KW-2478 is a safe HSP90α inhibitor with a BCR/ABL-specific inhibitory ability and a strong suppressive effect onimatinib-resistant CML cells

We detected the effect of KW-2478 on the growth of K562 and K562/G01 cells harbouring the BCR/ABL fusion protein. We discovered that KW-2478 inhibited CML cell proliferation in a dose- and time-dependent manner. The IC50 values of the KW-2478-treated K562 cells at 48 and 72 h were 5.195 µM and 4.659 µM, respectively. The IC50 values of the K562/G01 cells treated under the same conditions were 1.424 µM and 0.867 µM (Fig. 1a). We also evaluated the effect of KW-2478 on the proliferation of normal human peripheral blood mononuclear cells (PBMCs). Interestingly, KW-2478 did not have a strong inhibitory effect on the viability of normal human PBMCs, even when the concentration of KW-2478 reached 10 µM (Fig. 1b). We also calculated the IC50 values of other BCR/ABL-positive and BCR/ABL-negative cells. The IC50 values of the KG-1a, THP-1, KCL-22, K562, K562/G01, SUP-B15, 32DP and 32DP-T315I cells treated with KW-2478 for 48 h were 37.06 µM, 29.07 µM, 5.473 µM, 5.196 µM, 1.446 µM, 1.166 µM, 0.639 µM and 0.6138 µM, respectively. The results showed that the IC50 value of the BCR/ABL-negative cells (KG-1a and THP-1 cells) was much higher than that of the BCR/ABL-positive cells (KCL-22, K562, K562/G01, SUPB15, 32DP and 32DP-T315I cells). These results indicated that KW-2478 specifically inhibited BCR/ABL-positive cells (Fig. 1c). The results of the colony-forming assay showed that the number and size of colonies formed by K562 or K562/G01 cells were significantly decreased by KW-2478. We also discovered that at the same concentration of KW-2478, the inhibitory effect on colony formation in K562/G01 cells was more stronger than that in K562 cells. When K562/G01 cells were treated with 0.25 µM KW-2478, the number of colonies was significantly reduced compared with that in the control group. When the concentration of KW-2478 reached 0.5 µM, the colony number of K562/G01 cells showed a significant difference (Fig. 1d, e). Collectively, KW-2478 inhibited the growth of CML cells, and K562/G01 cells were more sensitive to KW-2478 than K562 cells. The above experiments proved that KW-2478 was a safe HSP90α inhibitor with a BCR/ABL-specific inhibitory ability and showed strong inhibition against imatinib-resistant cells.

Fig. 1

KW-2478 inhibited the growth of BCR/ABL-positive cells with no obvious effect on BCR/ABL-negative cells. a The effects of KW-2478 on the growth of K562 and K562/G01 cells were assessed by CCK-8 assay after incubation for 24, 48 and 72 h. b The survival rate of PBMCs from two healthy individuals was detected by CCK-8 assay after incubation with KW-2478 for 48 h. c The IC50 values of KG-1a, THP1, KCL-22, K562, K562/G01, SUP-B15, 32DP-T315I and 32DP cells were detected by CCK-8 assay after 48 h of treatment with KW-2478. d, e The proliferation of CML cells after KW-2478 treatment was detected using a colony formation assay, and the results were statistically analysed. The scale bar represents 100 μm. The values are the mean standard deviation of three independent experiments. Asterisks denote statistically significant differences compared with the control group

KW-2478 induces G2/M phase cell cycle arrest in imatinib-sensitive and imatinib-resistant CML cells

The effect of KW-2478 on K562 and K562/G01 cell cycle progression was investigated by flow cytometry. The results revealed that KW-2478 induced a dose-dependent increase in the G2/M-phase population and a decrease in the S-phase population. Compared to that of the control group (12.75%), the G2/M-phase cell population of the KW-2478-treated K562 cells reached 14.02%, 17.17% or 24.91% at concentrations of 2, 4 or 8 µM, respectively. The G2/M phase arrest caused by KW-2478 was more effective in K562/G01 cells. Moreover, the K562/G01 cell population at G2/M-phase increased to 47.97% at 8 µM KW-2478 (Fig. 2a, b). To further investigate the reason for the G2/M phase arrest induced by KW-2478, we detected the expression of G2/M phase-regulating proteins by western blotting. It has been reported that activated cyclin B1/CDC2 complexes are required for progression through G2/M. Moreover, p27 and p21 can inhibit the function of the cyclin B1/CDC2 complexes and block cell cycle progression [31]. Our results showed that the expression of p21 and p27 increased, while cyclin B1 expression was reduced considerably after KW-2478 treatment for 48 h (Fig. 2c). The results in this part showed that KW-2478 induced cell cycle arrest in G2/M phase by downregulating cyclin B1 expression and upregulating p21 and p27 expression, and the G2/M phase arrest was much more significant in K562/G01 cells than in K562 cells.

Fig. 2

KW-2478 induced cell cycle arrest in the G2/M phase in CML cells. a After K562 and K562/G01 cells were treated with KW-2478 for 48 h, the cell cycle distribution was determined using PI staining and flow cytometry. b Following flow cytometry detection, the cell proportions in each cell cycle phase were summarized. c Levels of expression of cell cycle-related proteins, such as cyclin B1, p21 and p27, were quantitatively analysed by western blots. The results are expressed as the mean ± SD. *P < 0.05, **P < 0.01

KW-2478 induces apoptosis by activating the caspase pathway in imatinib-sensitive and imatinib-resistant CML cells

Treatment with KW-2478 for 48 h resulted in an increased percentage of apoptotic cells compared to that of the control group (Fig. 3a, b). After treatment with the same dose of KW-2478, the percentage of apoptotic K562/G01 cells was substantially higher than that of K562 cells. DAPI staining was used to identify morphological alterations associated with apoptosis. CML cells were stained with DAPI dye and observed with fluorescence microscopy. In the control group, cell nuclei were round with uniform size, and chromatin was distributed homogeneously. However, chromatin condensation, marginalization and nuclear fragmentation were commonly detected in the KW-2478-treated cells (Fig. 3c). It has been reported that the activation of the caspase pathway in response to proapoptotic signals is crucial in the process of apoptosis triggered by a range of stimuli, which necessitates the activation of early caspases such caspase-8 and -9 [32]. Treatment of K562 cells with 8 µM KW-2478 and K562/G01 cells with just 1 µM KW-2478 resulted in PARP cleavage (Fig. 3d). In parallel, KW-2478 led to an increase in caspase-3 and caspase-9 cleavage in CML cells, which indicated their activation. The decrease in pro-caspase-8 levels indicated that its proapoptotic activity increased (Fig. 3d). These findings revealed that KW-2478 activated the caspase pathway and caused CML cell apoptosis, and the effect was stronger in imatinib-resistant K562/G01 cells.

Fig. 3

KW-2478 induced apoptosis by activating the caspase pathway in CML cells. a After K562 and K562/G01 cells were treated with KW-2478 for 48 h, flow cytometry was used to detect cell apoptosis after annexin-V-FITC/PI labelling. b The cell apoptosis rates were summarized after flow cytometry detection. c The morphology of apoptotic cells was examined after DAPI staining. The scale bar represents 10 μm. d The expression levels of cell apoptosis-related proteins, such as PARP, pro-caspase-8, cle-caspase-9 and caspase-3, were quantitatively analysed by western blots after exposure to KW-2478 for 48 h. β-Actin expression served as a loading control

Apoptosis induced by KW-2478 results from mitochondrial dysfunction

Caspase activation is linked to mitochondrial malfunction and cytochrome C release into the cytoplasm. JC-1 is a fluorescent dye that characterizes the dissipation of mitochondrial membrane potential (∆Ψm) by a considerable change of red to green fluorescence [33]. KW-2478 enhanced the emission of green fluorescence in CML cells, which demonstrated the dissipation of ∆Ψm during apoptosis (Fig. 3a). The ∆Ψm of K562/G01 cells was significantly decreased after incubation with 2 µM KW-2478, whereas the ∆Ψm of K562 cells decreased significantly only when the concentration of KW-2478 reached 8 µM (Fig. 4b). Furthermore, JC-1 fluorescence imaging confirmed this finding. Compared with those of the control group, K562 and K562/G01 cells showed stronger green fluorescence and weaker red fluorescence after treatment with KW-2478, which indicated a decrease in their ∆Ψm (Fig. 4c). The effect of KW-2478 on proteins involved in mitochondrial apoptotic pathways was then investigated. We first obtained the proteins of CML cells after removal of mitochondria. The quantities of cytochrome C released into the cytoplasm from the mitochondria increased after treatment with KW-2478 (Fig. 4d). These findings indicated that KW-2478 induced apoptosis partly by disrupting the mitochondrial membrane potential. The Bcl-2 family is an important regulator in mediating the release of cytochrome C. Bad inhibits the release of cytochrome C, while Bcl-2 and Bcl-XL promote its release [34]. The protein expression of the Bcl-2 family was studied to learn more about the mechanisms of apoptosis induced by KW-2478 in CML cells. The results revealed that Bad, a proapoptotic protein, had significantly upregulated expression. The expression levels of Bcl-2 and its closest relative, Bcl-XL, were decreased in the KW-2478-treated CML cells. These effects were dose dependent, and a lower concentration of KW-2478 caused a more obvious effect in K562/G01 cells (Fig. 4e). The results showed that KW-2478 promoted the release of cytochrome C into the cytoplasm and reduced mitochondrial membrane potential by regulating Bcl-2 family proteins, thereby inducing apoptosis through the mitochondrial pathway in CML cells.

Fig. 4

Apoptosis induced by KW-2478 is relevant to mitochondrial dysfunction. a The changes in mitochondrial membrane potential were detected by flow cytometry after CML cells were stained with JC-1. b The ratio of red fluorescence to green fluorescence was summarized after flow cytometry detection. c Representative images of JC-1 fluorescence imaging. The scale bar represents 50 μm. d Cytochrome C released from mitochondria was detected by western blot. e Western blotting was used to detect the expression level of the Bcl-2 protein family

KW-2478 inhibits the chaperone function of HSP90 and then weakens the BCR/ABL and MAPK signalling pathways

Our results clearly demonstrated that the levels of phospho-BCR/ABL, BCR/ABL, phospho-STAT5, and STAT5 in K562 and K562/G01 cells were diminished after KW-2478 treatment (Fig. 5a). To investigate how KW-2478 altered BCR/ABL expression, we examined the mRNA level of BCR/ABL in response to KW-2478. The results indicated that the mRNA level of BCR/ABL slightly increased, which suggested that KW-2478 affected BCR/ABL expression mainly through post-transcriptional mechanisms (Fig. 5c). The HSP90 family is an important chaperone for BCR/ABL, so we detected the effect of KW-2478 on the protein and mRNA expression of the HSP90 family. The results showed that HSP90α mRNA expression was upregulated after treatment with KW-2478, but no difference was observed in the protein levels of HSP90α, HSP90β and Grp94 (Fig. 5b, d). HSP70 is a member of the heat shock protein family and can be used as an indicator of the heat shock response (HSR) [35]. We found that HSP70 was barely affected by KW-2478, which indicated that KW-2478 did not induce a remarkable heat shock response (Fig. 5b). Next, we further investigated whether the downregulated protein expression of BCR/ABL by KW-2478 was caused by the chaperone function of HSP90α. K562/G01 cells were cultured with or without KW-2478 for 48 h, and then, the interaction between HSP90α and BCR/ABL was detected by coimmunoprecipitation. The results indicated that BCR/ABL interacted with HSP90α in the absence of KW-2478 (Fig. 5e), while the interaction between them was disturbed by KW-2478 (Fig. 5f). These experiments proved that HSP90α could bind to BCR/ABL in K562/G01 cells and that KW-2478 could obstruct their binding. Without the protection of HSP90α, BCR/ABL might degrade, which leads to the downregulation of the expression of BCR/ABL and its downstream molecules and reduced oncogenic properties of CML cells. The results of western blot experiments proved that KW-2478 did not affect the expression of Ras but significantly downregulated p-Raf1 expression. Moreover, decreased expression of Mek1/2 and p-Erk1/2 was observed after treatment with KW-2478 in CML cells (Fig. 5g). Mek1/2 and p-Erk1/2 are downstream molecules of Raf1, and all three proteins are important regulators in the MAPK pathway, which is critical for cell proliferation. These results indicated that p-Raf1 expression was directly downregulated by KW-2478 but not through Ras. Overall, KW-2478 simultaneously inhibited the BCR/ABL and MAPK signalling pathways by disturbing the chaperone function of HSP90α and then induced obvious growth suppression in CML cells.

Fig. 5

KW-2478 inhibited the chaperone function of HSP90α and then weakened the BCR/ABL and MAPK signalling pathways. a The expression levels of BCR/ABL, p-BCR/ABL, STAT5 and p-STAT5 in CML cells after KW-2478 treatment were detected by western blots. b The expression levels of HSP90α, HSP90β, Grp94 and Hsp70 in CML cells after KW-2478 treatment were quantitatively analysed by western blot. c The mRNA level of BCR/ABL was detected with qRT-PCR. β-Actin was used as an internal control. d The mRNA level of HSP90α was detected with qRT-PCR. e The interaction between HSP90α and BCR/ABL was analysed by Co-IP assays with no KW-2478 treatment. f The influence of KW-2478 on the interaction between HSP90α and BCR/ABL was analysed by Co-IP assays. g Molecules in the MAPK signalling pathway, such as Ras, phospho-Raf1, Mek1/2, phosphorylated and total Erk1/2, were analysed by western blots. The results are expressed as the mean ± SD. *p < 0.05

KW-2478 synergistically functions with imatinib to induce apoptosis and growth inhibition in imatinib-sensitive and imatinib-resistant CML cells

To investigate the synergistic effect of KW-2478 with imatinib, we treated K562 cells and K562/G01 cells with KW-2478, imatinib or both for 48 h, followed by a CCK-8 assay. Considering the different sensitivities of K562 and K562/G01 cells to KW-2478 and imatinib, we used different drug combinations. The combination exhibited a higher inhibition rate on either K562 or K562/G01 cells than KW-2478 or imatinib alone (Fig. 6a). The combination index (CI) was analysed with CompuSyn software. All values of the CI were than 1 indicating that the combined growth inhibitory effect was synergistic (CI < 1, = 1, and > 1 represent synergistic, additive, and antagonistic effects, respectively) (Fig. 6b, Additional file 1: Tables S3 and S4). Moreover, the CI values of the K562/G01 groups were lower than those of the K562 groups, indicating that the synergistic effect was more effective in K562/G01 cells. Then, flow cytometry analysis was used to examine the apoptotic index values to learn more about the influence of this combination on CML cell apoptosis. As indicated in Fig. 6c, d, the combination group had a greater apoptotic rate than the KW-2478 or imatinib alone groups. These experiments showed that KW-2478 had a synergistic effect with imatinib whether in growth inhibition or induction of apoptosis. The combination of imatinib and KW-2478 is a prospective treatment to address TKI resistance and intolerance in CML treatment.

Fig. 6

KW-2478 synergistically functioned with imatinib to induce apoptosis and growth inhibition in CML cells. a A CCK-8 assay was used to determine the combined effects of KW-2478 and imatinib on cell viability of K562 and K562/G01 cells. b The combination index of K562 or K562/G01 cells was analysed using CompuSyn software. c Following treatment with KW-2478, imatinib or both, cell apoptosis was detected by flow cytometry. d Statistical analysis of the apoptosis rates of K562 and K562/G01 cells treated with KW-2478, imatinib or both

KW-2478 has effective antitumor activity in imatinib-sensitive and imatinib-resistant CML-like mouse models

We confirmed that KW-2478 had an excellent antileukaemia effects in vitro, so we further investigated its antileukaemia effect in vivo. The in vivo antitumor activity of KW-2478 was examined in a NOD-SCID mouse model bearing human CML xenografts. We generated two CML mouse models. One was injected with K562 cells to mimic ordinary CML disease, and the other was injected with K562/G01 cells to mimic imatinib-resistant CML disease. The results showed that the mice that received KW-2478 treatment displayed lower white blood cell counts than the controls in both models (Fig. 7a). When the mice showed obvious symptoms, such as weight loss, listlessness, vertical hair and arched back or unstable gait, we sacrificed the mice. The spleen and liver were removed and weighed. As shown in Fig. 7b–d, the mice in the control groups had more severe hepatosplenomegaly than the mice in the KW-2478-treated groups. The infiltration of leukaemic cells in the liver and spleen was examined by HE staining and Wright’s staining. The results of HE staining revealed that the number of rod-shaped leukocytes and lobulated leukocytes that infiltrated the liver and spleen significantly decreased in the KW-2478 treatment groups compared to the control groups (Fig. 7e). The results of Wright’s staining showed less active proliferation in the bone marrow and less granulocytes infiltrated in the liver and spleen in the KW-2478-treated groups than in the control groups (Fig. 7f). We also detected the expression of BCR/ABL in cells from mouse livers, spleens and bone marrow to further confirm the infiltration of CML cells by immunofluorescence assays. We found that a high level of BCR/ABL expression was detected in these tissues from control group mice, while we barely detected the expression of BCR/ABL in the KW-2478-treated groups (Fig. 7g). We extracted bone marrow proteins from these mice and performed western blotting to detect the expression of BCR/ABL and HSP90 family proteins. The BCR/ABL protein level in the KW-2478-treated mice was lower than that in the control mice. There was no obvious difference in the protein levels of the HSP90 family and HSP70 between the KW-2478-treated groups and the control groups (Fig. 7h). As shown in Kaplan–Meier survival curves, the mouse survival time in the KW-2478 groups was prolonged. In the observation period of 90 days, 40% of the mice survived in the K562 control group, while 80% of mice survived in the KW-2478 treatment group. In addition, 40% of the mice survived in the K562/G01 control group, while 100% of the mice survived in the KW-2478 treatment group (Fig. 7i). Moreover, the remaining mice in the two control groups still showed CML, while the mice in the two KW-2478 treatment groups showed no signs of disease. We predicted that the control mice would die before the treatment mice if the observation period was extended.

Fig. 7

KW-2478 had effective antitumor activity in imatinib-sensitive and imatinib-resistant CML-like mouse models. a The maximum WBC counts of mice were recorded. b, c The liver and spleen weights of mice were measured. The results are expressed as the mean ± SD. *p < 0.05, **p < 0.01. d Images of the liver and spleen are displayed. e Murine liver and spleen infiltration was analysed by HE staining. The scale bar represents 50 μm. f Cells from mouse bone marrow, liver, and spleen were stained by Wright’s stain and then photographed under a microscope. The scale bar represents 20 μm. g The expression of BCR/ABL in each tissue was detected by immunofluorescence assays. The scale bar represents 10 μm. h The expression levels of BCR/ABL and HSP90 family proteins from mouse bone marrow cells were detected by western blots. i The survival curves of mice were analysed by the Kaplan–Meier method

In general, these results proved that KW-2478 prolonged the mouse lifespan and alleviated disease symptoms and tumour cell infiltration in both the K562 and K562/G01 mouse models.