C-MYC mRNA levels was performed by the SYBR Green real-time PCR method using Green PCR Core Reagents (Applied Biosystems) in an AB7700 Real-Time System (Applied Biosystems) as described previously (25). The primers used were the following: 5′-GCTGCGAAGTGGAAACCATC-3′ (sense) and 5′-CCTCCTTCTGCACACATTTGA-3′ (antisense) for CCND1; 5′-CGAGCATGACGAAGAGATCAT-3′ (sense) and 5′-TCGAAGGTTGGCCTATCTGA-3′ (antisense) for PIM1; 5′-TGCTCCATGAGGAGACACC-3′ (sense), 5′-CTTTTCCACAGAAACAACATCG-3′ (antisense) for C-MYC and 5′-TTCCCCATGGTGTCTGAGC-3′ (sense) and 5′-ATCTTCTTTTGCGTCGCCAG-3′ (antisense) for GADPH. Each sample was tested in triplicate, and relative gene expression data were analyzed by means of the 2−ΔCt method.
Western blot analysis
Cells were lysed with ice-cold lysis buffer (150 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1%v/v Triton X-100, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L β-glycerophosphate, 1 mmol/L Na3VO4, 1 μg/mL leupeptin, 10 nmol/L NaF, 1 mmol/L PMSF, 20 mmol/L Tris-HCl pH 7.5). Between 30 and 50 μg of total protein were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Amersham Bioscience). Primary antibodies were used against the phosphorylated forms of STAT5, p70S6k, S6, 4-EBP, BAD (1:1,000, Cell Signaling Technology), total PIM1, PIM2 (1:1,000, Cell Signaling Technology) and GAPDH—as a loading control (1:1,000, Santa Cruz Biotechnology). For signaling detection appropriated secondary (anti-rabbit IgG peroxidase–conjugated and anti-goat IgG peroxidase 1:3,000, Calbiochem) antibodies were used and the reaction was visualized by means of enhanced-chemiluminescence detection reagents (Amersham Biosciences) following the manufacturer's protocol.
In vivo xenograft model
Female immunodeficient mice were purchased from Janvier Laboratories (CB17 SCID, 5-weeks-old) and maintained under sterile and controlled conditions with food and water ad libitum. All animal research protocols were approved by the Animal Research Ethical Committee of the University of Oviedo. MV4-11 cells (5 × 106 cells with 25% Matrigel) were injected subcutaneously into the right flank. Once tumor size reached 200 mm3 animals were randomized in the different experimental groups of 8 mice, unless otherwise noted. Tumor volume was measured daily, and drug efficacy was expressed as the percentage tumor growth inhibition (%TGI).
For pharmacodynamic studies, immunodeficient mice were treated orally once the mean tumor volume was 400 mm3. Tumors were dissociated into single cell suspensions using a MACS Tissue Dissociation Kit and the GentleMACS Dissociator system (Miltenyi Biotec) for protein, RNA isolation and drug accumulation analysis. Blood samples were also collected from same mice via cardiac puncture and EC-70124 concentration was analyzed by LC/MS.
Statistical analysis
Experiments were repeated at least three times, and data were calculated as the average ± SE. One-way ANOVA followed by a Student Newman–Keuls multiple range test were carried out and statistical significance was accepted when P < 0.05.
Results
EC-70124 is a potent and selective inhibitor for AML kinases
Structures of EC-70124 and midostaurin are shown in Supplementary Fig. S1. Competition binding assays has revealed that this compound was highly active against several human kinases including some AML related kinases such us FLT3, JAK or PIM kinases (19–22). Here, we evaluated kinase inhibitor activity at several EC-70124 concentrations to evaluate the affinity to the targets. Data were compared with those of midostaurin from the same assay platform (KdELECT, Discoverex). Results indicated that EC-70124 potently inhibits several AML related kinases, including FLT3, JAK, SYK or PIM kinases (Kd < 10 nmol/L, Table 1). Importantly, 7 out of 9 FLT3 mutants tested are inhibited at even lower concentrations than wild-type FLT3.
Comparison of the EC-70124 and midostaurin dissociation constants for selected kinases
Moreover, EC-70124 shows higher potency than midostaurin for several AML-related kinases. Thus, EC-70124 is more potent than midostaurin for FLT3 and its mutants, 10x more potent for SYK, 10–70x more potent for JAK kinases and 2–3 orders of magnitude more potent for PIM kinases (Table 1). Interestingly, EC-70124 is less potent than midostaurin for KIT, a kinase regarded as anti-target, due to its involvement in the development of undesired myelosuppression in AML treatment (26).
EC-70124 inhibits AML cell-line growth
EC-70124 was tested for antiproliferative activity in a panel of 8 AML cell lines (Fig. 1A). All cell lines were sensitive to EC-70124 with IC50 values ranging from 13 to 520 nmol/L (Fig. 1A and B). FLT3-ITD cell lines MV4-11 and MOLM-13 show the lowest IC50 values whereas wild-type cells show the biggest ones. In agreement with the superior profile as inhibitor of AML-related kinases, EC-70124 also displayed greater inhibitory effect on FLT3wt cell lines compared to midostaurin, which was particularly active only in FLT3-ITD cell lines (Fig. 1B and C).
Effect of EC-70124 and midostaurin on cell viability in AML cell lines and patient samples. Eight AML cell lines were treated with EC-70124 (0–10 μmol/L; A) or midostaurin (0–10 μmol/L; B) for 48 hours, and viability was determined by MTT reduction. C, Results were analyzed using CalcuSyn software to calculate IC50 values for each drug THP-1 cell line showed a plateau effect, and therefore the IC50 value could not be determined for midostaurin. FLT3 status is indicated for each cell line. D, Western blots of cell-free extracts of the 8 cell lines showing protein levels for phosphorylated STAT5 and p70S6K, pBAD, and PIM kinases, and GAPDH as a control. E, Cell viability determined after EC-70124 (0–100 nmol/L) treatment for 48 hours of isolated blast from 6 AML patient samples. FLT3 status is indicated for each sample. F, Cell death in AML patient samples determined by Trypan blue staining after EC-70124 treatment for 48 hours. FLT3 status is indicated for each sample. *, P < 0.01 versus control group (vehicle-treated cells). ND, not determined.
Sensitivity to EC-70124 can be related with expression of AML-related kinase-targets (Fig. 1D). Phosphorylated STAT5 is expressed in the five of the eight most sensitive cell lines. On the other hand, the three least sensitive cell lines (NB-4, THP-1, and HL-60) did not express neither phosphorylated STAT5 nor any PIM isoforms.
EC-70124 was also effective in mononuclear cells obtained from bone marrow aspirates from newly diagnoses AML patients. As showed in Fig. 1E, 5 out of 6 samples tested were sensitive to nanomolar concentrations of EC-70124 in a dose-dependent way. In fact, EC-70124 treatment results in an increase in cell death as determined by trypan blue exclusion assay (Fig. 1F). Interestingly, best results were observed in FLT3 ITD samples, which agree with the results obtained in AML cell lines and could be related to the fact that FLT3 mutants are inhibited at even lower concentrations than wild-type FLT3.
EC-70124 can induce cell-cycle arrest and apoptosis
Cell-cycle analyses in MV4-11, MOLM-13, and MOLM-16 (treated at concentrations near the IC50 value for each line) showed dose-dependent increase in G0/G1 population, indicating cell-cycle arrest in response to EC-70124, especially in the two FLT3-mutant cell lines (Fig. 2A). Moreover, an increase in apoptotic response (sub-G1 population) was observed in the three cell lines (Fig. 2B), although it was considerably higher in MOLM-16 cells. On the contrary, less sensitive HL-60 cells hardly showed variations neither in the cell cycle nor in sub-G1 population even at high concentrations (Fig. 2A and B).
Effect of EC-70124 on cell cycle and apoptosis in MV4-11, MOLM-13, MOLM-16, and HL-60 cell lines. A, Cell-cycle distribution after 24 or 48 hours treatment as determined by propidium iodide staining. B, Percentage of cell at sub-G1 phase after EC-70124 treatment as indicator of apoptosis cell death. C, Effect of 24 to 48 hours treatment with EC-70124 on AML cell lines apoptosis determined by Annexin V staining. A midostaurin-treated group (50 nmol/L for MV4-11, 50 nmol/L for MOLM-13, 50 nmol/L for MOLM-16, and 1,000 nmol/L for HL60) has been represented for comparison. *, P < 0.01 versus control group (vehicle-treated cells).
Consistent with the increase in sub-G1 population, dose-dependent increase in apoptotic cells was also observed in MV4-11, MOLM-13, and MOLM-16, as determined by Annexin V–binding assay. According to the cell–cycle distribution data, induction of apoptosis is much higher on MOLM-16 cells. Apoptotic cells were barely detected in less sensitive HL-60 cells after treatment with EC-70124. Midostaurin, according to the previously described selectivity for FLT3-ITD mutant AML cells, only induces apoptosis in MV4-11 and MOLM-13 cells (Fig. 2C).
EC-70124 modulates AML-related kinases
Abnormalities in receptor tyrosine kinases or receptor-associated intracellular kinases are extremely frequent in cancer, including AML (27). Thus, FLT3 alterations occurs in about 30% samples and are associated with poor outcome (3). Receptor activation results in the induction of downstream prosurvival pathways including STAT5 and Akt/mTOR. Modulation of several proteins of these pathways was measured after EC-70124 treatment. Low concentrations of EC-70124 results in rapid inhibition of Akt/mTOR pathway—upon few hours of treatment—as determined by the phosphorylation status of the downstream targets p70S6K and its substrate S6 ribosomal protein and also p4EBP1 (Fig. 3A). High EC-70124 concentrations were needed in less sensitive HL-60 cells to obtain similar results in agreement to the IC50 data.
Effect of EC-70124 on FLT3 downstream signaling in vitro. A, Phosphorylation levels of key targets upon treatment with two concentrations of EC-70124 (based on the IC50) at 2 time points in MV4-11, MOLM-13, MOLM-16, and HL-60 cell lines. B, Phosphorylation levels of key targets upon treatment with two concentrations of midostaurin (based on the IC50) at 8 hours after treatment in MV4-11, MOLM-13, MOLM-16, and HL-60 cell lines. C, mRNA levels of STAT5 target genes in MV4-11, MOLM-13, MOLM-16, and HL-60 cell lines determined by RT-qPCR. Dashed line represents basal level of expression. *, P < 0.01 versus control group (vehicle-treated cells).
As mentioned above, STAT5 appears to be one of the main targets in constitutively activated FLT3 cell lines (28). In this sense, we found EC-70124 to induce a rapid decrease in phosphorylated STAT5 in MV4-11 and MOLM-13 cells (Fig. 3A). No inhibition was found in MOLM-16 cells or less sensitive HL-60 cells (no basal STAT5 phosphorylation detected, Fig. 1D). On the other hand, we found a decrease in phospho-BAD in all cell lines treated with EC-70124 (Fig. 3A). Several intracellular pathways can inhibit the pro-apoptotic activity of BAD via protein phosphorylation including PI3K/Akt and also PIM kinases. In fact, it has been well described that PIM1 kinase promote AML cell survival via phosphorylation of BAD at Ser 112. Thus, according to low binding assay Kd value for PIM1 kinase (Table 1), proapoptotic effect of EC-70124 could be related to the inhibition of PIM1 in cells with basal expression, but not in the PIM1-lacking HL-60 cell line (Fig. 1D). Because high concentrations are needed for cell viability decrease in HL-60 we cannot exclude that other intracellular pathways may be altered.
We found similar effects in phosphorylated STAT5 and Akt/mTOR targets in FLT3-ITD mutant AML cells treated with midostaurin (Fig. 3B). Interestingly, we did not found a decrease in phospho-BAD in these cells after midostaurin treatment. This supports our hypothesis that the inhibition of BAD can be mediated by inhibition of PIM1 since misdostaurin lacks activity on PIM1.
We also treated the AML cells with SGI-1776, a well-known FLT3-PIM dual inhibitor, and found similar effects on cell viability (Supplementary Fig. S2A), cell cycle (Supplementary Fig. S2B) and apoptosis (Supplementary Fig. S2C). Moreover, SGI-1776 treatment also resulted in inhibition of phosphorylation of same key targets than EC-70124 (Supplementary Fig. S2D).
In agreement to the inhibition of STAT5 phosphorylation and activation, we also showed a dose-dependent decrease in the mRNA expression levels of STAT5 targets as cyclin D1, c-myc and PIM1 in MOLM-13 and MV4-11 cells (Fig. 3C). No changes in STAT5 targets were found in MOLM-16 cells where EC-70124 did not induce inhibition of STAT5 phosphorylation or in HL-60 cells that did not express basal STAT5 phosphorylation (Fig. 3C).
EC-70124 is orally bioavailable, metabolically stable, and shows lower protein plasma binding than midostaurin
Permeation through Caco-2 monolayers has been established as a valid preclinical model to correlate the oral bioavailability potential of drugs in preclinical development (29). As showed in Table 2, we determined the permeability (Papp) of EC-70124 through Caco-2 monolayers in both the apical-to-basolateral (A–B) and basolateral-to-apical (B–A) directions. The resulting Papp A–B, Papp B–A and efflux ratio, indicates a high potential for oral absorption to EC-70124, without significant drug efflux.
Oral absorption potential by the Caco-2 assay for EC-70124
Incubation with human liver microsomes was carried out to assess half-life and intrinsic clearance (Clint). The results obtained, half-life >60 minutes and Clint < 10 mL/min/kg, indicate an elevated metabolic stability (Supplementary Fig. S3A). Also, incubation without cofactors (to address disappearance unrelated to cytochrome activity) show no changes in EC-70124 concentration. In addition, an in vivo experiment provided further metabolic stability data. Mice were dosed with EC-70124 (24 mg/kg, i.v.) and its plasma analyzed 3 hours after dosing by HPLC-UV. The HPLC-UV profile shows only one peak, corresponding to the non-metabolized drug (see Supplementary Fig. S3B for a representative example).
We determine the unbound fraction (fu) of EC-70124 and midostaurin upon binding to human plasma proteins using a high sensitivity binding assay. As it can be seen in Table 3, data obtained for EC-70124 shows a 3.9% fu, as compared with 1.1% fu for midostaurin in human plasma. The human plasma data for EC-70124 is in line with the fu obtained in rat and dog plasma (4.2% and 3.0%, respectively).
Protein plasma binding for EC-70124 and midostaurin
In agreement with binding assay results, co-incubation with human alpha-acid glycoprotein (which represents 1%–3% of plasma proteins) induced a significant increase in IC50 for midostaurin compared with EC-70124 (Supplementary Fig. S3C).
EC-70124 efficacy and pharmacodynamics in AML xenografts
In vivo activity of EC-70124 was evaluated in a MV4-11 xenograft model. Inhibition of tumor growth was observed using both intravenous and p.o. administration routes although oral administration results in a more sustained cytostatic effect (Fig. 4A, and Supplementary Fig. S4A). Both administration routes were well tolerated by all mice because no significant weight loss was detected (Supplementary Fig. S5A).
EC-70124 in vivo efficacy and pharmacodynamic analysis in AML xenograft models. A, CB-17 SCID mice were implanted subcutaneously with MV4-11 cells and treated with either vehicle or EC-70124 by oral gavage (po) daily or intravenous injection every 3 days to assess tumor growth. Forty mg/kg daily po doses for 2 weeks, followed by 20 mg/kg daily oral doses for another 2 weeks (20 doses total), were compared with 18 mg/kg i.v. doses every 3 days (8 doses total). *, Significant differences between treated and control groups (P < 0.01, t test). #, Significant differences between intravenous and po groups (P < 0.01, t test). B, Tumor growth determined in mice treated by oral gavage with two different schedules and dose (20 mg/kg daily vs. 80 mg/kg every other day). *, Significant differences between treated and control groups (P < 0.01, t test). #, Significant differences between 20 mg/kg and 80 mg/kg groups (P < 0.01, t test). C, Comparison of EC-70124 versus midostaurin effect on tumor volume using an equivalent total dose under different schedules: 13 doses of 80 mg/kg every other day for EC-70124 (1,040 mg/kg total) and 10 doses of 100 mg/kg daily for midostaurin (1,000 mg/kg total). Because tumors grew back in both treatment groups, mice were retreated once tumors reached 500 mm3 using same dose and schedule as in the first treatment cycle (treatment windows are represented on the bottom of the figure). D, Oral pharmacokinetic profile of EC-70124 obtained by monitoring plasma levels upon a single 80 mg/kg dose. Each data represented the average of 3 mice. E, Phosphorylation levels of key targets in MV4-11 xenograft model at different time points. Mice were implanted as above and treated orally once with 80 mg/kg of EC-70124 when the mean tumor volume was 400 mm3. Tumors were dissociated into single cell suspensions and harvested for immunoblotting and RNA at eight different time points between 15 minutes and 48 hours. Each time point represented the pool of 3 mice. F, mRNA levels of STAT5 target genes determined by RT-qPCR in MV4-11 tumor samples. Dashed line represents basal level of expression. *, P < 0.01 versus control group (vehicle-treated cells).
Oral administration efficacy was found to be dose dependent, as daily treatment at 20 mg/kg of EC-70124 led to a TGI of 91% after 28 days, whereas 80 mg/kg every other day resulted in complete regression of tumor growth in all mice before the end of treatment (Fig. 4B, and Supplementary Fig. S4B), and without apparent toxicity as judged by monitoring mice weight (Supplementary Fig. S5B). None of the high dose mice had reached 2,000 mm3 at the time all the mice in the low dose hit that mark (Supplementary Fig. S6A). It is of note that one mice from the high dose group was permanently cured (tumor free for >6 months).
We also compared EC-70124 with midostaurin, using an equivalent total dose under different schedules. In both cases, complete regression was observed; however, midostaurin-treated mice reached tumor suppression faster (day 20) than EC-70124 (day 35; Fig. 4C). Once fully regressed, 7/8 tumors grew back about 1 week later in both treatment groups. It is of note that 1/8 were permanently cured in each drug (mice tumor free for >6 months). Once tumors reached 500 mm3, mice were treated again with the corresponding drug at the same dose and schedule as in the first treatment cycle. Despite the high tumor volume, regression was again faster with daily midostaurin (−100% regression), while the EC-70124 group fell short of complete regression (−70% regression). Both drugs were well tolerated with no mice weight loss observed (Supplementary Fig. S5C). The longer schedule used for EC-70124, a positive sign of non-accumulative toxicity, is likely to translate into a survival benefit (Supplementary Fig. S6B).
EC-70124 and midostaurin were also evaluated in a MOLM-16 xenograft. Treatment with midostaurin (100 mg/kg daily) has no significant effect on tumor growth (45% TGI), whereas EC-70124 (80 mg/kg every other day) resulted in 95% TGI at day 32 (Supplementary Fig. S7A). Individual mice tumor volume plot at day 34 also shows better response of EC-70124 respect to midostaurin in the MOLM-16 xenograft model (Supplementary Fig. S7B).
Oral pharmacokinetic profile of EC-70124 was obtained by monitoring plasma levels upon a single 80 mg/kg dose. Maximum concentration in plasma was detected after 2 hours, reaching a peak that declines slowly over 48 hours, when plasma circulating levels dropped to 100 nmol/L. Drug analysis in tumor tissue shows much higher accumulation in MV4-11 tumor cells than in plasma (almost 10 times higher levels after 2 hours administration), and even at 48 hours concentration was high enough to maintain target inhibition (Fig. 4D). In agreement with results in cultured cells, a rapid inhibition of STAT5, 4EBP1, S6 and BAD phosphorylation was found in tumor tissue, indicating that same pathways identified in vitro are mediating the effects in vivo (Fig. 4E). Recovery profile of the proteins analyzed showed that the inhibition of STAT5, 4EBP1, and BAD was sustained for at least 48 hours after dosing, being S6 protein the only target that shows faster recovery after treatment. In addition, pharmacodynamic RNA expression analysis shows a decrease in expression levels of STAT5 target genes CCND1, C-MYC and PIM1, and their recovery profile rendered C-MYC as the only repressed gene at 48 hours, indicating a correlation with STAT5 inhibition and recovery profile (Fig. 4F).
Discussion
Several kinase inhibitors have been tested for AML treatment because deregulated tyrosine kinase activity has been implicated in pathogenesis of hematologic malignancies (2). EC-70124 is an indolocarbazole natural product, with in vitro antiproliferative activity in solid tumors at IC50 values >100 nmol/L [20–23]. Here we report antitumor activity of EC-70124 in AML. In a panel of 8 AML cell lines, 5 of them displayed IC50 < 100 nmol/L, indicating this hematological cancer could be a suitable indication for EC-70124, further supported by its antitumor activity in AML subcutaneous xenograft models, and in patient-derived AML primary cells.
Kinase affinity assays revealed that EC-70124 presents an excellent potential for AML, because its inhibition profile includes targets like FLT3—the most frequent mutated target in AML—both wild type, ITD, and kinase domain mutants, that have been related to kinase inhibitors resistance (30). However, despite promising preclinical results, FLT3 inhibitors have failed to meet expectations in advanced clinical trials (3) and novel inhibitors targeting not only FLT3 but also other pathways contributing to disease progression are needed. This perception is likely to become more attractive on the wake of the successful RATIFY trial (and subsequent FDA approval) that demonstrates that the dirty inhibitor midostaurin prolongs survival in a subset of AML patients (6). In this line, EC-70124, unlike midostaurin, is a potent inhibitor of other important AML related kinases such as PIM kinases. Moreover, EC-70124, is 100 times more potent against FLT3 than against KIT, responsible of undesired myelosuppression in patients (26). Ideally, FLT3 inhibitors should spare C-KIT activity, based on the evidence that FLT3 and C-KIT knocked out mice develop severe hematopoietic deficiencies (31).
In agreement to its kinase inhibition profile, EC-70124 exerts antitumor effects in several AML cell lines. Potent antiproliferative effect has been found in FLT3-mutated cells that also express high levels of PIM kinases, whereas modest results have been found in FLT3 wild-type cells that do not express PIM kinases. Comparable results were also observed in AML blasts derived from patients and treated with the drug, with best results in FLT3-mutated blast than FLT3 wild type blast. According to the kinase inhibition data, EC-70124 abrogation of PIM kinase activity in FLT3-ITD could happen not only at the transcriptional level through STAT5 but also by inhibiting PIM1 kinase itself, an important fact on the basis of the relation between PIM activity and resistance development. Actually, this fact represents the basis of the design of dual inhibitors as SGI-1776 (32), and more recently SEL24 (33). Moreover, patients often develop secondary point mutations on the FLT3 receptor at the residue D835, leading to TKI resistance. Both midostaurin and EC-70124 inhibit those TK mutations, but only EC-70124 inhibits JAK2 or SYK kinases, that have been implicated in other TKI resistance mechanism (34, 35).
Treatment with EC-70124 rapidly inhibits FLT3 downstream signaling pathways, inhibiting phosphorylation of STAT5, p70S6k and its substrate S6. Inhibition of p70S6k and its ligand occurs in all cells, whereas STAT5 inhibition only occurs in FLT3 mutant cells. MOLM-16 presents the highest apoptosis induction, with no reduction on STAT5 phosphorylation. Recently, a novel TK gene fusion (ELAVL1-TYK2) that controls the activation of STAT5 has been described for this cell line (36), which could explain the lack of inhibition of STAT5 phosphorylation in this case.
EC-70124 also inhibits phosphorylation of pro-apoptotic BAD. Phosphorylation of BAD at Ser 112 could be mediated by several intracellular pathways such us Akt, P90RSK and PIM kinases. In fact, it has been well described that PIM kinases promote AML cell survival via phosphorylation of BAD and constitutively activated FLT3 phosphorylates BAD partially through PIM (37). According to kinase binding assays, effects of EC-70124 on BAD phosphorylation can be related both to the inhibitory effect on FLT3 and PIM kinase activity. Moreover, the lack of activity in BAD phosphorylation of midostaurin (that lacks activity over PIM1), together with its inhibition after treatment with SGI-1776 (dual FLT3-PIM1 inhibitor) reinforced our hypothesis about EC-70214 affecting both FLT3 and PIM1. Importantly, Pim expression can be upregulated by STAT5, indicating a close relationship between FLT3 and PIM. Moreover, Pim kinase overexpression has been described to reduce the efficacy of FLT3 inhibitors (16). For this reason, design of safe strategies for concomitant FLT3/Pim kinase inhibition has been recently proposed as a promising strategy in AML (17, 33)
In addition to BAD, Pim kinases share other substrates with the Akt/mTOR pathway, modulating protein translation through p70S6k and 4-EBP1 that represents another convergent point in the FLT3 and Pim kinases pathway (38) that could be exploited by EC-70124. For example, 4EBP1 phosphorylation is downmodulated by EC-70124 treatment in AML cells, as well as pBAD. In the HL-60 cell line, phosphorylation of BAD could be mediated by the Akt/mTOR pathway because there is no basal PIM1 expression, leading to the higher IC50 obtained for EC-70124.
EC-70124 not only presents antileukemic effects in cultured cell lines but also reduces tumor growth in subcutaneous xenograft models without signs of toxicity. The compound is more efficacious by oral frequent administration, which can induce complete regression of the tumor even before the end of the treatment, than spaced intravenous injection.
Comparison between EC-70124 and midostaurin showed similar response for both drugs in the MV-4-11 xenograft model. However, EC-70124 demonstrates a higher efficacy than midostaurin slowing tumor growth in the FLT3wt MOLM-16 xenograft model, evidencing the benefits of dual PIM-FLT3 inhibition by EC-70124 beyond the FLT3-ITD tumor models. This in vivo result is in agreement to the in vitro IC50 values and PIM1 Kd for both drugs. The fact that Pim knockouts in mice show minimal phenotype changes highlights the suitability of PIM as a clean target for therapy (39). The more spaced schedule of EC-70124 allows longer progression-free periods as compared with equivalent doses of midostaurin.
Preclinical pharmacology studies demonstrate that EC-70124 seems stable, without significant degradation or appearance of new species. Instead, several metabolites are described for midostaurin (40). Moreover, protein plasma binding—typically very high for kinase inhibitors and cited often as a potential source of clinical failure (41)—shows higher free unbound fraction in EC-70124 than midostaurin, a potential benefit for the clinical translation of the drug. In this sense, presence of alpha-acid glycoprotein, known to bind midostaurin (42), caused a moderate reduction of the EC-70124 activity, as compared with midostaurin. Finally, data obtained showed also a high potential for human gut absorption according to the Caco-2 assay.
In summary, the present report describes the antileukemic activity and preclinical characterization of a novel multikinase inhibitor, EC-70124, based on its ability to interfere the complex oncogenic events activated in AML at several levels. Because treatments directed against single targets such as FLT3 have not demonstrated the expected effectiveness, the unique kinase inhibitor profile exhibited by EC-70124 makes it an ideal candidate for the treatment of AML, and future studies are warranted to evaluate its clinical efficacy.
Disclosure of Potential Conflicts of Interest
F. Moris has ownership interest (including patents) in EntreChem S.L. No potential conflicts of interest were disclosed by the other authors.
Authors' Contributions
Conception and design: P. Costales, C. Rodríguez, V. Martín, F. Morís
Development of methodology: N. Puente-Moncada, P. Costales, L.-E. Núñez, P. Oro, M.A. Hermosilla, N. Ríos-Lombardía, V. Martín
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): N. Puente-Moncada, P. Costales, I. Antolín, L.-E. Núñez, P. Oro, M.A. Hermosilla, J. Pérez-Escuredo, N. Ríos-Lombardía, A.M. Sanchez-Sanchez, E. Luño
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N. Puente-Moncada, P. Costales, I. Antolín, J. Pérez-Escuredo, A.M. Sanchez-Sanchez, C. Rodríguez, V. Martín
Writing, review, and/or revision of the manuscript: P. Costales, L.-E. Núñez, P. Oro, M.A. Hermosilla, J. Pérez-Escuredo, N. Ríos-Lombardía, E. Luño, C. Rodríguez, V. Martín, F. Morís
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P. Costales, F. Morís
Study supervision: P. Costales, F. Morís
Acknowledgments
EntreChem thanks Dr. Atanasio Pandiella for valuable input during the course of the project. P. Costales, J. Pérez-Escuredo, and N. Ríos-Lombardía were supported by Torres-Quevedo program from Ministry of Science and Innovation (Spain; references PTQ-11-04-507, PTQ-13-06-368, PTQ-12-05-407, respectively). C. Rodriguez was supported by Ministry of Science and Innovation (SAF2014-58468-R) and FICYT (GRUPIN14-081). N. Puente-Moncada was supported by an FICYT fellowship (BP13-108).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Footnotes
Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).
- Received June 19, 2017.
- Revision received October 19, 2017.
- Accepted December 15, 2017.
- ©2018 American Association for Cancer Research.