Annals of Hematology

, Volume 91, Issue 3, pp 331–344

3,4-Diarylmaleimides—a novel class of kinase inhibitors—effectively induce apoptosis in FLT3-ITD-dependent cells

Authors

    • Department of Hematology and Oncology, Medical CenterOtto-von-Guericke University
  • Thomas S. Mack
    • Department of Hematology and Oncology, Medical CenterOtto-von-Guericke University
  • Elena Razumovskaya
    • Experimental Clinical Chemistry, Wallenberg Laboratory, Department of Laboratory MedicineSkåne University Hospital, Lund University
  • Marie-Christine Blum
    • Department of Hematology and Oncology, Medical CenterOtto-von-Guericke University
  • Daniel B. Lipka
    • Department of Hematology and Oncology, Medical CenterOtto-von-Guericke University
  • Anne Ballaschk
    • Department of Hematology and Oncology, Medical CenterOtto-von-Guericke University
  • Jan-Peter Kramb
    • Department of PharmacyJohannes-Gutenberg-University
  • Stanislav Plutizki
    • Department of PharmacyJohannes-Gutenberg-University
  • Lars Rönnstrand
    • Experimental Clinical Chemistry, Wallenberg Laboratory, Department of Laboratory MedicineSkåne University Hospital, Lund University
  • Gerd Dannhardt
    • Department of PharmacyJohannes-Gutenberg-University
  • Thomas Fischer
    • Department of Hematology and Oncology, Medical CenterOtto-von-Guericke University
Original Article

DOI: 10.1007/s00277-011-1311-3

Cite this article as:
Heidel, F.H., Mack, T.S., Razumovskaya, E. et al. Ann Hematol (2012) 91: 331. doi:10.1007/s00277-011-1311-3

Abstract

FLT3 kinase has become an attractive drug target in AML with up to 30% of cases harboring internal-tandem-duplication (ITD) mutations. For these, conferring a worse prognosis and decreased overall survival, several FLT3 tyrosine kinase inhibitors (TKIs) are currently being tested in clinical trials. However, when using these drugs as monotherapy, the problem of short duration of remissions and high incidence of TKI resistance has emerged. Here, we investigated two members of a novel class of tyrosine kinase inhibitors, 3,4-diarylmaleimides, for their efficacy on mutated FLT3 kinase. These compounds inhibit FLT3 kinase in an ATP-competitive manner and effectively inhibit phosphorylation of downstream targets. 3,4-Diarylmaleimides (DHF125 and 150) induce apoptosis in FLT3-ITD-dependent cells lines and patient blasts at low micromolar concentrations. They are retained in the cytoplasm of exposed cells for more than 24 h and synergize with chemotherapy and midostaurin. Both 3,4-diarylmaleimides show inhbition of FLT3-ITD-related kinase autophosphorylation at distinct tyrosine residues when compared to midostaurin. In conclusion, this novel group of compounds shows differential inhibition patterns with regard to FLT3 kinase and displays a promising profile for further clinical development. Currently, experiments evaluating toxicity in murine models and unraveling the exact binding mechanism are under way to facilitate a potential clinical application.

Keywords

AMLFLT3Tyrosine kinase inhibitorTyrosine phosphorylation

Introduction

Acute myeloid leukemia (AML) is an aggressive malignant disease characterized by abnormal proliferation of immature hematopoietic cells. Despite advances in chemotherapeutic treatment within the last decade, only 30–45% of patients below the age of 60 can be cured. The majority of patients, above the age of 60, have a dismal prognosis with only 5–15% long-term survival [15]. Several prognostic factors have been established within the last decade. While cytogenetic changes are established prognostic markers survival and response to chemotherapy, several molecular markers have been defined and analyzed for their prognostic influence [6]. Besides overexpression of the brain and acute leukemia gene [7], the v-ets erythroblastosis virus E26 oncogene [8, 9] and high meningioma 1 gene [10], mutation of nucleophosmin-1 (NPM1) [11], CAATT/enhancer binding protein [12], mixed-lineage-leukemia gene, Wilms’ tumor 1 gene [13] and FMS-related tyrosine kinase (FMS-like tyrosine kinase 3–internal tandem duplications (FLT3-ITD)) [14, 15] have shown prognostic influence.

The FLT3 gene is encoding a receptor tyrosine kinase that is part of the receptor tyrosine kinase III family (together with KIT, CSF-1R and PDGFRalpha and beta). FLT3 is expressed on the blasts of over 90% of AML cases. Mutations of FLT3 divide into length mutations (ITD) and point mutations. FLT3-ITD is a heterogenic group of genetic alterations found in up to 30% of AML cases [1618]. These mutations confer a dismal prognosis with impaired progression free and overall survival [14, 15, 19, 20]. FLT3-ITDs lead to autophosphorylation and constitutive activation of the FLT3 receptor with consecutive phosphorylation of downstream signaling nodes such as STAT5, AKT, and ERK, presumably by disruption of the autoinhibitory region [21]. Transfection of murine hematopoietic cells with FLT3-ITD leads to growth factor independent proliferation and development of a myeloproliferative phenotype in a murine bone marrow transplant model [22].

Small molecule tyrosine kinase inhibitors are a newly established class of therapeutic drugs. In AML, several tyrosine kinase inhibitors are currently tested in advanced clinical trials. Using these drugs as monotherapy has revealed remarkable efficacy. However, at the same time the problem of short duration of remissions has emerged indicating rapid development of secondary resistance [23, 24]. In addition, up to 30% of patients may show primary resistance to currently available FLT3-TKIs. This imposes a strong need to further develop treatment strategies: besides immunotherapeutic approaches [25], development of novel “second-generation” inhibitors using different mechanisms of action to achieve increased efficacy are clearly warranted.

Here, we investigated two novel TKIs (3,4-diarylmaleimides, DHF125 and DHF150) previously described as angiogenesis inhibitors [2628], to determine their mechanism of action and efficacy in FLT3-ITD-positive cell lines as well as in primary ITD-positive AML blasts.

Materials and methods

DNA constructs and generation of transfected cells

A human FLT3-ITD construct (amino acid position, N598; amino acid sequence, NEYFYVDFREYE), subcloned into the pAL expression vector under control of the 5′ long terminal repeat of the Moloney murine sarcoma virus and the plasmid pMAM/BSD were used as previously described [24]. The pAL-vector construct was stably transfected into the murine IL3-dependent myeloid cell line 32D by electroporation.

Cells were cultured in RPMI 1640 supplemented with 10% fetal calf serum (FCS), 20 mM Hepes (pH 7.3), 50-mM β-mercaptoethanol, and 2 mM l-glutamine.

Protein extract preparation, immunoprecipitation, and Western blotting

Cells at 2 × 106 were incubated in the presence of different inhibitor concentrations and in combination of different inhibitors for 1 h at 37°C. Preparation of cellular lysates was performed as described previously [24]. Protein lysates were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and blotted on nitrocellulose membrane (Amersham, Freiburg, Germany). The following antibodies were used: anti-phospho-FLT3 (pTyr591), anti-phospho-STAT5 (pTyr694/699), anti-phospho AKT (pSer473), anti-phospho ERK (pThr202/pTyr204) (all Cell Signaling Technology™, Frankfurt, Germany), antiphosphotyrosine (Biomol, Hamburg, Germany; and Santa Cruz, Heidelberg, Germany), anti-FLT3, anti-STAT5 (Santa Cruz, Heidelberg, Germany), anti-AKT (Cell Signaling Technology, Danvers, MA, USA), and anti-β-actin. Densitometric analysis was performed using the program Gel-Pro Analyzer®. For immunoprecipitation, 1 × 107 inhibitor/DMSO-treated MV4;11 cells were used for protein extraction; 100 μL of protein lysates were incubated with 20 μL of FLT3 antibody (sc-479, Santa Cruz, Heidelberg, Germany) for 24 h at 4°C. AG beads were added for 1 h at 4°C, followed by multiple washing steps. Finally, AG beads were covered with 30-μL β-mercaptoethanol-containing loading dye, heated for 5 min at 95°C followed by a quick spin-down. The supernatant was subjected to SDS-PAGE gel electrophoresis and Western blotting. Phospho-specific antibodies were produced and applied for Western blotting experiments as previously described [29].

Kinase assays

For determination of in vitro FLT3 kinase activity, we used HTScan™ FLT3 Kinase Assay Kit (Cell Signaling Technology, Inc., Danvers, MA, USA). Analysis was performed according to the manufacturer’s protocol. For ATP competition experiments 10-fold of ATP concentration (200 μM instead of 20 μM) was used. Absorbance was measured using a standard ELISA reader at 450 nm.

A second kinase assay was performed on a commercial basis by Millipore UK Ltd. (Gemini Crescent; Dundee Technology Park, Dundee DD2 1SW, UK). Details of assay protocols can be found on Millipore’s website at www.millipore.com/drugdiscovery/dd3/assayprotocols.

Apoptosis assays

Transfected murine 32D cells (1 × 105 cells/well) were incubated in 2 ml RPMI 1640 medium with different inhibitor concentrations for 48 h at 37°C. Primary human cells (2 × 105 cells/well) were incubated 72 h at 37°C. After incubation, cells were washed with ice-cold PBS, pelleted, and mixed with 300 μl of propidium-iodide-buffer (containing 50 μg/ml PI in 0.1% sodium citrate plus 0.1% Triton × 100, Sigma) for 30 min at 4°C. Cell cycle analysis was performed as described previously [30] using FACSCanto™ flow cytometers (BD Biosciences, Heidelberg, Germany). For AnnexinV/SytoxRed staining, treated 32D cells were washed with ice-cold PBS, resuspended in AnnexinV buffer and stained with AnnexinV-FITC and SytoxRed (Invitrogen) for 30 min on ice.

Isolation of primary AML blasts and cell culture

Bone marrow and peripheral blood samples with heparin as anticoagulant were obtained from AML patients (patient characteristics are indicated below) or donors with no evidence of malignant bone marrow infiltration after informed consent in a study approved by the local ethics committee. Mononuclear cells (MNC) were isolated immediately by means of Ficoll-Hypaque (Seromed, Berlin, Germany) density gradient centrifugation. For immunoblotting freshly isolated MNCs were either lysed directly or after incubation in RPMI 1640 medium supplemented with 20 mM Hepes (pH 7.3), 50-mM β-mercaptoethanol and 2 mM l-glutamine containing varying amounts of different tyrosine kinase inhibitors (midostaurin (protein kinase C (PKC)412), DHF125, and DHF150). For cell cycle analysis, MNCs were maintained in RPMI 1640 medium supplemented as above plus 10% FCS.

Patient characteristics

Patient characteristics (apoptosis assay)

Patient no.

Age/gender

FAB

WBC/uL

% of Blasts

Cytogenetics

FLT3 mutation*

Patient 1

79/F

M5

75,800

83

46,XX

ITD

Patient 2

78/M

M4

52,300

62

46,XY

ITD

Patient 3

88/F

M5

181,000

51

46,XX

ITD

Patient 4

39/F

M3

30,000

50

46,XX, t(15;17)

ITD

FLT3 mutation status was confirmed using standard diagnostic primer as follows: FLT3-ITD-fw 5-GCAATTTAGGTATGAAAGCCAGC-3 and FLT3-ITD-rev 5-CTTTCAGCATTTTGACGGCAACC-3.

Patient characteristics (cellular uptake assay)

Patient no.

Age/gender

FAB

WBC/uL

% of Blasts

Patient 1

88/F

M5

137,400

90

Patient 2

75/M

M1

80,500

66

Patient 3

76/F

M5

3,700

70

Patient 4

39/F

M3

30,000

50

Patient 5

78/M

M4

52,300

62

Patient 6

30/F

M4

151,400

70

Colony assays

Bone marrow MNCs were isolated as indicated above. Bone marrow samples were obtained from patients in complete remission after leukemia or lymphoma treatment and without evidence of any malignant bone marrow infiltration; 1 × 106 cells were plated in 1.1 ml Methocult™ (GF H4534) “complete” methylcellulose medium with recombinant cytokines (StemCell Technologies, Vancouver, Canada) and incubated for 10 days at 37°C. Colony formation was counted on day 10.

Inhibitors

3,4-Diarylmaleimides (academically developed at the Department of Pharmacy, Johannes-Gutenberg-University Mainz [2628] and PKC412A (kindly provided by J. Roesel, Novartis Inc.) were dissolved and diluted in DMSO (Sigma, Munich, Germany). The chemotherapeutic drugs cytarabine and daunorubicin have been diluted in water. Equivalent doses of DMSO were added when used for combination treatment.

Detection of autofluorescence and cellular uptake

To determine uptake and persistence of 3,4-diarylmaleimides in AML blasts, primary cells were incubated at 37°C in a 5% CO2-humified incubator with either compound in RPMI 1640 medium supplemented with 10% FCS. Uptake was determined by FACS analysis (FACSCanto, BD Biosciences, Heidelberg, Germany) by mean fluorescence intensity (DHF125, PE-Cy5-A mean; DHF150, PE-A mean). For determination of persistence of 3,4-diarylmaleimides in primary blasts, cells were incubated for 10 min with either compound, washed twice with PBS, and incubated at 37°C in RPMI 1640 medium supplemented with 10% FCS for 24 h. Autofluorescence was detected as indicated above.

Fluorescence of 3,4-diarylmaleimides was detected by confocal laser microscopy using a Zeiss Axiovert 100 M microscope attached to an LSM confocal unit. 3,4-Diarylmaleimides were added to murine 32D cells or human AML-blasts 5 min prior to microscope imaging. Images were edited using Zeiss LSM Image Examiner (Version 3.2.0.115).

Quantification of synergism and antagonism in drug combinations

For definition of synergy and/or antagonism in drug combinations we used the CompuSyn™ software (Chou, TC; ComboSyn, Inc. Paramus, NJ, USA) as previously described [30].

Statistical calculations

Statistical analysis was performed using GraphPad PrismTM v4.00. All experiments were performed in triplicates unless otherwise stated. For statistical comparison of colony formation upon diarylmaleimide treatment, a paired Student’s t test was applied. Statistical significance was assumed for p < 0.05.

Results

3,4-Diarylmaleimides inhibit FLT3 kinase in an ATP-dependent manner and reduce phosphorylation of its downstream signaling molecules

3,4-Diarylmaleimides are a novel class of inhibitors that have been initially described as potent inhibitors of angiogenesis in an in vivo chick embryo assay. Molecular modeling studies suggested binding of these compounds at the ATP-binding site of the model kinase CDK2. Screening for inhibitory activity in a panel of 12 selected protein kinases revealed a high affinity to VEGF-R but not to other kinases investigated (e.g., IGF-R, EGF-R, ABL, PDGFR, CDK2, CDK4, GSK3-b, and Aurora A&B), indicating a narrow spectrum of target structures [27, 28]. Herein, we aimed to investigate whether two structurally related 3,4-diarylmaleimides, DHF125 and DHF150 (Fig. 1a), inhibit FLT3 kinase and its downstream effectors. The majority of tyrosine kinase inhibitors are known to inhibit mutated FLT3 kinase by competing with ATP binding in the ATP-binding pocket. This reduces or abrogates phosphorylation of FLT3-ITD downstream targets, such as STAT5, AKT, and ERK.
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-011-1311-3/MediaObjects/277_2011_1311_Fig1_HTML.gif
Fig. 1

3,4-Diarylmaleimides inhibit FLT3 kinase activity and phosphorylation of its downstream targets. a Chemical structures of both diarylmaleimide compounds investigated are displayed. b Using kinase assays, inhibition of mutant FLT3 kinase was evident at low micromolar concentrations. Using the HTScan assay, for DHF125 and DHF150, IC50 concentrations of ~350 and ~450 nM were detectable (left panel). In a second kinase assay performed by Millipore Inc., the IC50 for FLT3 kinase was 166 nM for compound DHF125 (right panel). c 32D-FLT3-ITD protein lysates show reduction of STAT5, ERK, and AKT phosphorylation upon incubation of cells (for 1 h) with PKC412 (100), compound DHF125 (0.5 and 1 μM) and compound DHF150 (1 μM), respectively

To directly address inhibition of FLT3 kinase activity, we performed FLT3 kinase assays. DHF150 and DHF125 effectively inhibited FLT3 kinase with an IC50 of ~350 and ~450 nM, respectively. As suggested by the initial kinase screen, we aimed to confirm the ATP-competitive mechanism of 3,4-diarylmaleimides. As inhibitors binding in the ATP-pocket can be outcompeted by escalation of ATP [31], we applied a tenfold ATP concentration (Fig. 1b, left panel). These experiments revealed that both DM compounds could be antagonized by ATP escalation, indeed indicating an ATP-competitive mechanism of action. Midostaurin used as a positive control, also revealed ATP-dependent inhibition of FLT3 kinase (data not shown). This FLT3 kinase-specific mechanism of inhibition was confirmed using a commercially available FLT3 kinase assay performed by Millipore. In this assay, DHF125 showed inhibition of FLT3 kinase with an IC50 of 166 nM (Fig. 1b, right panel). The differences between the two kinase assays applied may be best explained by the different methodology: antibody-based array technology versus detection of kinase activity by radioactive labeled ATP, respectively.

In order to monitor inhibition of FLT3-ITD targets, we investigated the main downstream signaling nodes—STAT5, ERK, and AKT by Western blotting. Upon incubation of 32D-FLT3-ITD cells with either DM compound, a marked decrease in phosphorylation was detectable for STAT5 and ERK (Fig. 2c, left panel) while a slight reduction was seen for AKT phosphorylation. PKC412 served as a positive control for inhibition of these signaling nodes. These findings were confirmed using the human MV4;11 cell line (Fig. 2c, right panel). Incubation with either PKC412 or compound DHF125 led to reduced phosphorylation of STAT5 and ERK. In contrast, the human cell line HEL, harboring an activating JAK2V617F mutation, did not show any reduction in phosphorylation of STAT5 or ERK when compared to DMSO-treated control.
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-011-1311-3/MediaObjects/277_2011_1311_Fig2_HTML.gif
Fig. 2

3,4-Diarylmaleimides induce apoptosis in FLT3-ITD-dependent cell lines. Percentage of apoptotic cells was measured after a 48-h incubation period with either compound by flow cytometry using AnnexinV/SytoxRed staining (a), PI-based cell cycle analysis (b, c) and caspase-3 cleavage (d) in Western blotting. Calculated IC50 concentrations for DHF125 and DHF150 using 32D-FLT3-ITD cells were 375 and 850 nM, respectively (b). Addition of growth factor (IL-3) leads to rescue of cells upon incubation with either compound. e 32D-FLT3-WT cells grown in IL-3 containing medium show no toxic effects up to 10 μM of inhibitor concentration, as displayed for DHF125

Thus, we were able to provide first evidence that both 3,4-diarylmaleimides inhibit FLT3 kinase and its downstream targets in an ATP-competitive manner.

3,4-Diarylmaleimides (DHF compounds) effectively induce apoptosis in FLT3-ITD-dependent cell lines

To determine the effects of 3,4-diarylmaleimides on mutated FLT3 tyrosine kinase in a cellular context, we transfected the murine IL3-dependent hematopoietic 32D cells with an FLT3 length mutation (FLT3-ITD)- or FLT3-wildtype (FLT3-WT)-cDNA construct. While FLT3-ITD-positive cells gain growth factor-independent proliferation, FLT3-WT cells generally remain IL3 dependent. 32D-FLT3-ITD-positive cells are known to undergo apoptosis upon incubation with FLT3 inhibitors like PKC412 in the absence of IL-3 [32]. However, when IL-3 is added, induction of apoptosis can be prevented bypassing FLT3-ITD dependency via growth factor receptor. Therefore, we incubated these cells with either DHF compound in the presence or absence of IL-3.

Induction of apoptosis was detectable at low micromolar concentrations (Fig. 2a), as detected by flow cytometry analysis using AnnexinV (as an early apoptosis marker) and SytoxRed (as a dead cell stain). Analyzing the subG1 fraction in cell cycle analysis upon incubation with DHF125 and DHF150, we were able to define an IC50 of 375 and 850 nM, respectively (Fig. 2b, c). To confirm the FLT3 specificity of these effects, FLT3-ITD-positive cells were also cultured in the presence of IL-3 allowing the cells to bypass FLT3-ITD dependency by activation of alternative signaling pathways. As depicted in Fig. 2b, induction of apoptosis was reversible in both cases, confirming FLT3-ITD specificity of the effects observed. Induction of apoptosis was further confirmed by Western blotting experiments as indicated by caspase-3 cleavage (Fig. 2d).

As 32D cells transfected with an FLT3-WT construct are cultured in growth factor containing medium, any induction of cell death would potentially display toxic off-target effects. To elucidate the “maximal tolerable” or toxic dose for 32D transfectants, we incubated 32D-FLT3-WT cells with increasing concentrations of either compound until cell death occurred. 32D-FLT3-WT cells showed no significant increase in apoptosis up to 10 μM of either DHF compound (Fig. 2e; displayed is data for DHF125). However, higher doses of up to 50 μM revealed increasing rates of cell death in terms of cellular toxicity. This provides first evidence for a possible therapeutic range of this novel class of inhibitors.

To evaluate a potential broader spectrum of kinase inhibition, we analyzed the effects of DHF compounds on JAK2-kinase that had not been included in previous screens. Ba/F3 cells harboring the V617F mutation leading to constitutive activation of JAK2-kinase were incubated with increasing concentrations of either inhibitor. No significant induction of apoptosis could be demonstrated using doses up to 5 μM while apoptosis could be induced using JAK inhibitor (data not shown).

This suggests that 3,4-diarylmaleimides effectively induce apoptosis in FLT3-ITD-dependent murine cell lines at low micromolar concentrations. These effects are specific for mutated FLT3 kinase and unspecific effects do not occur until doses above 50-fold the IC50.

3,4-Diarylmaleimides show synergy with chemotherapy and PKC412

Applied as monotherapy, various tyrosine kinase inhibitors have shown promising responses. However, in most cases, these responses were short lived and insufficient for disease control. Currently, advanced clinical trials evaluate tyrosine kinase inhibitors in combination with chemotherapy. As anthracyclines and cytarabine are standard drugs for the treatment of AML, we investigated the efficacy of a combination therapy with DHF inhibitors. Synergy was calculated using Chou–Talalay plots (CI Index).

Combination of the anthracycline daunorubicin with either DHF compound showed clear induction of apoptosis in 32D-FLT3-ITD cells (Fig. 3a). Increasing doses of either chemotherapeutic drug were combined with fixed inhibitor concentrations in a non-constant ratio design [30]. For daunorubicin, calculating the effective concentrations applied according to the Chou–Talalay model revealed synergy at higher concentrations, with slightly antagonistic/additive effects at lowest concentrations (Fig. S1 in the Electronic supplementary material). In contrast, simultaneous exposure of cells to cytarabine plus DHF inhibitors revealed synergistic effects at all concentrations applied (CI indices were calculated below 1 for all combinations).
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-011-1311-3/MediaObjects/277_2011_1311_Fig3_HTML.gif
Fig. 3

3,4-Diarylmaleimide compounds synergize with chemotherapy and midostaurin in vitro. Combined treatment of 32D-FLT3-ITD cells with diarylmaleimides and daunorubicin (a, left panel) or cytarabine (a, right panel) for 48 h resulted in markedly increased amounts of apoptotic cells. Co-incubation of cells with DHF125 or DHF150 and midostaurin (PKC412) (b, left panel) demonstrated increased efficacy with regard to apoptosis induction. Synergy could be calculated at all effective doses applied (b, right panel)

Recent reports suggest that FLT3-TKI can display distinct mechanisms of action [33, 34]. This offers a potential for combination therapy. Thus, we combined both DHF compounds with midostaurin to investigate potential synergy. Interestingly, co-incubation of PKC412 with DHF125 or DHF150, using a constant ratio design, revealed synergy (CI index, <1) at all concentrations applied (Fig. 3b). PKC412 doses used in this assay varied from 0.5 to 7 nM. In clinical phase 2 trials, PKC412 trough levels of up to 40 nM have been shown to be reached and have shown limited clinical efficacy with tolerable toxicity [24, 35, 36]. Thus, our results offer a rationale to clinically investigate TKI combination treatment in order to reach increased efficacy. Moreover, this suggests a distinct mechanism of action for DHF compounds, as identical binding would rather lead to competition of both compounds and potentially to antagonistic effects.

3,4-Diarylmaleimides induce apoptosis in primary AML blasts harboring FLT3-ITD mutations in vitro while leaving colony formation of normal stem and progenitor cells unaffected

In contrast to cell culture models, primary blasts may display a variety of genetic and epigenetic modifications, leading to increased resistance to apoptosis or proliferation.

To determine the extent of apoptosis induction in primary blasts upon inhibition of mutated FLT3 kinase, we incubated primary AML cells for 72 h with increasing concentrations of either DHF-inhibitor. Induction of apoptosis was detected by measuring the subG1-fraction after propidium iodine staining. When taken in culture primary patient blasts revealed induction of apoptosis up to 30%. Therefore, we investigated cell death exceeding baseline apoptosis. Incubation with DHF125 led to increase of apoptosis by 10–17% at 1–2 μM with no further effect upon dose escalation up to 5 μM. In contrast, using DHF150, the rate of apoptotic cells increased by 12–15%. Upon dose escalation (5 μM) 17–30% of primary blasts became apoptotic after 72 h of incubation (Fig. 4a). Although conducted in an approach limited by spontaneous apoptosis, we were able to detect induction of cell death in primary patient blasts upon incubation with 3,4-diarylmaleimides.
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-011-1311-3/MediaObjects/277_2011_1311_Fig4_HTML.gif
Fig. 4

3,4-Diarylmaleimide compounds induce apoptosis in FLT3-ITD harboring AML blasts at micromolar concentrations while leaving normal progenitor cells largely unaffected. diarylmaleimide compounds show induction of apoptosis in primary AML blasts. Primary cells were incubated for 72 h with different inhibitor concentrations. Apoptosis was determined as the amount of cells subG1 exceeding the baseline apoptosis rate of AML cells in culture (baseline apoptosis is graphed as 0%, every patient sample (single dot) is graphed as the mean of a triplicate). DHF125 led to increase of basal apoptosis by 10–17% while DHF150 elevated the rate of apoptotic cells by 17–30% (a). Using bone marrow cells from healthy donors, colony formation was analyzed upon increasing doses of diarylmaleimide inhibitors—in methylcellulose supplemented with cytokines—for 10 days. Impairment of colony formation was not detectable up to a dose of 5 μM of either compound. Incubation of these cells with 10 μM of either inhibitor led to almost complete loss of colony formation, suggesting toxicity on healthy hematopoietic stem and progenitor cells (b). Combination therapy of DM compounds with midostaurin (PKC412) did not reveal significant toxicity on colony formation at all doses applied (c)

To approach a possible therapeutic dose range, we investigated bone marrow cells of healthy donors or patients in stable remission after chemotherapy treatment. Toxicity was evaluated as impairment of colony formation. Usually, stem and progenitor cells in healthy human bone marrow have the ability to form colonies when plated in methylcellulose while further differentiated cells lack that ability. Therefore, choosing a colony formation assay seemed to be a relevant approach to investigate potential hematopoietic toxicity that would arise from impaired stem or progenitor compartments.

Addition of diarylmaleimide compounds showed no significant reduction in colony formation up to doses of 5 μM (p = 0.39 for DHF125—5 μM compared to DMSO control; p = 0.33 for DHF150—5 μM compared to DMSO control) (Fig. 4b). However, dose escalation of up to 10 μM revealed markedly impaired colony formation as readout for hematopoietic toxicity. This indicates that normal human stem and progenitor cells may not be affected by exposure to 3,4-diarylmaleimides concentrations of up to 5 μM of either compound (Fig. 4b). As we had investigated efficacy of a combination therapy using DM compounds with midostaurin (Fig. 3), we aimed to assess the impact of a combination therapy on colony formation of normal bone marrow. All doses investigated did not reveal any significant reduction in colony formation when compared to DMSO-treated controls (Fig. 4c)

These data are consistent with our results observed in murine cell lines (as described above). As FLT3-ITD-specific induction of apoptosis in murine cell lines and primary patient blasts occur at low micromolar concentration, these results can be interpreted as a first indicator for a potential therapeutic window.

3,4-Diarylmaleimides show rapid uptake and prolonged intracellular persistence in transfected cell lines and primary AML blasts

Exposure of malignant cells to tyrosine kinase inhibitors is one of the crucial issues in TKI treatment. Recent publications provide first evidence that prolonged exposure of kinases to the corresponding inhibitors is a crucial prerequisite for induction of cell death. Therefore, we aimed to investigate the retention of 3,4-diarylmaleimides in both, murine 32D cells and primary patient blasts. As 3,4-diarylmaleimides show autofluorescence that can be used as a correlate for inhibitor binding, we performed confocal fluorescence microscopy and flow cytometry analysis.

Fluorescence of exposed cells (as analyzed in the PE or PECy5 channel by flow cytometry) could be detected in primary patient blasts (Fig. 5a) and murine 32D cells upon incubation for less than 1 min. When co-incubated with an APC-anti-CD33 antibody, most of the blasts showed accumulation of either inhibitor (Fig. 5a). To determine the kinetics of drug uptake and persistence, we incubated blasts with either DHF compound for 20 min. Increase in fluorescence was analyzed as described above. After 20 min, cells were washed twice with PBS and analyzed for fluorescence while cultured in RPMI 1640 medium at different intervals for 24 h. As increase in fluorescence could be due to inhibitor binding to the outer membrane of cells, we performed confocal fluorescence microscopy to determine whether 3,4-diarylmaleimides were accumulated in the cytoplasm. In both, murine 32D-FLT3-ITD cells as well as primary patient blasts, accumulation of 3,4-diarylmaleimide compounds could be detected in the cytoplasm within 5 min after exposure (Fig. 5b). Thus, confocal fluorescence microscopy confirmed that detection of fluorescence using flow cytometry was associated with rapid cytoplasmatic uptake and retention of either compound investigated.
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-011-1311-3/MediaObjects/277_2011_1311_Fig5_HTML.gif
Fig. 5

3,4-Diarylmaleimides are incorporated and retained in primary patient blasts. Primary blasts of FLT3-ITD-positive AML samples were incubated with anti-CD33, APC-conjugated antibodies (according to previous positivity of patient samples in clinical flow cytometry analysis) and 1 μM of either DM inhibitor (a). Increase in fluorescence intensity was detected in the majority of patient blasts within 5 min of incubation as evaluated by flow cytometry. Confocal laser microscopy revealed cytoplasmatic accumulation in 32D-FLT3-ITD cells as well as in primary patients AML blasts (b). Cells treated with DHF125 or DHF150 (c) reached a peak in fluorescence intensity after 5–10 min of incubation. After washout, fluorescence level remained stable within a range of 200–600% of baseline fluorescence for either compound for up to 24 h

Besides analysis of plasma levels in drug development, pharmacokinetics on a cellular level moves also into focus. As cells can retain inhibitors or even accumulate them in the cytoplasm, drug exposure can be prolonged and more effective. Therefore we aimed to address that specific question by incubating primary patient blasts for 20 min with 1 μM of either inhibitor, followed by washout and analysis of fluorescence levels compared to unstained controls.

For all patients investigated, a rapid uptake within 5–10 min with increase of mean fluorescence intensity of up to 2,600% of baseline values could be demonstrated. Of note, there was a large variability in uptake and accumulation of inhibitors as detected by increase of fluorescence between the different patient samples investigated. After washout, fluorescence levels decreased, but remained elevated up to 200–500% of baseline control within 24 h (Fig. 5c). This was detectable for both compounds with a comparable patient variability. However, none of the samples investigated showed a drop below 100% of baseline values.

In summary, these data show prolonged availability of both diarylmaleimides in the cytoplasm of leukemic cells for a sustained period of up to 24 h. This could be a first indicator for a promising clinical profile, as 3,4-diarylmaleimide may be retained or even accumulated in malignant blasts and thereby ensure a prolonged exposure of mutated FLT3 kinase.

3,4-Diarylmaleimides show a distinct profile for inhibition of FLT3-ITD-induced autophosphorylation

Activity of FLT3 tyrosine kinase inhibitors has been frequently measured using reduction of ITD-induced autophosphorylation as readout in Western blotting experiments. However, up to date few data exist on detailed analysis of the phosphorylation status of individual tyrosine residues upon inhibitor treatment. For example, it is not known, whether specific inhibition of autophosphorylation at distinct residues is required for inhibition of downstream targets and induction of apoptosis. The majority of publications investigating established FLT3 kinase inhibitors used total tyrosine phosphorylation (using the 4G10 antibody) as a correlate for kinase inhibition [32, 37]. In addition, analysis of Y591 phosphorylation has also been used. Recently, phosphorylation of tyrosine residues Y589, Y591, Y599, Y726, Y768, and Y955 has been shown to be involved in constitutive activation of FLT3 kinase induced by presence of ITD length mutations [29].

Given the fact that 3,4-diarylmaleimides acted synergistically with midostaurin in apoptosis assays, we hypothesized the possibility of distinct inhibition patterns using 3,4-diarylmaleimide compounds when compared to midostaurin. Indeed, Western blotting experiments using an FLT3-Y591 detecting phospho-specific antibody confirmed inhibition of autophosphorylation at this particular residue by midostaurin but not by any of the 3,4-diarylmaleimide compounds (Fig. 6a). This was consistent in murine 32D-FLT3-ITD cells (left panel) as well as in the human AML cell line MV4;11 harboring an FLT3-ITD mutation (right panel).
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-011-1311-3/MediaObjects/277_2011_1311_Fig6_HTML.gif
Fig. 6

3,4-Diarylmaleimides inhibit FLT-ITD-mediated autophosphorylation at distinct tyrosine residues compared to midostaurin. a Phosphorylation of FLT3 at Y591 (as detected by the P-Flt3-antibody applied) was abrogated by 100 nM of midostaurin (PKC412). In contrast, incubation with doses up to 1 μM of either DM compound did not affect autophosphorylation at Y591. b Immunoprecipitation experiments reveal distinct inhibition of phosphorylation at residue Y589, which is reduced by midostaurin but neither of the compounds investigated. For residues Y599, Y726, Y768, Y842, and Y955, neither midostaurin nor diarylmaleimides showed any effect on phosphorylation levels. Exclusively Y793 seems to be affected by both, diarylmaleimides and controls. Total FLT3 tyrosine phosphorylation (as detected by 4G10 antibody) showed clear reduction for midostaurin while diarylmaleimides merely caused a slight reduction

In order to investigate ITD-relevant tyrosine residues, we performed additional FLT3-ITD immunoprecipitation experiments in 32D-FLT3-ITD cells. Interestingly, investigating seven additional tyrosine residues, we found residues affected (1) by midostaurin treatment but not by 3,4-diarylmaleimides, (2) affected by neither group of compounds or (3) affected by both, midostaurin and 3,4-diarylmaleimides (Fig. 6b). None of the tyrosine residues was affected by 3,4-diarylmaleimides but not by midostaurin treatment.

In detail, tyrosine residue Y589 showed reduced phosphorylation upon midostaurin but was not affected by 3,4-diarylmaleimide treatment. Apparently, residues Y842, Y726, Y768, Y599, and Y955 were not significantly affected by any tyrosine kinase inhibitor treatment applied potentially indicating non-informative tyrosine residues in terms of TKI treatment. Residue Y793 showed reduction of autophosphorylation upon incubation with both 3,4-diarylmaleimides and PKC412. Total phosphorylation of FLT3, as determined using the 4G10 phospho-tyrosine antibody revealed a clear inhibitory effect of midostaurin as well as a minor reduction using DHF inhibitors (Fig. 6b). This effect could be seen best in the Y793 panel as no residual activity would be detected after stripping off pY793 and re-probing with 4G10.

These results implicate several novel findings: when investigating distinct tyrosine residues for their involvement in inhibitory activity of FLT3 kinase inhibitors, several residues involved in autophosphorylation caused by the presence of ITD mutations were not affected by midostaurin. Thus it is tempting to speculate that these residues (namely Y842, Y726, Y768, Y599, and Y955) do not play a crucial role in midostaurin-mediated reduction of kinase activity. Moreover, 3,4-diarylmaleimides seem to carry out their inhibitory activity without affecting tyrosine residues Y589 and Y591, which are believed to play a pivotal role in FLT3-ITD-mediated transformation. Thus, in future studies, the phosphorylation status of these residues will be investigated using a broad spectrum of kinase inhibitors to gain a deeper insight into TKI-induced regulation of ITD-mediated autophosphorylation.

Discussion

AML is a malignant disease with high biological and genetic heterogeneity and survival rates of 30–40% in patients <60 years upon myelosuppressive chemotherapy [24]. However, the majority of patients affected by AML are above the age of 60, with a long-term disease-free survival of only 5–15% after myelosuppressive chemotherapy treatment. Besides cytogenetic risk-stratification, molecular markers—such as NPM1 or FLT3 mutations—have shown prognostic impact [14, 15, 19]. Therefore, inhibition of mutated FLT3 tyrosine kinase may improve response rates to chemotherapy and overall survival. Currently, advanced clinical studies are testing this hypothesis, using FLT3 tyrosine kinase inhibitors such as PKC412 or lestaurtinib (CEP-701) in combination with myelosuppressive chemotherapy. The majority of tyrosine kinase inhibitors currently investigated in clinical phase 2–3 trials for FLT3-ITD-positive AML originally have not been designed as “FLT3-inhibitors.” Examples are PKC412 being developed as an inhibitor of PKC, BAY 43-9006 (Sorafenib) as an inhibitor of Ras/Raf or CEP-701 (Lestaurtinib) as an inhibitor of Trkl. As a ‘beneficial side effect’, these compounds have shown to inhibit FLT3 kinase mostly at low micromolar or nanomolar concentrations. Currently, second-generation inhibitors, developed and designed for inhibition of FLT3 kinase are being tested in clinical and pre-clinical settings. [38] [33]. Here, we describe a novel class of inhibitors that was initially detected in a screening approach for angiogenesis inhibitors. These 3,4-diarylmaleimides are obviously potent inhibitors of mutant FLT3 kinase and display several characteristics that make them a promising group for further clinical development: (1) 3,4-diarylmaleimides are retained in the cytoplasm of exposed cells and suggest a prolonged intracellular exposure of mutant FLT3 kinase, (2) they synergize with chemotherapeutic drugs and notably with midostaurin which may be explained (3) by a distinct inhibition pattern with regard to the phosphorylation status of specific FLT3 tyrosine residues.

Exposure of malignant cells to tyrosine kinase inhibitors is one of the crucial issues in TKI treatment. Several inhibitors did not reach advanced clinical studies due to pharmacokinetic/-dynamic problems and inadequate exposure of target cells for TKI. In addition, uptake into malignant cells displays an important issue in the clinical use of tyrosine kinase inhibitors. Recently, irreversible commitment to apoptosis upon short-term high-dose exposure to TKIs has been suggested as a general mechanism [39].

Using a combination therapy of DM compounds with midostaurin, we could achieve synergy and decent activity against 32D-FLT3-ITD cells. The concentrations applied for PKC412 are achievable, even as trough levels (12.4–157 nM corrected for 99% protein binding) during long-term treatment [24]. Although no studies have been conducted using 3,4-diarylmaleimides in vivo so far, other FLT3 kinase inhibitors (such as Sorafenib) have been shown to reach peak plasma levels up to 10 μM (when given twice daily at 400 mg) [40], suggesting a potential therapeutic window.

More importantly, our experiments provide first evidence for prolonged retention of both diarylmaleimide inhibitors investigated in leukemic cells for a period of up to 24 h, resulting in apoptotic cell death. Thus, both compounds investigated seem to meet the requirements for sufficient intracellular target inhibition.

In general, monotherapy treatment of relapsed or refractory leukemia using tyrosine kinase inhibitors in early clinical trials showed short-lived responses. However, first results from trials using TKI in combination with chemotherapy revealed promising duration of remission rates and suggested further development in clinical phase 3 studies. These trials are currently under way and underline the importance of potential synergistic effects when applied simultaneously or sequentially. 3,4-Diarylmaleimides revealed synergistic effects with both, cytarabine and daunorubicin, which can be considered standard therapeutics for AML treatment. The concept of combining different kinase inhibitors could be a novel approach to overcome resistance in AML therapy.

Recently, several tyrosine residues of FLT3 have been investigated for their differential phosphorylation in FLT3-WT, FLT3-D835Y, and FLT3-ITD kinase [29]. While tyrosine residues Y589, Y591, Y599, Y726, Y768, and Y955 are phosphorylated sites in FLT3-ITD as well as in FLT3-D835Y mutants and FLT3-WT cells upon stimulation with its physiological ligand (Flt3 ligand), Y793 and Y842 seem to be exclusively phosphorylated in FLT3-D835Y mutants [29]. Moreover, functional analysis of tyrosine residues within an FLT3-ITD background has already been performed [41, 42]. Tyrosine to phenylalanine substitution of the FLT3 tyrosine residues Y589 and Y591 on an ITD background led to significant reduction of STAT5 phosphorylation while not affecting tyrosine kinase activity or whole FLT3 tyrosine phosphorylation. As tyrosines Y589 and Y591 were identified as two candidate STAT5 SH2 docking phosphorylation sites in the juxtamembrane (JM) domain, this finding raised the possibility that they might display occult STAT5 binding sites and that ITD mutations expose these sites by conformational disruption of the JM domain. Single amino acid exchanges from Y to F at known tyrosine phosphorylation sites (Y591F, Y726F, Y955F, and Y969F) and even some combined exchanges (such as Y726F/Y955F and Y955F/Y969F) did not lead to significant reduction of STAT5 phosphorylation [41]. Consistent with this finding, single Y to F mutations of Y589, Y591 or Y597 in the ITD background did not reduce FLT3-transforming potential of ITD mutations [42]. However, combined mutations of certain phosphorylation sites, showed an influence on phosphorylation of STAT5 or reduced transforming capacity of ITD mutation (Y589F/Y591F, Y589F/Y597F, Y589F/Y599F, Y591F/Y597F, and Y591F/Y599F).

In our hands, DM inhibitors clearly inhibited or even abrogated phosphorylation of FLT3 downstream targets (such as STAT5) and inhibited FLT3 kinase activity. However, inhibition of phosphorylation at residues affected by midostaurin (such as FLT3-Y591 and FLT3-Y598) was not evident in Western Blottting. This indicates that residual phosphorylation of single tyrosine residues may have no influence on transforming potential or phosphorylation status of important downstream targets such as STAT5. Our data suggest that even Y591 which is thought to be involved in STAT5 binding within an FLT3-ITD background and which has been shown to reduce phosphorylation of downstream targets when mutated to phenylalanine in combination with cooperating phosphorylation sites may not be crucial for kinase inhibitor function.

Further studies investigating efficacy in cells resistant to currently available FLT3-TKIs and evaluation of toxicity in mouse models are clearly warranted and currently under way. Additional studies investigating differential inhibition patterns of FLT3 autophosphorylation at distinct tyrosine residues and experiments focusing on the exact binding mechanism of 3,4-diarylmaleimides are future aims but were beyond the scope of this report.

Acknowledgement

We thank Fian Mirea and Ann-Kathrin Borrmann for technical assistance and Dr. D. Strand (Department of Gastroenterology, University Hospital Mainz) for support with confocal microscopy. This work was supported by grants from the Johannes-Gutenberg-University (MAIFOR, Mainz Research Funding no. 9728249, to F.H.H.) and the German Cancer Aid (DKH 108218 and DKH 108401 (TP6) to T.F.).

Disclosures

F. Heidel, J.-P. Kramb, S. Plutitzki, G. Dannhardt, and T. Fischer filed a patent on the use of 3,4-diarylmaleimides in leukemia.

Supplementary material

277_2011_1311_Fig7_ESM.jpg (32 kb)
Fig. S1

Synergy of DHF compounds with either daunorubicin or cytarabine as investigated in apoptosis assays. Co-incubation of cells with DHF125 or DHF150 and daunorubicin (left panel) or cytarabine (right panel) demonstrated increased efficacy with regard to apoptosis induction. Synergy could be calculated using concentrations above the IC10 for daunorubicin and all effective doses applied for cytarabine. (JPEG 32 kb)

277_2011_1311_MOESM1_ESM.tiff (1.1 mb)
High-resolution image (TIFF 1,092 kb)
277_2011_1311_Fig8_ESM.jpg (49 kb)
Fig. S1

Synergy of DHF compounds with either daunorubicin or cytarabine as investigated in apoptosis assays. Co-incubation of cells with DHF125 or DHF150 and daunorubicin (left panel) or cytarabine (right panel) demonstrated increased efficacy with regard to apoptosis induction. Synergy could be calculated using concentrations above the IC10 for daunorubicin and all effective doses applied for cytarabine. (JPEG 32 kb)

277_2011_1311_MOESM2_ESM.eps (100 kb)
High-resolution image (EPS 99 kb)

Copyright information

© Springer-Verlag 2011