Activation of ERK1/2 by MOS and TPL2 leads to dasatinib resistance in chronic myeloid leukaemia cells

Abstract The development of BCR::ABL1 tyrosine kinase inhibitors (TKIs), such as dasatinib, has dramatically improved survival in cases of chronic myeloid leukaemia (CML). However, the development of resistance to BCR::ABL1 TKIs is a clinical problem. BCR::ABL1 TKI resistance is known to have BCR::ABL1‐dependent or BCR::ABL1‐independent mechanisms, but the mechanism of BCR::ABL1 independence is not well understood. In the present study, we investigated the mechanism of BCR::ABL1‐independent dasatinib resistance. The expression and activation level of genes or proteins were evaluated using array CGH, real time PCR, or western blot analysis. Gene expression was modulated using siRNA‐mediated knockdown. Cell survival was assessed by using trypan blue dye method. We found that dasatinib‐resistant K562/DR and KU812/DR cells did not harbour a BCR::ABL1 mutation but had elevated expression and/or activation of MOS, TPL2 and ERK1/2. In addition, MOS siRNA, TPL2 siRNA and trametinib resensitized dasatinib‐resistant cells to dasatinib. Moreover, expression levels of MOS in dasatinib non‐responder patients with CML were higher than those in dasatinib responders, and the expression of TPL2 tended to increase in dasatinib non‐responder patients compared with that in responder patients. Our results indicate that activation of ERK1/2 by elevated MOS and TPL2 expression is involved in dasatinib resistance, and inhibition of these proteins overcomes dasatinib resistance. Therefore, MOS, TPL2 and ERK1/2 inhibitors may be therapeutically useful for treating BCR::ABL1‐independent dasatinib‐resistant CML.


| INTRODUCTION
Chronic myeloid leukaemia (CML) is a myeloproliferative tumour of haematopoietic stem cells that affects 1-2 in 100,000 people, accounting for 15% of all adult leukaemia cases. The Philadelphia chromosome, a chimeric chromosome resulting from a reciprocal translocation of the Abelson murine viral homologue 1 (ABL1) gene on chromosome 9 and the breakpoint cluster region (BCR) gene on chromosome 22, has been implicated as a major factor in the development of CML. 1 The Philadelphia chromosome generates BCR::ABL1 protein, which is positive in approximately 95% of CML patients. 2 BCR::ABL1 protein constitutively activates, Janus kinase (JAK)/signal transducer and activator of transcription (STAT), phosphoinositide 3-kinase (PI3K)/Akt, and mitogen-activated protein kinase kinase 1/2 (MEK1/2)/extracellular signal-regulated kinase 1/2 (ERK1/2) pathways, which are involved in CML cell proliferation and survival. 2 BCR:: ABL1-dependent activation of Src kinase also contributes to the pathogenesis of CML. 3 BCR::ABL1 tyrosine kinase inhibitors (TKIs) such as imatinib, nilotinib and dasatinib are currently the first-line agents for the treatment of CML. 4 Compared to conventional interferon therapy, these inhibitors have markedly improved the 5-year survival rate of chronic phase CML patients to more than 90%. 5 However, approximately 30% of patients who receive BCR::ABL1 TKIs develop resistance or intolerance, which is a major clinical problem. 6 A cause of BCR::ABL1 TKI resistance is point mutations in the kinase domain within the BCR::ABL1 protein, which interfere with the binding of BCR::ABL1 TKIs to the BCR::ABL1 protein, thereby reducing the sensitivity of the drug. 7 Currently, there are more than 1000 reported point mutations in the kinase domain of BCR::ABL1. Dasatinib is 300 times more potent than imatinib and is effective against most of these mutations. 8 However, dasatinib has been reported to be ineffective against the T315I mutation, and this mutation accounts for approximately 24% of dasatinib-resistant patients. 9 Third-generation BCR::ABL1 TKIs, ponatinib and rebastinib, have been developed to treat T315I mutations. 10 This has improved the prognosis in patients with resistant mutations in the BCR::ABL1 kinase domain. However, approximately 50% of patients who are resistant or intolerant to dasatinib do not have point mutations. 9 In addition to point mutations in the kinase domain within the BCR::ABL1 protein, other factors that may cause CML to become resistant to dasatinib include activation of the Gab2 or Akt pathway, reduction of FOXO1 or PTEN expression, and activation of Fyn or Lyn kinase. [11][12][13][14] However, the mechanism of BCR::ABL1-independent dasatinib resistance remains unclear.
In this study, we established dasatinib-resistant cell lines using BCR::ABL1-positive human CML cell lines K562 and KU812 and investigated the mechanism of dasatinib-acquired resistance.

| Cell viability assay
Cell survival and proliferation by treatment with various reagents was measured using the trypan blue dye staining assay, as previously described. 17,18 2.3 | Array comparative genomic hybridization Genomic DNA was analysed using OncoScan FFPE Assay (Affymetrix) according to the manufacturer's protocol, as previously described. 19

| Statistical analysis
All data are presented as the mean ± standard deviation of several independent experiments, and the data were analysed using analysis of variance with Dunnet test. Significant differences were determined at p < 0.05.
Other materials and methods is referred to the supplementary materials and methods.  Figure 1A). In addition, the survival of K562 and K562/DR cells was investigated using IC50 values for dasatinib and other BCR::ABL1 TKIs, such as imatinib, nilotinib, bafetinib, ponatinib, rebastinib, GNF-2 and GNF-5.
We found that K562/DR cells showed a markedly higher IC50 value than K562 cells ( Figure 1B; Figure S1A,B).
We also examined the ABL1 gene mutation related to BCR::ABL1 TKIs resistance in K562/DR cells. The NGS analysis did not reveal any mutations in the ABL1 gene in K562/DR cells ( Figure S1C).

| Constitutive activation of ERK1/2 in K562/ DR cells did not affect other targeted signalling molecules by BCR::ABL1 TKIs
We first examined the activation of signalling molecules directly inhibited by dasatinib and their downstream components to analyse the factors that contribute to the acquisition of dasatinib resistance. We found that the activation levels of ERK1/2 were higher in K562/DR cells than in K562 cells, but activation of BCR-ABL1, Src, c-Kit, Akt, JNK, p38MAPK, STAT1, STAT3 and STAT5 did not differ between K562 and K562/DR cells ( Figure 1C). is a member of the WNT family and is concerned in tumour cell growth and survival via activation of β-catenin. 21 MOS and TPL2 are upstream signal molecules of ERK1/2, and TPL2 overexpression is involved in resistance to imatinib in CD34+ CML cells. 22,23 In addition, we confirmed that MET, cytoplasmic β-catenin, nuclear β-catenin, phosphorylated TPL2, TPL2 and MOS were higher in K562/DR cells than in K562 cells, but there was no change in NIK expression between K562 and K562/DR cells ( Figure 2B). We also found that the amplification levels of genomic DNA and mRNA of TPL2, MOS and WNT16 in K562/DR cells were higher than those in K562 cells ( Figure 2C,D)  in K562/DR cells, but the activation level of ERK1/2 was higher than that of K562 cells ( Figure 3A,B). In addition, MOS or TPL2 siRNAs alone elevated the sensitivity of K562/DR cells to dasatinib ( Figure 3C). WNT16 siRNA markedly inhibited WNT16 mRNA and nuclear β-catenin expression in K562/DR cells but did not affect the viability of K562/DR cells ( Figure S2). Combined treatment with MOS and TPL2 siRNAs significantly inhibited ERK1/2 phosphorylation and overcame dasatinib resistance in K562/DR cells ( Figure 3D-F).
We also analysed whether MOS and TPL2 expression were different between dasatinib non-responder and responder patients using the GSE33224 dataset. We found that the expression levels of MOS in dasatinib non-responders were higher than those in dasatinib responders ( Figure 4C), and the expression level of TPL2 in dasatinib non-responders tended to increase more than that in dasatinib responders, although the difference was not significant ( Figure 4C). Therefore, high expression of MOS and TPL2 was correlated with dasatinib resistance in patients with CML.
To validate these findings, we examined whether the expression of MOS, TPL2, and phosphorylated ERK1/2 contributed to dasatinib    Figures S3 and S4A,B). Moreover, trametinib overcame dasatinib resistance in KU812/DR cells ( Figure S4C). It has been reported that the activation of NF-κB promotes TPL2 expression in imatinib-resistant CML cells. 23 We also investigated the activation/ expression of NF-κB p65 in dasatinib-resistant CML cells. There was no change in phosphorylated or total NF-κB p65 protein levels between parental and resistant cells ( Figure S5).

| DISCUSSION
In the present study, we established dasatinib-resistant K562/DR and KU812/DR cells that were resistant to dasatinib and other BCR::ABL1 TKIs, such as nilotinib, ponatinib, rebastinib and GNF-5. In addition, K562/DR and KU812/DR cells did not harbour a kinase domain mutation in BCR::ABL1 correlated with BCR::ABL1 TKI resistance. It has been reported that treatment for BCR::ABL1 mutations is described in therapeutic guidelines, but there is no description of what to do in the absence of BCR::ABL1 mutations. 24 Our results indicated that the Moreover, overexpression of MOS by p110 CUX1 induced ERK1/2 phosphorylation, cell proliferation and tumorigenesis in pancreatic cancer. 26 It has also been reported that high expression of MOS is associated with poor prognosis in patients with lung adenocarcinoma than in patients with low expression. 27 In addition, TPL2 overexpression was associated with imatinib resistance in patients with CD34-positive CML cells. 23 Elevated expression of TPL2 in BRAF V600E-mutated melanoma cells is known to confer resistance to BRAF inhibitors, and TPL2 expression is higher in patients with relapse after BRAF inhibitor therapy. 28 These findings indicate that overexpression of MOS and TPL2 is involved in dasatinib resistance in CML cells.
We showed that the activation of ERK1/2 via MOS and TPL2 overexpression is implicated in dasatinib resistance in CML, but our study has some limitations. First, the GSE dataset was used in this F I G U R E 5 Schematic summary of dasatinib resistance mechanism in chronic myeloid leukaemia (CML) cells. Activation of ERK1/2 by the overexpression of MOS and TPL2 was associated with BCR:: ABL1-independent dasatinib resistance in CML cells. Moreover, the inhibition of MOS and TPL2, ERK1/2 activation by siRNAs targeting MOS and TPL2, or trametinib treatment overcame dasatinib resistance. study to examine MOS and TPL2 expression in dasatinib-nonresponder patients with CML, but the sample size was small. More samples are needed to determine the extent to which the overexpression of MOS and TPL2 contributes to dasatinib resistance in patients with CML. Second, we did not use CML cells from patients with dasatinib-resistant CML overexpressing MOS and TPL2 to confirm whether the abrogation of MOS, TPL2 and ERK1/2 overcomes dasatinib resistance. This is important to elucidate the mechanism of dasatinib resistance in CML and to establish a therapeutic strategy, which needs to be clarified in a subsequent study.
In conclusion, we demonstrated that overexpression of MOS and TPL2 is associated with BCR::ABL1-independent dasatinib resistance in CML cells and that inhibition of MOS and TPL2 or ERK1/2 activation by MEK1/2 inhibitor overcomes dasatinib resistance ( Figure 5).
These findings suggest that a combination of MOS, TPL2 or MEK inhibitor and dasatinib may be a useful strategy for the treatment of BCR::ABL1-independent dasatinib-resistant CML.