|Home | About | Journals | Submit | Contact Us | Français|
Bcr-Abl is the predominant therapeutic target in chronic myeloid leukemia (CML), and tyrosine kinase inhibitors (TKIs) that inhibit Bcr-Abl have been successful in treating CML. With progression of CML disease especially in blast crisis stage, cells from CML patients become resistant to imatinib mesylate (IM) and other TKIs, resulting in relapse. Because Bcr-Abl is known to drive multiple signaling pathways, the study of the regulation of stability of Bcr-Abl in IM-resistant CML cells is a critical issue as a possible therapeutic strategy. Here, we report that a new dual-kinase chemical inhibitor, ON044580, induced apoptosis of Bcr-Abl+ IM-sensitive, IM-resistant cells, including the gatekeeper Bcr-Abl mutant, T315I, and also cells from blast crisis patients. In addition, IM-resistant K562-R cells, cells from blast crisis CML patients, and all IM-resistant cell lines tested had reduced ability to form colonies in soft agar in the presence of 0.5 µM ON044580. In in vitro kinase assays, ON044580 inhibited the recombinant Jak2 and Abl kinase activities when the respective Jak2 and Abl peptides were used as substrates. Incubation of the Bcr-Abl+ cells with ON044580 rapidly reduced the levels of the Bcr-Abl protein and also reduced the expression of HSP90 and its client protein levels. Lysates of Bcr-Abl+ cell lines were found to contain a large signaling network complex composed of Bcr-Abl, Jak2, HSP90, and its client proteins as detected by a gel filtration column chromatography, which was rapidly disrupted by ON044580. Therefore, targeting Jak2 and Bcr-Abl kinases is an effective way to destabilize Bcr-Abl and its network complex, which leads to the onset of apoptosis in IM-sensitive and IM-resistant Bcr-Abl+ cells. This inhibitory strategy has potential to manage all types of drug-resistant CML cells, especially at the terminal blast crisis stage of CML, where TKIs are not clinically useful.
In chronic myelogenous leukemia (CML), Bcr-Abl, the fusion protein derived from the Philadelphia chromosome, is the constitutively activated protein tyrosine kinase, which is largely unregulated.1-3 It is widely known that Bcr-Abl drives several important signaling pathways—the Ras, PI-3 kinase, STAT5, STAT3, and Jak2 pathways that cause oncogenesis in CML.4-10 Since these important pathways are controlled by Bcr-Abl, it is considered the critical target molecule for CML therapy. Imatinib mesylate (IM) is an effective inhibitor of the Bcr-Abl tyrosine kinase and is the first-line treatment of CML since about 75% of early chronic phase CML patients favorably respond to IM treatment. During longer term treatment with IM, progression of the disease and drug resistance can develop in patients for several reasons.11-20 Continuous targeting of Bcr-Abl can lead to blastic transformation21 due to activation of other oncogenes and inactivation of tumor suppressor genes. The remission rate of the accelerated phase is 50%, and for the blast crisis phase, the remission rate is 20%.17,22 Alterations of tumor suppressors such as PP2A, mutation of p53, inactivation of tyrosine phosphatases (Shp1), and overexpression of new proteins (e.g., SET) lead to the terminal blast crisis stage and ultimately death of the patients. More potent forms of IM (i.e., Nilotinib, NS-187) have been developed for the treatment of IM-resistant patients,23 but they fail to kill cells from the blast crisis stage. The dual-kinase inhibitor dasatinib (Bcr-Abl and Lyn) is successful in the induction of apoptosis of several IM-resistant Bcr-Abl mutant cells in blast crisis patients,24 but dasatinib fails to kill T315I Bcr-Abl mutant cells. Dasatinib-resistant CML has been reported, as 20 of 21 patients treated with dasatinib developed resistant CML cells containing the T315I mutation.25,26 Several other second-generation drugs were developed for CML therapy, but each drug has its own limitations.27 Although overcoming IM resistance can be achieved for some forms of IM resistance caused by mutations in BCR-ABL, specific drugs for the T315I BCR-ABL IM-resistant mutant have not yet been developed, nor are drugs available to treat blast crisis CML. The untreated chronic phase may last for several years, the accelerated stage lasts for only 4 to 6 months, and the terminal blast crisis stage, characterized by rapid expansion of either myeloid or lymphoid differentiation-arrested blast cells (blast crisis), lasts for only a few months.17,18 No successful therapeutic strategy of blast crisis exists at the present time. Allogeneic stem cell transplantation with high chemotherapy has been found to be successful in a small percentage (10%) of patients. New target molecules and specific inhibitor(s) need to be developed to treat advanced stages of CML, particularly in blast crisis patients.
Since Bcr-Abl is considered the primary therapeutic target molecule in CML, the stability and regulation of Bcr-Abl in CML cells is one of the critical issues for development of new therapeutic strategies required to overcome drug resistance. Neviani et al.28 demonstrated that Bcr-Abl regulates its own stability by inhibiting PP2A-Shp1 phosphatases by inducing expression of tumor suppressor protein SET.28,29 Our previous studies demonstrated that Jak2 is a major downstream signaling molecule in CML. It has been shown that Jak2 interacts with Bcr-Abl,9 induces high-level c-Myc expression,30 induces tyrosine phosphorylation of Gab2 on YxxM sequences needed for activation of PI-3 kinase,31 is part of a Bcr-Abl network involving proteins such as Akt and GSK3β,31 and regulates SET protein in Bcr-Abl+ cells.32 Jak2 also maintains Lyn kinase in its functionally active form in Bcr-Abl+ cells through a Jak2-SET-PP2A-Shp1 signaling loop where PP2A-Shp1 remained inactive by Jak2-activated SET expression.32 These results indicate that Jak2 is one of the important signaling molecules in Bcr-Abl+ cells.
HSP90, a major molecular chaperone, is known to interact with proteins involved in transcriptional regulation and signal transduction pathways for maintaining the stability and functional conformation of signaling proteins.33-36 HSP90 acts as a biochemical buffer against genetic instability during cancer. HSP90 is responsible for the maturation and functional stability of a plethora of polypeptides called client proteins. HSP90 is overexpressed in leukemia and also in many other cancers, and it is assumed that in cancer, the requirement of HSP90 is critical since most of the client proteins of HSP90s are active participants in signal transduction pathways of cancer cells.33,36-38 These qualities and functional aspects of HSP90 make it a potential target for anticancer drugs. Although several small molecules have been identified as anti-HSP90 candidates during past years, none of them has yet been successful in the clinic.39,40 Gorre and colleagues14 first showed that inhibition of HSP90 expression by 17-AAG caused reduction of wild-type and mutant Bcr-Abl proteins, leading to inhibition of growth. Later, Blagosklonny et al.41 demonstrated that BCR-ABL+ cells were induced to undergo apoptosis upon treatment with 17-AAG. These qualities and functional aspects make HSP90 a potential target for the development of anticancer drugs.
In the current study, we have shown that ON044580 shows strong apoptotic activities in Bcr-Abl+ cells and overcomes drug resistance. These apoptotic events were initiated in part due to destabilization of the Bcr-Abl protein from where major signaling pathways originate. We have further demonstrated that ON044580 disrupted a high molecular weight Bcr-Abl/Jak2/HSP90 network structure. These results were obtained due to the unique Jak2 and Bcr-Abl kinase inhibitory properties of ON044580, which make it a novel and potentially useful compound for CML therapy.
ON044580, α-benzoyl styryl benzyl sulfide, is a new compound synthesized by Dr. Reddy’s group42 that is not an adenosine triphosphate (ATP) competitor like many of the tyrosine kinase inhibitors such as IM but inhibits the catalytic activities of Abl (and Bcr-Abl) and Jak2. We present results on the role of ON044580 in modulating Bcr-Abl-driven cell signaling pathways and its effects on cell viability, apoptosis, and colony formation in soft agar.
To examine the effects of ON044580 on Abl and Jak2 kinases, we performed in vitro kinase assays with purified recombinant Abl (45-kDa Abl kinase) and Jak2 kinase (JH1-JH2) using Abl tide substrate for assays with Abl kinase and Jak2 peptide containing the Tyr 1007 activation site for the Jak2 kinase, respectively. IM inhibited the phosphorylation of Abl tide by recombinant Abl about 85%, whereas ON044580 at 5 µM and 10 µM reduced the Abl kinase activity by 50% and 75%, respectively (Fig. 1a). In the Jak2 kinase assay with JH1-JH2 domains, ON044580 strongly reduced Jak2 kinase activity in a dose-dependent-manner (Fig. 1b). As a positive control TG101209, an authentic Jak2 inhibitor43 was used that strongly reduced phosphorylation of the Jak2 peptide. These studies indicate that both recombinant Abl kinase and Jak2 kinase are strongly inhibited by ON044580, suggesting that ON044580 is a dual-kinase inhibitor (Figs. 1 a and andbb).
To further investigate the effects of ON044850 on the Jak2 kinase, we performed in vitro autophosphorylation assays of Jak2 using Bcr-Abl+ cell lysates. Our previous findings indicate that Jak2 is associated with the C-terminus of Bcr-Abl.9 On the basis of that observation, for the Jak2 kinase assay, we immunoprecipitated Bcr-Abl from detergent-extracted Bcr-Abl+ 32D cell lysates with Abl-specific antibody (P6D). After repeated washing of the immunoprecipitates, the kinase assays were performed using the protocol described for Jak2 kinase.9,44 The kinase supernatant was analyzed by Western blotting using anti-pTyr (4G10) to detect tyrosine-phosphorylated P210 BCR-ABL (Fig. 1c) and anti-pJak2 (Tyr1007/1008) to detect activated Jak2 (Fig. 1d). We observed that both Bcr-Abl kinase and Jak2 kinase activities were reduced in the presence of ON044580 (Figs. 1 c and anddd).
We incubated Bcr-Abl+ IM-sensitive (BaF3p210) and IM-resistant cells (BaF3p210 T315I and BaF3p210 E255K cells) with different doses of ON044580 for 16 hours. Cell lysates were prepared by detergent extraction, and the lysates were analyzed by Western blotting using anti-pTyr antibody (4G10). We observed that the levels of both pTyr Jak2 and pTyr Bcr-Abl were sharply reduced with 16-hour incubation (Fig. 1e). However, the Bcr-Abl protein was found to rapidly disappear from the lysate within 2 hours of 10 µM ON044580 treatment, whereas Jak2 protein levels were not affected during these 2-hour treatments. The dose needed to reduce the Bcr-Abl protein levels began at 2.5 µM and was complete at 10 µM (Suppl. Figs. S1 a and b). These studies indicate that treatment of Bcr-Abl+ cells with ON044580 may affect either the stability or solubility of Bcr-Abl.
We addressed the question of whether treatment of Bcr-Abl+ cells with ON044580 affected downstream signaling molecules of Bcr-Abl. To examine this possibility, we incubated Bcr-Abl+ 32D cells for 6 hours using 10 µM ON044580 and for 16 hours with increasing amounts (0-10 µM) of the inhibitor. The detergent-extracted lysates were analyzed by Western blotting using several antibodies. We observed that in addition to the reduction of Bcr-Abl, pTyr Jak2, STAT3, and Akt levels were also reduced during 6-hour incubation of Bcr-Abl+ cells with ON044580 (Fig. 2a). We further observed that a 16-hour incubation of Bcr-Abl+ cells with ON044580 reduced not only Jak2 and STAT3 levels but also pTyr705 and pSer727 STAT3 levels. Interestingly, Lyn was unaffected (Fig. 2b). It is known that Bcr-Abl, Jak2, and STAT3 are the client proteins of HSP90,45-48 but Lyn has not been reported to be a client protein of HSP90. Thus, our results also suggest that Lyn is not a client protein of HSP90.
It is known that tyrosine phosphorylation of STAT3 plays a key role in the dimerization of STAT3, nuclear translocation, and binding to specific DNA consensus sequence of STAT3, whereas serine phosphorylation of STAT3 is essential for maximum transcriptional activity.49,50 Since Tyr 705 STAT3 phosphorylation was reduced by ON044580, it was expected that DNA binding of STAT3 to its consensus sequence would be interrupted. Therefore, we examined the binding of STAT3 to its consensus sequence by electrophoretic mobility shift assays (EMSA). STAT3, obtained from nuclear extracts of ON044580-treated Bcr-Abl+ 32D cells (16 hours), was allowed to interact with its radiolabeled consensus STAT3 oligonucleotide DNA sequence.51 Bcr-Abl+ cells treated with ON044580 had strongly reduced the STAT3-specific DNA binding activity in a dose-dependent manner (Fig. 3a). The assay signal for STAT3 is specific because competition with nonradioactive consensus sequences strongly competed with the radioactive target oligonucleotides in a dose-dependent manner (Fig. 3b, right panel). Similarly, addition of STAT3 antibody to the nuclear lysate caused a mobility shift of the STAT3 complex (not shown), indicating that the signals for STAT3 in EMSA (Fig. 3a) are STAT3 specific.
HSP90 is reported to be a chemotherapeutic target molecule for many cancers, including CML.35,36,48,52 Some of the critical signaling molecules in Bcr-Abl+ cells are client proteins of HSP90.14,47,3 We examined whether ON044580 regulated the expression of HSP90 at the transcriptional level. For this, we performed RT-PCR assays using HSP90 primers. We treated 32Dp210 cells with ON044580 for 16 hours. We note that the HSP90α promoter has a binding site for STAT3 (not shown). Of interest, ON044580 at 10 µM strongly reduced HSP90α transcripts at 16 hours of treatment (Fig. 3c), which coincides with the amount of ON044580 required to inhibit STAT3 binding to its consensus sequence (Fig. 3a). HSP90α protein levels in IM-sensitive and IM-resistant cells were also reduced by incubation of cells with 5 and 10 µM ON044580 for 16 hours. However, T315I cells were partially resistant to HSP90 reduction by ON044580 at 16 hours despite the high sensitivity to ON044580 to reduction of activated STAT3 (Fig. 3d). Nevertheless, these results suggest that Jak2 kinase may regulate expression of HSP90α through Jak2’s ability to activate STAT3 in Bcr-Abl+ cells (Figs. 3 a--ee).
From our previous studies with various co-immunoprecipitation experiments, we showed that immunoprecipitation of one member of the Bcr-Abl signaling pathway co-precipitated other members of the pathway. Therefore, we predicted the presence of a large molecular network complex in Bcr-Abl+ CML cells.31 To identify, characterize, and estimate the relative size of the Bcr-Abl/Jak2 network complex, we performed gel filtration column chromatography as a means to determine whether the Bcr-Abl/Jak2 network complex could be detected in a high molecular weight region of the column eluant. In collaboration with our Proteomics Core Facility, we optimized and calibrated the gel filtration column with different marker proteins ranging up to 8 million molecular weight (Suppl. Fig. S1e). Cell lysates of Bcr-Abl+ 32 D cells (32Dp210) were fractionated on the gel filtration column and eluted with a buffer containing NP-40 and glycerol. Fractions were analyzed by Western blotting with various antibodies so as to detect several proteins thought to be present in this network complex (Fig. 4a). We detected several signaling proteins, including HSP90, in the same fractions of the column eluant (e.g., fraction 12), suggesting the presence of high molecular weight protein complexes (Fig. 4a), which were estimated to be in the 4 to 6 million Da molecular size fraction. The Bcr-Abl/Jak2 network proteins included pTyrJak2 (1007/8), pLyn (Tyr 396), Lyn, Akt, STAT3, GSK3β, pErk, and HSP90; several column fractions contained these high molecular weight complexes (Fig. 4a and Suppl. Fig. S1e). Similar results were obtained with lysates of K562 cells (Suppl. Fig. S1d). The decrease in levels of Bcr-Abl and several other signaling proteins by treatment with ON044580 (Fig. 2a) suggested that this dual-kinase inhibitor might disrupt the network structure. To determine whether the elution pattern of the network would be affected by ON044580 treatment, we incubated 32Dp210 cells with 10 µM ON044580 for 3 hours and loaded the cell lysate into the column. We observed that the Bcr-Abl/Jak2/HSP90 network complex was disrupted, as Bcr-Abl protein was severely reduced in amount, as were other members of the network. Importantly, HSP90 and other the client proteins eluted at a much lower molecular size (Fig. 4b). Although the levels of Jak2, STAT3, and Akt were reduced in the column fractions of ON044580-treated lysates, the levels of HSP90 remained almost unchanged but eluted at a much lower molecular size, as the position of the HSP90 protein shifted from elution at the higher molecular weight fractions (e.g., 12-15) to the lower size fractions (fractions 24-27), indicating that network had been disrupted. These results suggest the following: (1) that the Bcr-Abl/Jak2 network is bound to HSP90 and (2) that decrease in Bcr-Abl and inhibition of both Bcr-Abl and Jak2 kinases lead to disruption of the network structure by separation of Bcr-Abl and Jak2 from its signaling partners. We hypothesize that HSP90 client proteins such as Bcr-Abl are more susceptible to proteolytic degradation when the network structure is disrupted by treatment with ON044580. Under identical conditions, lysates of Bcr-Abl+ 32D cells treated with 10 µM imatinib for 6 hours did not show degradation/dissociation of signaling molecules (Suppl. Fig. S1c). A hypothetical model for disruption of the network by ON044580 is shown in Figure 4c.
Our studies demonstrate that ON044580 strongly inhibits Jak2 and Abl kinase activities, and as a result, the levels of downstream signaling molecules are reduced, and the large Bcr-Abl/Jak2/HSP90 network complex is disrupted. We next examined how these inhibitory effects on the Bcr-Abl/Jak2/HSP90 network structure affected cell survival. For that purpose, we did cell viability/proliferation assays (MTT), apoptosis assays, and colony formation assays. We first assessed the effects of ON044580 on cell viability and proliferation by MTT assays. IM-sensitive Bcr-Abl+ cells (32Dp210) and IM-resistant cells (e.g., K562R) were inhibited by ON044580, as the viability was reduced in a dose-dependent manner (the IC50 of ON044580 for 32Dp210 and K562R cells was 3-5 µM; Suppl. Figs. S2 a and b). Apoptosis assays on several Bcr-Abl+ IM-sensitive and IM-resistant hematopoietic cell lines (Figs. 5 a--cc and Suppl. Figs. S2 c and d, S3 a and b) were conducted by staining with annexin and propidium iodide followed by flow cytometric analysis. Results from this study showed that ON044580 was a potent inducer of apoptosis at concentrations of 1 to 5 µM. IM-sensitive Bcr-Abl+ cells (32Dp210) and BaF3-p210 cells were very sensitive to ON044580 to apoptosis induction, and 5 µM of ON044580 induced >80% apoptosis. IM-resistant cells such as T315I mutant cells and E255K and K562-R cells, although resistant to IM, were very sensitive to apoptosis induction by ON044580 (Fig. 5b). The T315I mutant is termed the gatekeeper mutation,11 and all known kinase inhibitors that target the ATP binding domain of the Bcr-Abl tyrosine kinase fail to induce apoptosis in T315I cells (Fig. 5b). Therefore, it is quite significant that ON044580 induced apoptosis in T315I mutant cells. Similar results were obtained with the E255K IM-resistant mutant of Bcr-Abl (not shown).
After examination in IM-sensitive and IM-resistant Bcr-Abl+ cell lines, we tested the ability of ON044580 to kill cells from blast crisis CML patients, which are largely resistant to many drugs. As can be seen in Figures 5 d--ff and Supplement Figures S3 a-e, white blood cells from the peripheral blood of blast crisis CML patients are quite resistant to IM (5 and 10 µM) but are very sensitive to ON044580. Most interestingly, primary CML cells are very sensitive to low doses (1-2 and 0.5 µM) of ON044580. We observed that blast crisis patient cells, some of which are resistant to IM, are induced to undergo apoptosis by ON044580 with values ranging from 70% to 90% (Figs. 5 d--ff and Suppl. Fig. S3 c-e).
Anchorage-independent growth is a cell culture surrogate for tumor behavior in mice. We assessed the ability of ON044580 to inhibit colony formation in soft agar cultures. Cells were seeded into soft agar culture medium at the single cell level. Cultures were allowed to incubate for two weeks in the presence of different doses of ON044580. Colonies were stained, photographed, and counted to assess the remaining colony number after the drug treatment. Cells that were both IM sensitive and IM resistant were tested (Figs. 6 a--d).d). In general, colony formation was completely inhibited at 0.5 µM ON044580. Importantly, IM-resistant forms of Bcr-Abl+ cells were also inhibited at similar concentrations (Figs. 6 c and andd).d). The results showed that ON044580 severely inhibited colony formation at levels between 0.1 and 0.5 µM (right panels of Figs. 6 a--d).d). These results suggest that oncogenic ability of IM-sensitive and IM-resistant Bcr-Abl+ cells are inhibited by ON044580 at lower concentrations compared to the concentrations required for apoptosis and MTT assays.
In this article, we investigated the mode of action and functional properties of a new non-ATP competitive kinase inhibitor, ON044580, in Bcr-Abl+ mouse hematopoietic cell lines, IM-resistant cell lines, and cells from blast crisis CML patients. Our studies (see Fig. 1) and those of Jatiani et al.42 indicate that ON044580 is a dual-kinase inhibitor that inhibited both Bcr-Abl and Jak2 kinases. Importantly, ON044580 induced apoptosis in IM-sensitive and IM-resistant cells and cells from the late stage of CML patients (Figs. 5 a--e).e). Our findings further showed that ON044580 induced rapid disappearance of Bcr-Abl protein from the detergent-soluble fraction of leukemic cells (Fig. 2a and Suppl. Fig. S1a), which affects downstream signaling of Bcr-Abl (Figs. 2 a and andb)b) and disrupts the Bcr-Abl/Jak2/HSP90 network complex (Figs. 4 a--c).c). The rapid disappearance of Bcr-Abl from Bcr-Abl+ cells caused by ON044580 makes it a novel compound with potential for clinical application in CML. The possible mechanism of a rapid decrease of Bcr-Abl protein by ON044580 is not yet established, but preliminary experiments with a potent Jak2 inhibitor suggest that Jak2 inhibition only is sufficient for rapid disappearance of Bcr-Abl from the detergent-soluble fraction of cell. Since proteosomal inhibitors (MG132 and lactocyteine) failed to protect the rapid disappearance of Bcr-Abl from the detergent-soluble fraction within 2 to 4 hours (data not shown), we predict that upon inhibition with ON044580, Bcr-Abl and Jak2 dissociate from the network complex, and Bcr-Abl rapidly migrates to a detergent-insoluble compartment of the cell. How ON044580 treatment may accomplish this is under study in our laboratory. Nevertheless, upon dissociation of Bcr-Abl and Jak2 from the network complex, oncogenic signaling would be greatly reduced, and the leukemogenic properties of CML cells would similarly be greatly reduced.
The Bcr-Abl/Jak2 dual-kinase inhibitory effects of ON044580 are a critical aspect of this compound. Thus, unlike IM, where resistant mutations arise in BCR-ABL, ON044580 has the capacity to also inhibit the Jak2 kinase, which induces apoptosis in IM-resistant Bcr-Abl mutant cells, including those expressing the gatekeeper mutant T315I (Fig. 5). In addition, signals produced from both Bcr-Abl and Jak2 in IM-sensitive cells will be downregulated (Figs. 2 a and andb)b) by treatment with ON044580.
In our previous study, we reported that Bcr-Abl is associated with several signaling proteins and forms a signaling network that includes Jak2, Gab2, Akt, and GSH3β.31 In these studies, we showed in co-immunoprecipitation experiments that Bcr-Abl was associated with various members of its downstream signaling targets. For example, immunoprecipitation of Bcr-Abl+ cells with anti-Jak2 detected Jak2, Akt, GSK-3β, and Bcr-Abl. In addition, immunoprecipitation with antibody against GSK-3β co-precipitated Bcr-Abl, and immunoprecipitation with an Akt antibody also co-precipitated Bcr-Abl. Normal serum controls established the specificity of these co-immunoprecipitation experiments. We concluded that Bcr-Abl and members of the signaling network described by Samanta et al.31 were present in a network complex. The gel filtration experiments in Figure 4 and Supplement Figure S1 support this conclusion and indicate that the network is quite large in terms of molecular size, possibly more than 6 million Da.
It is known that HSP90 is a therapeutic target molecule for solid tumor cancers and CML.14,54 It is also reported that the critical signaling molecules, such as Bcr-Abl, Jak2, Akt, pErk, and STAT3, are physically associated with HSP90,14,47,55,56 which plays a role in their conformational maturation and functional performance and also provides protection from proteases.47,56,57 Any disturbance of HSP90 synthesis will eventually lead to proteolytic degradation of the client proteins.58-60 Our studies in Bcr-Abl+ cells indicate that ON044580 treatment, because of its ability to inhibit both Jak2 and Bcr-Abl kinases (Figs. 1 and and4d),4d), led to inhibition of STAT3, and since STAT3 appears to control HSP90α transcription, this eventually leads to a decrease in HSP90α transcription and decreased HSP90α protein levels (Figs. 3 c--e).e). Whether STAT3 is the direct cause of HSP90 transcription is not known at this point, but further experiments are planned to clarify the pathway from STAT3 to HSP90. Importantly, we noticed that the promoter of HSP90α contains binding sites for STAT3 (at -125 bases; our observation) and also for NF-κB.61 Jak2 inhibition also leads to downregulation of NF-κB.31 We also note that leukemia cells express predominantly the HSP90α form.41,62 However, the reduction of HSP90α protein levels follows the initial inhibitory event of ON044580, which is the decrease in Bcr-Abl levels within 2 to 3 hours of ON044580 treatment. Our findings also showed that Bcr-Abl, Jak2, and HSP90α exist in a high molecular weight network structure (estimated to be about 6 million in size) (Fig. 4a) (a similar structure was seen in K562 cells; Suppl. Fig. S1d) that houses a number of other signaling proteins, including Akt, Erk, GSK3β, and STAT3 (Fig. 4a). Treatment of Bcr-Abl+ 32D cells with ON044580 for 3 hours caused destruction of this large network structure (Fig. 4b). It is of interest that IM treatment of Bcr-Abl+ 32D cells had little effect on the Bcr-Abl/Jak2/HSP90 network complex during a 6-hour treatment (Suppl. Fig. S1c). Thus, these results suggest that Jak2 inhibition by ON044580 is the critical inhibitory activity caused by ON044580 that leads to rapid destruction of the Bcr-Abl/Jak2/HSP90 network complex. However, ON044580, by also decreasing levels of Bcr-Abl, might contribute to the rapid destabilization of the network complex. It is likely that this destruction of the Bcr-Abl/Jak2/HSP90 network structure will induce apoptosis and death of the leukemia cell.
Our mechanistic studies (Fig. 4c) regarding ON044580 treatment of Bcr-Abl+ cells is likely to explain why it induces apoptosis in the IM-sensitive and IM-resistant Bcr-Abl+ cells such as Bcr-Abl mutant T315I and E255 cells, as well as the IM-resistant CML cell line (K562-R) and drug-resistant CML blast crisis patient cells (Figs. 5 a--e).e). We have proposed a model that describes the roles of Bcr-Abl and Jak2 in signaling pathways that operate in CML cells and the effects of ON044580 (Fig. 4d). In this model, our findings suggest that both Bcr-Abl and Jak2 have important roles in activating STAT3, and importantly, ON044580 treatment of leukemia cells will downregulate both Bcr-Abl and Jak2 kinase-induced effects. The reduction of STAT3 will lead to reduced transcripts of HSP90α, which in turn will reduce HSP90α protein levels. Although these inhibitory effects on HSP90 protein expression probably play an important role in the final apoptotic consequences of ON044580 treatment, we emphasize that the rapid reduction of Bcr-Abl protein levels and the inhibition of the Jak2 kinase are the primary events that initiate the destruction of the Bcr-Abl/Jak2/HSP90 signaling complex. On the basis of our findings, we propose that targeting Jak2/Bcr-Abl/HSP90 is an excellent strategy for inducing apoptosis in drug-resistant CML cells of all types, including advanced stages of CML-like blast crisis, and thus ON044580 may have potential for treatment of many forms of drug-resistant CML.
In all our major experiments, we used IM-sensitive Bcr-Abl transformed mouse cell lines 32Dp210 and BaF3p210 cell lines. For examination in IM resistance, we used Abl point mutant cell lines 32D/BaF3-T315I and BaF3-E255K and Lyn upregulated human CML line K562-R. All the cell lines were cultured in RPMI 1640 supplemented with penicillin and streptomycin and 10% fetal bovine serum. K562-R cells were cultured in 5 µM imatinib.
The cells were washed with cold phosphate-buffered saline (PBS) followed by washing with low salt buffer (pH 7.4), 25 mM NaCl, and 1 mM DTT. The cells were lysed in lysis buffer (1% NP-40 in 20 mM HEPES [pH 7.4]), 150 mM NaCl, 1 mM EDTA, and protease inhibitors 1 mM AEBSF/PMSF, 2 µg/mL aprotinin, 2 µg/mL leupeptin, 50 mM NaF, and 500 µg/mL benzamidine.
For immunoprecipitation, 1 µg primary antibody was added to 400 µg cell extract and mixed with 400 µL lysis buffer for 90 minutes at 4°C. After that time period, 35 µL 50% suspension of protein A/G agarose was added, and the rotation was continued for another 90 minutes. After that incubation, the agarose beads were washed twice with lysis buffer and twice with low salt buffer. Then the beads were treated with 30 µL 2x sample buffer and boiled for 5 minutes. For direct Western blot, 50 µg cell lysate was mixed with an equal amount of 2x sample buffer and boiled for 5 minutes, loaded into wells, and separated in 4% to 20% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gradient gel using protein markers. The proteins were transferred onto a PVDF membrane and blocked by either 3% bovine serum albumin (BSA; for detection of pTyr) or 5% nonfat dry milk diluted in TRIS-buffered saline with tween-20 (TBST). Primary antibodies were diluted in TBST (at room temperature; monoclonal) or in milk (polyclonal) and incubated for 1 hour. After washing 3 times with TBST, the appropriate secondary antibodies conjugated to horseradish peroxidase were incubated at room temperature for 1 hour. After washing, the membrane was developed using ECL/ECL plus reagents (Amersham Biotech Company, Piscataway, NJ). β-Actin (Sigma, St. Louis, MO) was used as a loading control.
EMSA was carried out following the method of Samanta et al.51 After treatment with the appropriate inhibitors, cytosolic and nuclear extracts were prepared. The nuclear extract was used either immediately after preparation or stored at −70°C. For each EMSA, 6 to 8 µg of nuclear extract protein was incubated with poly (dI:dC), 10% NP-40, and 32P-labeled consensus oligonucleotide was annealed to make it double stranded. From ON044580-treated cells, nuclear extracts were prepared and incubated with gamma radiolabeled 32P STAT3 consensus DNA binding site prepared as above at 37°C. Then the whole contents were separated in a 6.6% polyacrylamide gel.
(5′-AGCTTCATTTCCCGTAAATCCCTA-3′ consensus STAT3 binding site
Autoradiography of the dried gel provided the results of the experiments. Since there were inhibitory effects of ON044580 for this binding, the amount of STAT3 protein bound to the radioactive consensus sequence was reduced in a dose-dependent manner in ON044580-treated leukemia cells. DNA binding specificity was determined by competition with unlabeled wild-type or mutated oligonucleotides. The mixture was incubated at 37°C for 15 minutes, and the reaction was terminated by addition of 4 µL 2x DNA loading dye and analyzed in a 6.6% polyacrylamide gel. After autoradiography, STAT3-DNA complexes were detected using a PhosphoImager (Molecular Dynamics, Sunnyvale, CA).
From Bcr-Abl+ CML cells, total cellular RNA was prepared by the TRIzol method following the manufacturer’s protocol (GIBCO, Carlsbad, CA). RT was carried out using 500 ng total RNA in a first-strand cDNA synthesis reaction with superscript reverse transcriptase as recommended by the manufacturer (Invitrogen, Carlsbad, CA). The sequence for HSP90α is as follows: forward 5′-GCGGCAAAGACAAGAAAAAG-3′ and reverse 5′-CAAGTGGTCCTCCCAGTCAT-3′. GABDH was used as an internal control. The sequences for GABDH are as follows: forward 5′-CATGATGGCTTCCTTAGA TGCCCAG-3′ and reverse 5′-CCGTGTGTCATGTAG TGAACCTTTAAG-3′, and an expected product size was 316 bp. PCR reaction was carried out by adding 1 µL RT product into a 25-µL volume reaction mixture containing 1x buffer and 200 µM of each dNTPs, oligonucleotide primer, and 0.2 U AmpliTaq polymerase. For amplification of DNA, cDNA was denatured at 94°C for 1 minute and subjected to primer annealing at 60°C for 1 minute, followed by DNA extension at 72°C for 1 minute for 30 cycles in a thermal cycler (Applied Biosystems, Foster City, CA). Amplified products were analyzed by DNA gel electrophoresis in 1% agarose and visualized by the Alpha Imager 3400 (Alpha Innotec, Santa Clara, CA).
The protein separation column selected for this purpose was 50 cm length × 0.7 cm diameter (Econo column, Bio-Rad, Hercules, CA), and the column material selected for this purpose was Superose 6 prep grade gel filtration (Amersham-Biosciences, part of GE Healthcare, Piscataway, NJ), which can achieve high-resolution separations across an exceptionally broad molecular weight range. The bed volume of the column was 17.5 mL, and the void volume was 6.0 mL. The composition of the elution buffer was 30 mM HEPES (pH 7.4) containing 150 mM NaCl, 10% glycerol, and 0.5% NP-40. Elution rate was 4.56 mL/h. The column was standardized with the mixture of protein markers containing keyhole limpet hemocyanin (KLH; MW 8.5 million Da), blue dextran (2 million Da), β-amylase (200 kDa), BSA (66 kDa), and cytochrome C (12.4 kDa) (Suppl. Fig. S1e). The fractions were collected in 500-µL microfuge tubes in a fraction collector. The elution of the markers detected in 280 nm was plotted against the log of the molecular weight of the standard proteins. From this standard elution pattern, the size of the Bcr-Abl protein network was estimated to be between 2 and 6 million. In a preequilibrated column, we loaded 150 µL (~3 mg) protein onto the column, and the proteins were separated into 40 tubes, each containing approximately 500 µL column eluant. All the column fractions were stored at −20°C. From each column fraction, 25 µL was taken for analysis by Western blotting with various antibodies. Elution analysis of fractions 8 to 24 were performed in 3 premade gradient SDS-PAGE gels (4%-20%). The proteins were transferred to PVDF membranes. The membranes were blocked with BSA for detection of pTyr, and for detection of Bcr-Abl and other proteins, blocking was carried out with 5% milk, and Western blot was carried out as described earlier.
A detergent-extracted cell-free lysate was prepared from the Bcr-Abl-positive cell lines 32Dp210 or K562. Lysis buffer used for this cell-free extract was 1% NP-40 in 20 mM HEPES (pH 7.4), 150 mM NaCl, 1 mM EDTA, and protease inhibitors 1 mM PMSF, 2 µg/mL aprotinin, 2 µg/mL leupeptin, 50 mM NaF, 1 mM sodium venadate, and 500 µg/mL benzamidine. The cells were usually incubated with the lysis buffer for 30 minutes in ice with mild mixing with a vortex mixer. The cell lysate was centrifuged at 4°C for 10 minutes. The supernatant was removed and kept in a separate tube, and protein concentration was measured by a Bio-Rad reagent (Bio-Rad) using BSA as a standard (Pierce, Rockford, IL).
Jak2 kinase assay was carried out following the methods of Xie et al.9 and Sandberg et al.44 Cell-free lysate of 32Dp210 was prepared by treating the cells with lysis buffer containing 20 mM Tris-HCl, 100 mM NaCl, 1% NP-40, and protease inhibitors. Detergent-extracted lysate was aliquoted to each Eppendorf tube containing 500 µg protein/500 µL lysis buffer and prescreened with protein G agarose conjugate. The supernatant was incubated with 50 µL Abl antibody (P6D) for 1 hour followed by 30 µL protein G agarose beads for another 1 hour for co-immunoprecipitation for Bcr-Abl/Jak2. After washing with lysis buffer followed by washing with kinase buffer (containing 50 mM HEPES [pH 7.6], 100 mM NaCl, 5 mM MgCl2, 5 mM MnCl2), the agarose beads were suspended in kinase buffer. Different amounts of ON044580 were added and incubated for 10 minutes, and the reaction was initiated by addition of 2.5 mM ATP. The reaction was continued for 30 minutes at room temperature, and the reaction was stopped by addition of 2x sample buffer. The signals for kinase reaction were detected in Western blotting with pJak2 Tyr1007/1008 antibody.
Autophosphorylation of Bcr-Abl kinase was performed following the method of Bartholomeusz et al.63 by immunoprecipitating Bcr-Abl with P6D antibody. Immunoprecipitates were incubated with various amounts of ON044580. Kinase reactions were initiated with addition of cold ATP, Mg++, and 1 mM dithiothreitol (DTT) at 30°C for 30 minutes. Kinase activity was detected by Western blotting with anti-pTyr antibody (4G10).
Recombinant Jak2 kinase and Abl kinase were assayed in vitro following modified methods.
Recombinant Jak2 kinase assay: Recombinant Jak2 (JH1 and JH2 domains) was preincubated for 10 minutes with different amounts of ON044580 in an incubation mixture as described above for the cold kinase assay. After 10 minutes, the reaction was initiated with cold ATP (5 µM), 10 µCi/assay 32P gamma ATP, and 5 µM Jak2 peptide substrate (994-Asp-Phen-Gly-Leu-Thr-Lys-Val-Leu-Pro-Glu-Lys-Glu-Tyr1007-Tyr1008-Lys-Val-Lys-Glu-Pro-Gly-Glu-Ser-Pro-Iso-Phen-1019) as originally described,9 and the incubation was continued for 10 minutes at 30°C. The reaction tubes were kept in ice, 250 µg BSA was added, and finally an equal volume of trichloroacetic acid (TCA, 40%) was added and incubated for 30 minutes. After centrifugation, the pellet was washed with 20% TCA twice, and the pellet was used for counting 32P gamma ATP incorporated in the pellet in a gamma counter.
Recombinant Abl kinase assay: For Abl kinase assay, 20 ng recombinant Abl kinase was mixed with the same kinase buffer used for the Jak2 kinase assay. A different amount of ON044580 was added to the incubation mixture and preincubated for 10 minutes. The reaction was initiated by adding the substrate for Abl kinase (5 µM Abltide), 5 µM unlabeled ATP, and radiolabeled ATP (10 µCi). The reaction was stopped by addition 5 µL of 3% phosphoric acid from the mixture, and 10 µL of the mixture was dropped on Whatman filter paper (2.5-cm disk, grade p81) in triplicate. The disks were washed in 180 mM phosphoric acid 2 times with shaking and finally washed with 100% methanol. Then the count present in each disk was measured by a scintillation counter.
Apoptosis measurement was carried out following the manufacturer’s protocol using the Annexin V/PI method in a flow cytometer (BD Pharmingen, San Diego, CA). The cells were incubated with either IM or ON044580 in different doses for 48 hours. Then the cells were processed for measurement of apoptosis following the manufacturer’s protocol (BD Pharmingen).
Colony formation assay was carried out following the method described.28
Cells from CML donors were obtained under an approved institutional protocol. CML cells were separated by centrifugation through Histopaque 1077 (Sigma), and the cells were suspended in RPMI medium with 10% fetal bovine serum for 48 hours in the presence of 5 and 10 µM imatinib and 2.5 to 10 µM ON044580 and incubated for 48 hours. Cells were processed for flow cytometry with Annexin V and propidium iodide staining to measure late-stage apoptosis.
The authors are thankful to Dr. Kobayashi of Molecular Pathology Dept. for his help in the optimization and calibration of the gel filtration column and Dr. Susmita Samanta, Dept. of Genetics, Institute of Molecular Medicine for her help in some of the experiments.
Dr. E.P. Reddy is a stockholder, board member, and consultant for Onconova Therapeutics, Inc. Drs. Samanta, Chakraborty, Wang, Schlette, and Arlinghaus declare no conflicts of interest.
This work was supported in part by grants from the DOD (W91ZSQ-5309-N7 [RBA] and W81XWH-06-1-0267 [EPR]) and Leukemia Spore grant CA100632 (AKS), Ladies Leukemia League (AKS), and the National Heart, Lung, and Blood Institute (HL080666) (EPR).
Supplementary material for this article is available on the Genes & Cancer Web site at http://ganc.sagepub.com/supplemental.