Preclinical efficacy for a novel tyrosine kinase inhibitor, ArQule 531 against acute myeloid leukemia

ARQ 531 was synthesized and provided by ArQule, Inc., (Burlington, MA). Quizartinib, midostaurin, and dasatinib were purchased from LC Laboratories (Woburn, MA). Gilteritinib and venetoclax were purchased from Chemietek (Indianapolis, IN). Entospletinib (GS-9973) was purchased from Selleckchem (Houston, TX). All chemical compounds were dissolved in Dimethyl sulfoxide (DMSO) to obtain a 10 mM stock solution.

MOLM-13, MV4-11, and OCI-AML3 AML cell lines were purchased from the Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures (Germany). THP-1 and U937 AML cell lines were purchased from the American Type Culture Collection (Manassas, VA). HS-5 GFP were obtained from Dr. William Dalton (H. Lee Moffitt Cancer Center & Research Institute).

All cell lines were cultured in Roswell Park Memorial Institute 1640 medium (Life Technologies, Carlsbad, CA) supplemented with 10% fetal bovine serum (VWR Life Science Seradigm, Radnor, PA), 2 mM L-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin (Life Technologies, Carlsbad, CA). The luciferase expressing MOLM-13 cell line was a generous gift from Dr. Ramiro Garzon [38]. Drug-resistant MOLM-13 cells (D835 mutation) and Ba/F3 murine Pro-B cell lines expressing FLT3-ITD or FLT3-TKD or FLT3-ITD-TKD were a generous gift from Dr. Sharyn Baker (College of Pharmacy, OSU) and cultured as previously described in Zimmerman et al. [39]. All cell lines were routinely tested for mycoplasma contamination and authenticated using short tandem repeat (STR) profiling.

The pMSCV-IRES-Tomato (pMIT-Empty) and pMIT-SYK-TEL vectors were a generous gift from Dr. Kimberly Stegmaier (Dana-Farber Cancer Institute) and previously described in Puissant et al. [30]. HEK293T were transfected with the pMIT vectors and packaging vectors, VSV and GP2. The resulting supernatant was harvested and concentrated for spin infection of Ba/F3 expressing FLT3-ITD (Ba/F3-FLT3-ITD). After recovery, cells were sorted using FACS Aria III flow cytometer.

AML patient samples were obtained from The Ohio State University Comprehensive Cancer Center Leukemia Tissue Bank under an Institutional Review Board-approved protocol.

For AML cell lines and primary patient samples, genomic DNA was extracted using the DNeasy Blood and Tissue Kit (QIAGEN, Hilden, Germany). The mutational status of 91 protein-coding genes was determined using Illumina TruSeq Custom Amplicon library preparation. Libraries were pooled and sequenced using the MiSeq platform (Illumina, San Diego, CA). DNA library preparations were performed according to the manufacturer’s instructions. Illumina amplicon sequenced reads were aligned to the hg19 genome build using the Illumina Isis Banded Smith-Waterman aligner. Single nucleotide variant and indel calling were performed using MuTect and VarScan, respectively. The MuCor algorithm was used as the baseline for integrative mutation assessment. We only considered non-synonymous variants not listed in either the 1000 Genome database or dbSNP142-common variants. After variant calling, variants were annotated using SnpEff and vcfanno along with the dbsnp, COSMIC, 1000 genomes, and 6500 exomes variant databases. The Mucor3 algorithm was used as the baseline for integrative mutation assessment. All called variants underwent visual inspection of the aligned reads using the Integrative Genomics Viewer.31. Variants in regions of high discrepancy, low quality, tandem repeats, or mononucleotide runs were excluded.

Antibodies purchased from Cell Signaling Technology (CST, Danvers, MA) including anti-FLT3 (Cat.#3462), anti-phospho-FLT3 Tyr589/591 (Cat.#3464), anti-STAT5 (Cat.#9363), anti-phospho-STAT5 (Cat.#9351), anti-ERK (Cat.#4695), anti-phospho-ERK (Cat.#4377), anti-Src family kinase (Cat.#2109), anti-phospho-Src family kinases Tyr416 (Cat.#2101), anti-SYK (Cat.#13198), anti-phospho-SYK Tyr525/526 (Cat.#2711), anti-BTK (Cat.#8547), anti-phospho-BTK Tyr223 (Cat.# 87457), anti-MCL1 (Cat.# 94296), anti-BCLXL (Cat.# 2764), and anti-HSP90 (Cat.#4874). Anti-β-Actin (Cat.# SC-1616) was purchased from Santa Cruz Biotechnology (Dallas, TX) and anti-GAPDH (Cat.# MAB374) from EMD Millipore (Burlington, MA). HSP90, β-Actin, and GAPDH were used as loading controls.

Cell lines were serum starved overnight followed by a 10-min treatment. Cells were washed in ice-cold PBS then lysed in 1X Cell Lysis Buffer (CST) supplemented with proteinase and phosphatase inhibitors cocktail (Roche). Pierce BCA Protein Assay was used to quantify the protein lysates (Thermo Fisher Scientific, Waltham, MA). For Immunoblotting, 15–25 μg of protein lysate was prepared in 2X or 4X Laemmeli’s sample buffer (Bio-Rad Laboratories, Hercules, CA) and heated for 5 min at 100 °C. Samples were separated on 8% sodium dodecyl sulfate polyacrylamide gel and transferred onto nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA). The membranes were blocked for 1 h in Blocker BLOTTO Blocking Buffer (Thermo Fisher Scientific, Waltham, MA) or 5% non-fat milk followed by overnight incubation at 4 °C with primary antibody followed by HRP-conjugated secondary antibodies and visualized with chemiluminescent Pierce ECL Substrate (Thermo Fisher Scientific, Waltham, MA) on X-ray films.

Biochemical inhibition of FLT3-WT and FLT3-ITD fusion was measured using recombinant protein constructs of kinase domains (Reaction Biology). ARQ 531 was tested in a 10-point concentration mode with 3-fold serial dilutions starting at 1 μmol/L in the presence of 10 μmol/L ATP, and the IC₅₀ was determined.

A total of 5000–30,000 cells from the lines above were treated in a 96-well plate for 72 h. CellTiter 96 Aqueous MTS reagent (Promega, Madison, WI) was added to cells followed by a 3-h incubation at 37 °C, 5% CO2. The absorbance was measured at 490 nm using DTX 880 plate reader and the half maximal inhibitory concentration (IC₅₀) was calculated using GraphPad Prism 7 (GraphPad Software, La Jolla, CA) and SAS/STAT software (SAS v9.4 for Windows, SAS Institute, Inc., Cary, NC). For the MOLM-13 resistant and the Ba/F3-FLT3-ITD/TKD cell lines, cells were treated for 72 h and cell viability was evaluated using MTT assay (Roche Diagnostics. Mannheim. Germany) according to the manufacturer’s instructions.

Cells were treated for 72 h then stained for 20 min in 1X Annexin Binding buffer (BD Biosciences, Franklin Lakes, NJ) containing Annexin V-FITC and Propidium Iodine (Leinco Technologies, Fenton, MO). Live and apoptotic cells were measured using FC 500 flow cytometer and analyzed on Kaluza software (Beckman Coulter, Pasadena, CA).

Primary AML patient samples were treated and plated at 5000–10,000 cells in 1.1 ml methylcellulose supplemented with recombinant human stem cell factor, interleukin 3, granulocyte colony-stimulating factor, and granulocyte-macrophage colony-stimulating factor (MethoCult™, STEMCELL Technologies, Vancouver, Canada) per dish in duplicates. Colonies were counted after 14 days when using primary patient samples.

All experiments were conducted after approval of The Ohio State University Institutional Animal Care and Use Committee. Mice were randomized and enrolled for daily oral gavage (Q.D.P.O) treatment of vehicle or 50 mg/kg ARQ 531. Mice were treated until reaching early removal criteria (ERC) defined as 20% weight loss, lethargy, hind limb paralysis, moribund, anorexia, dehydration, hunched posture, severe diarrhea, breathing difficulties, or any abnormal neurological symptoms.

For in vitro cytotoxicity experiments (Fig. 1b–d), repeated measures models were used to account for correlations among observations from the same experiment. The percentage of live cells was compared at each concentration with DMSO. Estimated IC₅₀ with 95% confidence intervals (CI) were calculated using 4-parameter logistic models. Mixed effects models were used in experiments described in Additional file 1: Table S4 and Fig. 2c to account for correlations within a single patient. Outcome data were first log-transformed to stabilize variances and reduce skewness. For survival experiments (Figs. 5 and 6a), Kaplan-Meier curves were produced and group differences for Fig. 5 and assessed by the log-rank test and for Fig. 6a, p values have been adjusted for multiple comparisons using Holm’s procedure. p values from experiments with multiple endpoints were adjusted for multiple comparisons using either Holm’s procedure or Dunnett’s test (all comparisons against vehicle/DMSO). All analyses were performed using SAS/STAT software (SAS v9.4 for Windows, SAS Institute, Inc., Cary, NC).

To investigate the activity of ARQ 531 in AML, we tested its ability to inhibit cell proliferation of AML cell lines and primary patient samples. ARQ 531 demonstrated anti-proliferative activity in FLT-ITD-mutated cell lines (MOLM-13 and MV4-11) in addition to the NPM1c⁺, DNMT3A-mutated OCI-AML3 cell line at low μM concentrations, while U937 and THP-1 cell lines were resistant to concentrations up to 10 μM (Additional file 2: Table S1 and Fig. 1a). To assess the direct effect of ARQ 531 on cell survival, we quantified the levels of live (Annexin−PI−), apoptotic (Annexin+PI−), and dead cells (Annexin+PI+) after ARQ 531 treatment. ARQ 531 induced cell apoptosis in MOLM-13, MV4-11, and OCI-AML3 cell lines while gilteritinib was cytotoxic in the FLT3-ITD cell lines only (Fig. 1b–d). These data suggest efficacy for ARQ 531 in both FLT3-ITD and FLT3-WT cells.

As one of the main resistance mechanisms to FLT3 inhibitors is the acquisition of TKD mutations, we tested the anti-proliferative activity of ARQ 531 in the MOLM-13 TKI-resistant (MOLM13-Res) cell line previously described in Zimmerman et al. [39]. In this model, MOLM-13 cells have gained a D835Y mutation via continuous exposure to the FLT3 inhibitor tandutinib [39]. Compared with the parental MOLM-13 cells, ARQ 531 shows a modest shift in IC₅₀ (1.3 μM vs. 3.4 μM in MOLM-13 and MOLM13-Res cells, respectively; Additional file 3: Table S2 and Fig. 1e). Similarly, ARQ 531 retains anti-proliferative activity in the IL-3 independent Ba/F3 murine cell lines harboring either a FLT3-ITD mutation, a resistance conferring TKD mutation, or a combination of both (IC₅₀ ranging from 1.2 to 7.6 μM; Additional file 3: Table S2 and Fig. 1f). The same murine cell lines have been previously shown to have a more robust response to midostaurin compared with the more specific FLT3 inhibitor quizartinib [39, 40], suggesting that a multi-kinase inhibitor such as ARQ 531 is likely to be effective in quizartinib-resistant samples.

To verify these results in primary cells, AML patient blasts (Additional file 4: Table S3) were co-cultured with HS5-GFP stroma cells for 96 h. Similar to gilteritinib, ARQ 531 demonstrated cytotoxicity in FLT-ITD samples in addition to FLT3-WT (Fig. 2a, b and Additional file 1: Tables S4). Moreover, reduced colony formation was observed at clinically attainable levels of ARQ 531 in both FLT3-ITD and FLT3-WT samples primary AML blasts (Fig. 2c and Additional file 5: Table S5) suggesting this agent has anti-clonogenic activity independent of FLT3 status.

The equal cytotoxicity in FLT3-ITD and FLT3-WT cell lines suggests that the ARQ 531 mechanism of action is independent of FLT3 status. In vitro kinase profiling utilizing recombinant protein kinases revealed that ARQ 531 exhibited potent inhibition of SFK, including LCK, YES, HCK, LYNa, FGR, FYN, and FRK [36] in addition to BTK and FLT3 (WT and ITD; 82 nmol/L and 330 nmol/L, respectively). Thus, we tested the ability of ARQ 531 to downregulate the levels of phosphorylated SFK (Tyr416), phospho-FLT3 and downstream targets, phospho-STAT5, and phospho-ERK in AML cell lines including the FLT3-ITD MOLM-13 and MV4-11 and the FLT3-WT OCI-AML3 (Fig. 3). ARQ 531 inhibited the phosphorylated levels of SFK, FLT3, STAT5, and ERK at concentrations as low as 500 nM in the FLT3-ITD cell lines, MOLM-13 (Fig. 3a) and MV4-11 (Fig. 3b). This is consistent with other FLT3 inhibitors quizartinib (selective FLT3 inhibitor), gilteritinib (AXL/FLT3 inhibitor), and midostaurin (multi-kinase inhibitor), although the more selective inhibitors are effective against FLT3 at a lower concentration range (50–100 nM). Importantly, ARQ 531 potently inhibited the phosphorylation of Tyrosine 416 of the SFK in all three cell lines, MOLM-13, MV4-11, and OCI-AML3 (Fig. 3c) similar to the positive SFK inhibitor control, dasatinib, while the other FLT3 inhibitors had little to no effect on SFK phosphorylation (Fig. 3c) in the FLT3-WT OCI-AML3 cell line. This suggests a role for SFK modulation in the mechanism of ARQ 531 in AML cell lines, distinguishing it from other FLT3 or multi-kinase inhibitors. To confirm the functional inhibition of SFK by ARQ 531, we measured the level of phosphorylated SYK, as well as an integral downstream target of the SFK member, LYN. SYK is indispensable for myeloproliferative disease development and its transformation to AML [30] and a critical regulator of FLT3-ITD driven neoplasia due to its role in FLT3 transactivation [30]. In all three cell lines, ARQ 531 inhibited the phosphorylated levels of SYK at the tyrosine 525/526 activation loop site in the SYK kinase domain which is a marker for SYK function [41]. Interestingly, the selective FLT3 inhibitors, quizartinib and gilteritinib in addition to the multi-kinase inhibitor midostaurin also down-modulated SYK phosphorylation only in the FLT3-ITD cell lines, but not in the FLT3-WT OCI-AML3 cell line (Fig. 3c). This implies that the SYK down-modulation in the ITD cell lines is the result of abrogating FLT3-ITD-dependent SYK activation as previously suggested by Puissant et al. [30], whereas inhibition of SYK signaling can occur directly in the FLT3 WT OCI-AML3 cell line.

Several reports have shed light on the involvement of BTK in AML, especially in the context of FLT3-ITD mutation [32, 33, 42‐44], and ARQ 531 potently inhibits BTK in the AML cell lines independent of FLT3 status (Fig. 3d). Although both ARQ 531 and ibrutinib both potently inhibit BTK phosphorylation, only ARQ 531 demonstrated in vitro anti-proliferative activity in MOLM-13 cell line (Additional file 6: Figure S2). Alternatively, the Syk inhibitor entospletinib while having no effect on BTK phosphorylation also potently inhibited proliferation in MOLM13 cells. These data suggest that the activity of ARQ 531 is attributed to the broad kinase inhibitory profile, beyond BTK.

Our data suggests the potential clinical benefit of ARQ 531 as a multi-kinase inhibitor targeting FLT3 upstream regulators, SFK and SYK and subsequently FLT3 and BTK. The effect does not appear to be FLT3 dependent, and further studies are warranted to identify mutational subsets most sensitive to ARQ 531.

We next determined if SYK inhibition is essential for the response to ARQ 531. We constitutively overexpressed the SYK-TEL fusion protein that has been previously utilized to study the role of SYK in response to therapeutic agents [30, 31]. The SYK-TEL construct maintains the SYK SH2 domains allowing the SYK and FLT3 interaction, and also contains a portion of the TEL/ETV6 transcription factor protein which results in enhanced FLT3 activation [30].

The Ba/F3 murine pro-B cell line expressing FLT3-ITD (IL-3 independent) [39] was transduced with pMIT-SYK-TEL or empty vector control to determine if the potency of ARQ 531 is reduced or abrogated with SYK overexpression. As expected, phosphorylated SYK was increased in the presence of SYK overexpression and we observed a reduced ability of ARQ 531 to decrease SYK and STAT5 phosphorylation in the pMIT-SYK-TEL expressing cell line compared with the pMIT-Empty control (Fig. 4a). Subsequently, there was more than twofold increase in the IC₅₀ in the Ba/F3-FLT3-ITD pMIT-SYK-TEL cell line compared with the parental (Additional file 7: Table S6 and Fig. 4b) suggesting that the overexpression of SYK reduced ARQ 531 potency. Similarly, both midostaurin and gilteritinib exhibited the same pattern of reduced potency with SYK-TEL expression (Additional file 7: Table S6 and Fig. 4c, d). Altogether, ARQ 531 potency is reduced in the event of SYK overexpression suggesting that SYK is an important target for ARQ 531 activity.

We next sought to investigate the in vivo activity of ARQ 531 in the previously described MOLM-13 xenograft model [38]. Seven days post-engraftment, mice were randomized to receive vehicle (n = 7) or 50 mg/kg ARQ 531 (n = 8) daily, a dose shown previously effective in TCL1 adoptive transfer mouse model of CLL [36]. Animals were monitored for ERC and overall survival (OS), with a subset (n = 3) injected with luciferin and imaged weekly for tumor burden (Additional file 8: Figure S1). Despite the aggressiveness of the MOLM-13 disseminated xenograft model, ARQ 531 demonstrated a modest yet significant median 2-day survival advantage (23 days vs. 21 days, log-rank p value = 0.002; Fig. 5 and Additional file 9: Table S7).

BCL-2 inhibition using venetoclax has shown modest activity as a monotherapy in AML [45], and venetoclax resistance has been reported especially in the context of MCL-1 upregulation [46]. Therefore, combination approaches aimed to synergistically enhance venetoclax activity and overcome resistance have been explored, such as combinations with a BRD4 inhibitor [47], an MCL-1 inhibitor [45‐47], or an MDM2 antagonist [48]. We examined if ARQ 531, to its strong inhibition of phosphorylated ERK which is essential for MCL-1 expression [49], is synergistic with venetoclax. In the MOLM-13 FLT-ITD cell line, we observed MCL1 downregulation with ARQ 531 treatment and a stronger inhibition when combined with venetoclax (Fig. 6a). Next, we determined potential additivity or synergy for the ARQ 531 + venetoclax (ARQ + Ven) combination compared with ARQ 531 monotherapy using the Combenefit software [50], which implements the classical mathematical models Highest Single Agent (HSA), Loewe, and the Bliss (Fig. 6b). As an additional check, we determined if the estimated decrease in proliferation in the presence of ARQ + Ven was greater than the decrease in proliferation with ARQ alone added to the decrease in proliferation with Ven alone (that is, the additive effect of ARQ, Ven) via separate mixed effects models for each dose combination. We considered all combinations with p < 0.05 as potential candidates for synergy (Fig. 6c and Additional file 11). Collectively, our in vitro analysis suggests enhanced activity for ARQ 531 in combination with the BCL-2 inhibitor, venetoclax. Subsequently, we tested if the ARQ + Ven combination enhanced ARQ 531 in vivo again using the aggressive MOLM-13 disseminated AML xenograft model. We observed a prolonged survival in the ARQ + Ven cohort in comparison with either agents alone (ARQ + Ven, 29 days; ARQ 531, 27 days; venetoclax, 24 days; vehicle, 25 days; overall log-rank test p value < 0.001; Fig. 7a and Additional file 9: Table S7) and reduced tumor burden (Fig. 7b, Additional file 10: Figure S3). Altogether, ARQ + Ven exhibited an increased survival advantage compared with ARQ 531 single therapy in an aggressive AML xenograft model.