The ubiquitin-proteasome system (UPS) is the principle pathway for diverse intracellular protein degradation [16]. Proteasome is a large proteolytic complex that consists of a 20S catalytic complex and two 19S regulatory subunits. The 20S proteasome is composed of two identical outer α-rings and two identical inner β-rings, each composed of seven distinct subunits. The β1, β2, and β5 subunits mediate the caspase-like, trypsin-like, and chymotrypsin-like activity, respectively [16]. Proteins that are to be degraded are tagged with ubiquitin chains and bind to a receptor on the 19S complex. Once recognized by the regulatory complex, the ubiquitin chain is removed and the protein is denatured and presented to the 20S proteasome for degradation [16]. Though UPS is critical to normal cell survival and function, proteasome has been shown to be an appropriate therapeutic target for cancer. Bortezomib (BOR)/Velcade/PS-341 [17] and the proteasome inhibitor I (PSI, Z-Ile-Glu(OtBu)-Ala-Leucinal)[18] are two inhibitors of the β5 subunit and the chymotryptic activity of the proteasome. One of the results of proteasome inhibition is the accumulation of the normally proteasome-degraded IκB in cytoplasm, leading to inhibition of the translocation of NFκB from cytoplasm to nucleus. BOR prolonged life span[19] and was shown to be superior to high-dose dexamethasone for relapsed MM patients[20]. PSI was shown to be a potent apoptosis inducer for myeloma and leukemic cells [21], [22]. Interestingly, BOR and PSI targeted the BCR-ABL oncoprotein and induced apoptosis of CML cells sensitive or resistant to IM, and exerted synergic effects with histone deacetylase inhibitors and cyclin-dependent kinase inhibitor flavopiridol [22]–[26]. However, the in vivo efficacy of proteasome inhibitors on CML remains obscure, and whether proteasome inhibitors could exert synergistic/additive effects with IM needs more in-depth analysis.

We show that IM in combination with proteasome inhibitor significantly prolongs life span of BALB/c mice bearing BCR-ABL/GFP-expressing murine hematopoietic cells (Figure 1, Figure S1 and S2), and suppresses tumor growth in nude mice harboring K562 cells (Figure 2). In vitro, IM/BOR and IM/PSI exhibit an enhanced inhibition of long-term colony forming activity and short-term cell growth of CD34+ cells from CML patients at CP or BC (Figure 3), cause potentiated proliferation inhibition in K562 and 32D cells expressing BCR-ABL (Figure 4 and ​and5),5), and exert significantly potentiated apoptotic effects on CML cells (Figure 6 and Figure S5). Heaney et al [60] recently demonstrated that proteasome may be a relevant target for quiescent CML stem cells following tyrosine kinase inhibitor therapy, while proteasome inhibitor are capable of inducing CML stem cell specific apoptosis. Hence, combining tyrosine kinase inhibitor and proteasome inhibitor in treating CML might probably provide beneficial effects to patients including relapsed ones.

Gatto et al[25] showed that sequential administration of PS-341 and IM (<0.5 µM) caused synergistic apoptotic effects on KBM-5 cells, while antagonistic effects were detected if IM was used at a higher concentration (≥0.5 µM). In addition, antagonistic effects were observed when PS-341 and IM were added simultaneously. Since KBM-5 cell line was derived from a patient with myeloid blastic phase, and K562 cells were derived from a patient with CML in erythroid blast phase, they might respond differently to a treatment protocol. An interesting finding in this work is that in CD34+ cells from patients at blastic phase, treatment with IM/BOR and IM/PSI significantly inhibits BFU-E but not CFU-GM (Figure 3A, lower panel), suggesting that cells from CML at blastic phase represent a heterozygous population which might respond diversely to drug treatment, and erythroleukemia cells seem to be more sensitive to IM/BOR combination. However, the exact mechanisms underlying the difference in response of KBM-5 and K562 cells to IM/BOR combination warrant further investigation.

Neither IM/BOR nor IM/PSI appears to increase systemic toxicity in our animal tests since the body weights and overall appearance of mice being given the combination of drugs are not different from controls or the mice receiving only one drug. Recently, IM [at 620±166 (400–800) mg/d] was shown to cause cardiotoxicity in some individuals [13], [14], [57], and unexpected cardiotoxicity was reported in patients received BOR (chemotherapy was used prior to or concomitantly with BOR) [61]–[63]. We show that though IM at high dose induces apoptosis in a small proportion of cardiomyocytes in samples from nude mice, BOR alone as well as BOR in combination with low dose IM does not impair the heart (Figure 2). If these results could be translated into clinical practice, IM at a dose of 100–120 mg orally per day in combination with BOR could be tried.

Compared to normal cells, cancer cells often bear higher Δψm and evade mitochondrial apoptosis [40]. Normally, in response to cellular stress, the cell's mitochondria are triggered to release cyto C into the cytosol which then binds to Apaf-1 and initiates the formation of apoptosome, leading to the activation of casp-9 and subsequent casp-3. The release of cyto C is tightly regulated by pro- (e.g., Bax and Bak) and anti-apoptotic (e.g., Bcl-2 and Bcl-XL) members of Bcl-2 family. In CML, BCR-ABL upregulates Bcl-2 [64] and Bcl-XL [65] through activation of STAT5, and inhibits release of cytochrome C [66] and prevents caspase activation even after cyto C release [67], hence confering resistance to apoptosis to CML cells. Interestingly, IM/BOR and IM/PSI cause collapse of Δψm, downregulation of pBCL-2, increase of cytoplasmic cyto C and activation of casp-9, -8 and -3 (Figure 6). It is well-known that IM acts as a specific inhibitor of BCR-ABL. BOR and PSI significantly enhance IM-triggered suppression of pBCR-ABL and inhibition of its tyrosine kinase activity in vitro and in vivo (Figure 2D and Figure 7, A through C). In consistence with a previous report [42], we show that activation of caspases by IM/BOR and IM/PSI leads to catabolism of BCR-ABL, where caspase inhibitor not only reduces apoptosis but also inhibits degradation of BCR-ABL (Figure 7D). IM/BOR and IM/PSI also downregulate pSTAT5 (Figure 7A). These data suggest that the combinatory regimens on one hand target the mitochondria, downregulate Bcl-2 and activate caspases, on the other hand inhibit BCR-ABL/STAT5 which might in turn potentiate downregulation of Bcl-2 and activation of caspases. Furthermore, activated caspases can enhance BCR-ABL catabolism and inactivation. Therefore, IM/BOR and IM/PSI may trigger a positive feedback apoptotic signaling network, leading to a significant amplification of apoptotic effects of each agent.

Dysregulation of Wnt-β-catenin signaling underlies multiple human malignancies [68]. In CML, BCR-ABL triggers tyrosine phosphorylation and hence stabilization and activation of β-catenin [39], which enhances the self-renewal and leukemic potential of CML stem/progenitors cells [69], [70]. We show that proteasome inhibitors and IM exert opposite effects on β-catenin: BOR and PSI inhibit its degradation and activate its CRT activity, while IM causes its inactivation (Figure 8A and B). Interestingly, the ultimate result of IM/BOR and IM/PSI on β-catenin is its inactivation (Figure 8A and B), and the expression of two β-catenin targets, c-Myc and cyclin D1, was downregulated (Figure 7A and Figure 8A and C), suggesting that IM dominates the effect of IM/BOR and IM/PSI on Wnt-β-catenin pathway. Casp-3 was shown to play an important role in IM-induced β-catenin catabolism [39], while PP2A reduced expression of β-catenin and inhibited transcription of its target genes [71]. Hence, BCR-ABL inactivation, caspases activation and PP2A restoration may contribute to β-catenin inactivation, which may facilitate eradication of CML stem/progenitor cells. Intriguingly, our results do show that IM/BOR and IM/PSI inhibit short term cell growth and long term colony forming activity of CD34+ stem/progenitor cells from CML patients (Figure 3). BTK which is involved in IM-resistance, was shown to use a positive autoregulatory feedback mechanism to stimulate transcription from its own promoter via NFβB [49]. Accumulation of IκB (Figure 8A) and inhibition of DNA binding activity of NFκB (Figure 8D) by IM/BOR and IM/PSI might lead to inhibition of BTK. These results suggest that combined use of IM and proteasome inhibitor may be helpful in reducing relapse and overcoming IM-resistance.

The state of phosphorylation of proteins is governed by the coordinated and competing actions of protein kinases and phosphatases. BCR-ABL bears dual functions to interfering with normal signal transduction. The fusion protein has constitutively active tyrosine kinase activity, and it inhibits phosphatases including PP2A through BCR-ABL-induced expression of SET protein [35]. PP2A is also inactivated by CIP2A through stabilization of c-Myc [51], which is regulated by E2F1 [43] and β-catenin [45]. We found that proteasome inhibitor represses the β5 subunit and inhibits chymotryptic activity of the 26S proteasome (Figure 9D and E), leading to accumulation of Ub-PP2A (Figure 9, A through C). In vivo, IM/BOR also causes upregulation of PP2A (Figure 2D). Accumulation of PP2A is further confirmed in Kasumi-1, U266 and A549 cells treated with BOR (Figure S7). Of course, inhibition of BCR-ABL/SET and CIP2A might also contribute to PP2A re-activation. As a result, PP2A activity is increased (Figure 9B). PP2A activator FTY720 [53] synergizes with IM in inducing apoptosis (Figure 9F), mimicking effects of proteasome inhibitors. Suppression of PP2A by OA and PP2A-specific siRNA inhibits combination regimen-induced apoptosis, and results in upregulation of BCR-ABL (Figure 9, G through I). Intriguingly, downregulation of SET, CIP2A, c-Myc, E2F1, and β-catenin forms a complex positive feedback signal network for BCR-ABL inactivation and PP2A activation. These signals may amplify effects of IM and proteasome inhibitor, facilitating apoptosis induction by the combination regimens.

In summary, we report here combined use of IM and BOR/PSI modulates several signal pathways and forms positive feed back loops for CML cell apoptosis (Figure 10), providing potential benefits for optimizing clinical CML remedy.

Table S1

Figure S1

Figure S2

Figure S3

Figure S4

Figure S5

Figure S6

Figure S7