Introduction
Platelet-derived growth factor receptor α (PDGFRα) is a class III receptor tyrosine kinase with an extracellular domain, a single transmembrane domain, and a cytoplasmic tyrosine kinase domain [
1]. Upon ligand binding, the activated receptor drives multiple downstream pathways such as signal transducer and activator of transcription (Stat), Src kinases, mitogen-activated protein kinases, and phosphatidylinositol-3 kinase to coordinate cell proliferation, differentiation, survival, adhesion, and cell migration [
2]. Gain-of-function mutations in PDGFRα have been found in chronic myeloid leukemia (CML), gastrointestinal stromal tumors (GISTs) and chronic eosinophilic leukemia (CEL) [
3]. A typical example is the fusion gene of FIP1-like 1 (FIP1L1)-PDGFRα created by an 800-kb cryptic interstitial deletion in chromosome 4q12, which is pathogenic for a subset of CEL [
1]. FIP1L1-PDGFRα encodes a ligand-independent and constitutively active tyrosine kinase that is sensitive to imatinib [
4‐
6]. However, acquired resistance to imatinib can occur because of point mutations in the ATP binding site (e.g., T674I and D842V) [
7‐
9]. T674I FIP1L1-PDGFRα is a “gatekeeper” mutation: substitution of the gatekeeper threonine (T674) with a bulky amino acid (I) blocks access by imatinib and second-generation tyrosine kinase inhibitors (TKIs), such as nilotinib and dasatinib, to a hydrophobic pocket inside the ATP binding site [
10]. This gatekeeper mutation is analogous to the T315I mutation in Bcr-Abl [
1,
5,
11,
12]. The prognosis is poor for CEL patients harboring T674I FIP1L1-PDGFRα although it is rare in incidence.
To search for novel TKIs to overcome imatinib resistance, midostaurin, EXEL-0862 and sorafenib have been investigated both in vitro and in vivo in cells harboring T674I FIP1L1-PDGFRα [
2,
9,
13]. Thus far, the first clinical trial of sorafenib for T674I FIP1L1-PDGFRα-positive CEL showed a transient hematological response, but patients died of rapid emergence of additional sorafenib-resistant point mutations in FIP1L1-PDGFRα [
7]. Independently, Metzgeroth et al. reported limited clinical activity of sorafenib and nilotinib in T674I FIP1L1-PDGFRα-positive CEL patients [
14]. Therefore, imatinib-resistant CEL remains a therapeutic challenge.
There has been exciting recent advance in third-generation TKIs (ponatinib, HG-7-85-1 and DCC-2036) efficacious against the gatekeeper mutants [
15‐
18]. In vitro screening assay has demonstrated that ponatinib, the first TKI effective against T315I Bcr-Abl, is also a potent inhibitor of KIT, PDGFRα, Flt3, Src, VEGFR and FGFR [
15,
19,
20].
We investigated the molecular docking of ponatinib to T674I PDGFRα in silico. In vitro and in vivo study then confirmed that ponatinib is a potent inhibitor of CEL cells bearing wild-type (WT) or T674I FIP1L1-PDGFRα.
Discussion
Acquired resistance to TKIs presents a therapeutic challenge. Gatekeeper mutants (e.g. T315I Bcr-Abl, T670I KIT and T674I PDGFRα) are particularly multi-drug resistant. In the present study, ponatinib potently inhibited the phosphorylation of the WT and gatekeeper mutant T674I FIP1L1-PDGFRα and their downstream signaling. Our molecular docking analysis revealed that ponatinib could target native or T674I FIP1L1-PDGFRα in the DFG-out (inactive) binding mode, similar to ponatinib docking in T315I Bcr-Abl. This characteristic of ponatinib may be related to its imidazo [1,2b] pyridazine core that occupies the pocket for adenine in the enzyme, whereas the methylphenyl group occupies the hydrophobic pocket behind the gatekeeper residue of the enzyme [
15]. Encouraged by the
in silico simulation results, we evaluated the efficacy of ponatinib against imatinib-resistant CEL cells both in vitro and in vivo.
Ponatinib potently inhibit the viability of EOL-1 cells expressing WT FIP1L1-PDGFRα, with an IC
50 value of 0.004 nM. This efficacy agrees with recent results [
20] showing an inhibitory effect in EOL-1 cells, with an IC
50 of 0.5 nM. In the same study, ponatinib inhibited malignant cells expressing Bcr-Abl, Flt3, KIT, FGFR1, with IC
50 values from 2 to 36 nM [
20]. We showed that ponatinib had an inhibitory effect on imatinib-resistant leukemic BaF3-T674I FIP1L1-PDGFRα cells, with an IC
50 of 2.5 nM, which is comparable to the potency in BaF3-T315I Bcr-Abl cells, with an IC
50 of 11 nM [
15]. Clonogenicity assay confirmed that ponatinib restrained the proliferation of BaF3-WT or -T674I FIP1L1-PDGFRα cells at low nanomolar concentrations. Further, our in vivo data revealed that ponatinib, at an oral dose of 30 mg/kg/day, potently abrogated the growth of xenografted imatinib-resistant BaF3-T674I FIP1L1-PDGFRα cells, with PDGFRα signaling highly suppressed (Figure
6). A pharmacokinetics study in mice indicated that orally administrated ponatinib as a single oral dose of 30 mg/kg, which was well tolerated, resulted in mean plasma concentrations of 782 and 561 nM at 2 and 6 h post-dosing, respectively [
15]. Such plasma levels highly exceed the in vitro IC
50 values for all 3 lines of FIP1L1-PDGFRα-expressing cells, so ponatinib may efficiently inhibit the growth of FIP1L1-PDGFRα-positive cells with clinically achievable doses.
Ponatinib induced remarkable apoptosis in both imatinib-sensitive and -resistant CEL cells, as evidenced by Annexin V binding, activation of caspase-3, and specific cleavage of PARP. The apoptosis was triggered by the mitochondrial-dependent pathway because of release of AIF and cytochrome c to the cytosol after treatment with ponatinib. The levels of survivin, Bcl-X
L and Mcl-1were decreased in ponatinib-mediated apoptotic CEL cells. The transcription of survivin and Bcl-X
L is regulated by Stat3, Stat5 and Erk1/2 [
31,
32]. The decreased expression of survivin and Bcl-X
L caused by ponatinib treatment is likely associated with the inhibition of Stat3, Stat5 and Erk1/2. However, future experiments can further define the precise mechanisms.
Mcl-1, a pro-survival and anti-apoptotic protein with relatively short-half life in the Bcl-2 family, is expressed in malignant hematological cells and protects cells against apoptosis in response to chemotherapeutic agents including TKIs [
22,
33]. The decrease in Mcl-1 by ponatinib in CEL cells may facilitate apoptosis, because silencing Mcl-1 with siRNA significantly potentiated the ponatinib-mediated apoptosis in EOL-1 cells, which is in line with the finding that decreased Mcl-1 level sensitizes leukemia cells to tyrosine kinase inhibitors [
22]. Forced overexpression of Mcl-1 protected CEL cells against apoptosis in response to ponatinib. Regarding the mechanism underlying the decrease in Mcl-1 level by ponatinib, our results do not support the involvement of the ubiquitin-proteosome and transcription-dependent pathways. Rather, our data support a caspase-3-dependent mechanism, which is consistent with findings for other small-molecule tyrosine kinase inhibitors [
9]. Of note, the resulting truncated form of Mcl-1
128–350 (p28) cleaved by activated caspase-3 can potentiate apoptosis [
24,
34]. Mcl-1
128–350 (p28) after ponatinib treatment likely produces a positive feedback to apoptosis. Although Bim has been reported as the primary death effector in TKIs-treated CML cells [
35], no appreciable change was noted in ponatinib-treated CEL cells in the present study. However, because Mcl-1 has been demonstrated to neutralize Bim through complex formation to prevent apoptosis, our observed decline in Mcl-1 by ponatinib would increase the Bim/Mcl-1 ratio, which may release Bim to promote apoptosis [
36].
The observed decrease in β-catenin induced by ponatinib exposure may be important because of the fundamental functions of β-catenin in cell proliferation, differentiation and apoptosis. [
25]. Besides accumulating in a Wnt/GSK3β-dependent way, β-catenin may also accumulate after phosphorylation by tyrosine kinases (e.g., Bcr-Abl, RET, KIT, Flt3, PDGFRα) [
26‐
28,
37]. Inactivating PDGFRα by treating EOL-1 cells with ponatinib inhibited tyrosine phosphorylation (Y654) and concurrently decreased levels of β-catenin, as reflected by immunoblotting and immunofluorescent staining (Figure
5). Silencing PDGFRα also lowered β-catenin level, which further supports the specificity of the effect of PDGFRα on the levels of β-catenin. The turnover rate is enhanced in EOL-1 cells pretreated with ponatinib. Of note, inactivation of PDGFRα by ponatinib decreased β-catenin level in both cytosolic and nuclear pools. Ponatinib also decreased β-catenin level in xenografts in in vivo experiments.
The decrease in β-catenin level by ponatinib has functional consequences. First, TCF/LEF-dependent gene transcription was impaired in EOL-1 cells treated with ponatinib. Second, the expression of β-catenin-dependent genes such as c-Myc and cyclin D1 was decreased. Third, the binding of β-catenin and DNA was also decreased, as revealed by EMSA. Because β-catenin plays a crucial role in controlling self-renewal and differentiation in both normal and cancer stem cells [
25], a decrease in β-catenin level may be an important aspect of the antineoplastic mechanism of ponatinib. An interesting hypothesis for future research is whether decreased β-catenin can facilitate the eradication of cancer stem cells.
While this manuscript was under review, Lierman et al. reported that ponatinib was active against imatinib-resistant FIP1L1-PDGFRα mutants [
38]. Our results corroborated and extended their findings by providing a mechanism for the induction of apoptosis and evidence for in vivo efficacy.
Materials and methods
Reagents
Ponatinib (purity > 95%, HPLC) was synthesized in our lab. Imatinib and sorafenib were purchased from Alexis Biochemicals (Plymouth Meeting, PA). 4′, 6-diamidino-2-phenylindole (DAPI) was from Invitrogen. Cycloheximide (CHX) and propidium iodide (PI) were from Sigma-Aldrich. TOPflash/FOPflash system consisting of optimized TCF binding sites (TOP) or mutated sites (FOP) controlling the expression of a luciferase reporter gene was from Upstate Biotechnology (Lake Placid, NY). pCMV5-flag-human Mcl-1 and pcDNA3-β-catenin were kindly provided by Dr. Mien-Chie Hung (The University of Texas MD Anderson Cancer Center, Houston, TX) [
40]. ON-TARGETplus SMARTpool small interfering RNA (siRNA) duplexes against human Mcl-1 or PDGFRα, and Non-Targeting Pool siRNA control were from Dharmacon RNA Tech. (Lafayette, CO) [
8,
41].
Cell culture and cell growth measurement
The EOL-1 cell line harboring the FIP1L1-PDGFRα fusion oncogene was purchased from DMSZ (Braunschweig, Germany). BaF3 cells expressing WT or T674I FIP1L1-PDGFRα were cultured as described previously [
8,
9].
Cell viability was assessed by MTS assay (CellTiter 96 Aqueous One Solution reagent, Promega, Shanghai) [
40,
42].
Clonogenicity assay was performed as described [
40]. In brief, 2×10
5/ml cells were treated with drugs or diluent (DMSO, control) for 24 h, then washed with PBS and seeded in methylcellulose medium (Methocult M3231, Stem Cell Technologies, Vancouver, Canada) [
40]. After incubation for ~7 days at 37°C and 5% CO
2, colonies with >50 cells were counted [
40].
Preparation of whole cell lysates and cytosolic fraction
Most experiments of immunoblotting involved whole lysates prepared with RIPA buffer unless otherwise stated [
42,
43]. To measure the levels of AIF and cytochrome
c in the cytosol, the cytosolic extract was prepared with digitonin extraction buffer [
42,
43].
Preparation of cytoplasmic and nuclear fractions
Distribution of β-catenin was determined in the cytoplasmic and nuclear fractions as we previously described [
44].
Immunoblotting
Immunoblotting involved use of whole cell lysates prepared in RIPA buffer [
8,
9,
40]. Antibodies and their sources were as follows: antibodies against apoptosis-inducing factor (AIF), Mcl-1 (S-19), Bax and Bcl-X
L (Santa Cruz Biotechnology, Santa Cruz, CA); antibodies against poly(adenosine diphosphate [ADP]-ribose) polymerase (PARP), X-linked inhibitor of apoptosis (XIAP), caspase-3, active caspase-3, cytochrome c (clone 6H2.B4), survivin, and C-terminal β-catenin (BD Biosciences Pharmingen, San Jose, CA); phospho-β-catenin (Y654) (Abcam, Cambridge, MA); antibodies against phospho-PDGFRα (Y1018), phospho-Erk1/2 (T202/Y204) and Erk1/2 (Cell Signaling Technology, Beverly, MA); antibodies against phospho-Stat3 (Y705), phospho-Stat5A/B (Y694/Y699), Stat3, Stat5, Bcl-2 and PDGFRα (Upstate Technology, Lake Placid, NY); anti-Bim (Stressgen, Ann Arbor, MI); anti-actin (Sigma-Aldrich, Shanghai). Anti-mouse immunoglobulin G and anti-rabbit immunoglobulin G horseradish peroxidase-conjugated antibodies were from Pierce Biotechnology (Rockford, IL, USA).
Plasmids or small interfering RNA transfection
EOL-1 cells were transfected with plasmids or siRNA duplexes with use of Nucleofector (Amaxa, Gaithersburg, MD) by use of the Cell Line Nucleofector Kit T (Amaxa) and program O-17 [
8]. At 24 h after transfection, EOL-1 cells were adjusted to 2×10
5/ml and exposed to ponatinib treatment, then underwent cell death assay and immunoblotting.
Luciferase assay
EOL-1 cells (2×10
5) were transfected with TOPflash or FOPflash plasmid (0.5 μg) and pEF
Renilla-luc (10 ng) by electroporation. At 24 h, cells were incubated with or without ponatinib for 24 h. Luciferase activity was then measured with the dual-luciferase assay kits (Promega, Shanghai) as described [
45].
Electrophoretic mobility shift assay (EMSA)
EMSA involved the LightShift Chemiluminescent EMSA kit (Pierce Biotechnology, Shanghai). The oligonucleotides for TCF/LEF were from Promega (Shanghai) with sequences as described [
46]: forward, 5-TGCCGGGCTTTGATCTTTG-3; reverse, 5-AGCAAAGATCAAAGCCCGG-3. In brief, oligonucleotides for TCF/LEF were labeled with biotin by use of the biotin 3′-end DNA labeling kit (Pierce Biotechnology, Shanghai). In total, 5 μg of nuclear extracts was incubated for 20 mins with 1 μg/μl poly(dI-dC) and biotin end-labeled target nucleotides in 20-μl reaction mixtures. The resulting bound complex was separated from free oligonucleotides on 6% native polyacrylamide gel and transferred to a nylon membrane. After cross-linking, blocking, and reacting with substrates, the membranes were exposed to X-ray film to detect biotin-labeled DNA. The binding specificity was examined by competition with a 200-fold excess of the unlabeled oligonucleotide probe (cold competitor) [
40,
47].
Transmission electron microscopy
The cells were treated with or without ponatinib, and then fixed with 2% glutaraldehyde plus 2% paraformaldehyde in 0.1 M cacodylate buffer (pH 7.3). After washing, and postfixing, the samples were dehydrated and embedded in Spurr’s low-viscosity medium [
48]. Ultrathin sections of the samples stained with uranyl acetate and lead citrate were examined under a JEM 1010 transmission electron microscope [
48].
Immunofluorescence staining
EOL-1 cells were treated with or without ponatinib for 24 h, and then harvested by use of Cytospin onto glass slides. Immunofluorescence staining was as described [
42]. DyLight 488 conjugated-goat-anti-mouse immunoglobulin was purchased from Pierce Biotechnology (Rockford, IL).
Apoptosis assessment
Apoptosis was evaluated by using an Annexin V-fluorescein isothiocyanate (for EOL-1) or Annexin V-phycoerythrin (for BaF3 cells expressing FIP1L1-PDGFRα) apoptosis detection kit (BD Biosciences Pharmingen, San Jose, CA) and analyzed by using a FACSCalibur flow cytometer [
8,
9,
40].
Cell cycle analysis
Control or ponatinib-treated cells were fixed with 66% ethanol overnight. DNA content was analyzed by flow cytometry after cells were stained with 50 μg/ml PI and 2.5 μg/ml RNase in PBS solution for 30 mins [
42,
43].
Tumor xenograft experiments
Male nu/nu BALB/c mice were bred at the animal facility of Sun Yat-sen University. An amount of 1×107 BaF3-T674I PDGFRα cells supplemented with 50% matrigel was inoculated subcutaneously on the flanks of 4~6-week-old male nude mice. Tumors were measured every other day with calipers. Tumor volumes were calculated by the following formula: a2 × b × 0.4, where a is the smallest diameter and b is the diameter perpendicular to a. Ponatinib was initially dissolved in DMSO and then adjusted to the appropriate doses with vehicle [30% Cremophor EL/ethanol (4:1), 70% PBS], and imatinib was dissolved in sterile double-distilled water. Mice in each group were treated once daily by oral gavage with ponatinib, imatinib or the same amount of vehicle. The body weight, feeding behavior and motor activity of each animal were monitored as indicators of general health. Tumor xenografts were immediately removed, weighed, stored and fixed after animals were killed. All animal studies were conducted with the approval of the Sun Yat-sen University Institutional Animal Care and Use Committee.
Immunohistochemical staining
Formalin-fixed BaF3-T674I FIP1L1-PDGFRα-cell xenografts were embedded in paraffin, sectioned (4-μm thick), then immunohistochemically stained by using the anti-Ki67 MaxVision kit (Maixin Biol, Fuzhou, China). Color was developed with 0.05% diaminobenzidine and 0.03% H
2O
2 in 50 mM Tris–HCl (pH 7.6), and slides were counterstained with hematoxylin [
47].
Homology modeling
The kinase domain sequence was identified from the Human Kinome database [
49] by sequence alignment of the kinase domain within the full-length PDGFR
α sequence (NCBI protein database, GI: 1736333) and the site to be mutated by use of CLUSTAL X [
50].
Prime module in Maestro (Schrödinger Inc., v7.5) was chosen to build homology models for the native kinase domain sequence and the mutated sequence. We performed a BLAST search against the PDB database to choose a suitable template; the 1.6 Å X-ray structure of KIT kinase (PDB code: 1 T46) in complex with imatinib was chosen (identity 61%, E-value 4.1e-65) [
51]. After SSP modification and alignment editing in Prime, native and mutated kinase models were built with default parameter sets, followed by loop refinement and energy minimization to eliminate and correct disallowed torsion angles and unfavorable atom-atom conflicts.
Molecular docking
The compound ponatinib was prepared by the Ligprep module, then the Glide module. Then docking simulations were performed to test binding of ponatinib to the native kinase model and the mutated one with the same default parameter sets. The grid-enclosing box was centered on the centroid of the aligned ligand (imatinib) and defined to enclose residues located within 20 Å around the ATP binding site; a scaling factor of 1.0 was set to van der Waals (VDW) radii of the receptor atoms with the partial atomic charge < 0.25. The Extra-Precision (XP) mode of Glide was used to dock ponatinib into the ATP binding site with default parameters, and the top 10 docked poses were reserved for the binding mode analysis.
Statistical analysis
All experiments were performed at least 3 times, and results are reported as mean ± 95% confidence intervals, unless otherwise stated. GraphPad Prism 5.0 (GraphPad Software, San Diego, CA) was used for statistical analysis. A P < 0.05 was considered statistically significant.
Acknowledgements
This study was supported by grants from the National Natural Science Fund of China (no. 81025021, 81373434, 91213304, 90713036, U1301226 to J. Pan), the National Basic Research Program of China (973 Program grant no. 2009CB825506 to J. Pan), the Research Foundation of Education Bureau of Guangdong Province, China (grant cxzd 1103 to J. Pan), the Research Foundation of Guangzhou Bureau of Science and Technology, China (grant to J. Pan), the National Hi-Tech Research and Development Program of China (863 Program grant no. 2008AA02Z420 to J. Pan).
Competing interest
The authors declare that they have no competing interests.
Authors’ contributions
YJ designed, performed experiments and analyzed data; KD synthesized ponatinib; HL and MX performed molecular docking analysis; XS and CW performed experiments of apoptosis. JP designed, performed research, analyzed data, directed the whole study and wrote the manuscript. All authors read and approved the final manuscript.