shRNA library screening identifies nucleocytoplasmic transport as a mediator of BCR-ABL1 kinase-independent resistance

Abstract

The mechanisms underlying tyrosine kinase inhibitor (TKI) resistance in chronic myeloid leukemia (CML) patients lacking explanatory BCR-ABL1 kinase domain mutations are incompletely understood. To identify mechanisms of TKI resistance that are independent of BCR-ABL1 kinase activity, we introduced a lentiviral short hairpin RNA (shRNA) library targeting ∼5000 cell signaling genes into K562(R), a CML cell line with BCR-ABL1 kinase-independent TKI resistance expressing exclusively native BCR-ABL1. A customized algorithm identified genes whose shRNA-mediated knockdown markedly impaired growth of K562(R) cells compared with TKI-sensitive controls. Among the top candidates were 2 components of the nucleocytoplasmic transport complex, RAN and XPO1 (CRM1). shRNA-mediated RAN inhibition or treatment of cells with the XPO1 inhibitor, KPT-330 (Selinexor), increased the imatinib sensitivity of CML cell lines with kinase-independent TKI resistance. Inhibition of either RAN or XPO1 impaired colony formation of CD34(+) cells from newly diagnosed and TKI-resistant CML patients in the presence of imatinib, without effects on CD34(+) cells from normal cord blood or from a patient harboring the BCR-ABL1(T315I) mutant. These data implicate RAN in BCR-ABL1 kinase-independent imatinib resistance and show that shRNA library screens are useful to identify alternative pathways critical to drug resistance in CML.

Figure 1

K562^R and AR230^R cells are resistant to imatinib despite inhibition of BCR-ABL1 kinase. (A) Whole cell extracts were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and probed with antibodies directed against ABL1 and phospho-ABL1; α-tubulin was analyzed as a control. (B) Bar graphs represent imatinib IC₅₀ for K562^S, K562^R, AR230^S, and AR230^R cells measured by treating these cells with increasing concentrations of imatinib and quantifying cell proliferation by MTS assay after 72 hours. Errors bars represent standard error of the mean (SEM); *P < .05.

Figure 2

Experimental design and reproducibility. (A) Flow diagram showing the experimental designs of the lentiviral shRNA screen and validation experiments. (B) K562^S and K562^R cells were infected with the Cellecta HM1 library and cultured in puromycin (see “Materials and methods”). Independent experiments were performed at the Huntsman Cancer Institute (R1, R2, R3, and HCI) or Cellecta. Comparison of the fold-change of barcodes between experiments using either K562^S or K562^R cells reveals a high level of correlation.

Figure 3

Validation of selected candidates from the lentiviral screen. (A) K562^S and K562^R cells were lentivirally infected with tetracycline-inducible constructs for expression of the top-scoring shRNA for each indicated gene. Expression of the candidate gene was measured by qRT-PCR 72 hours after addition of doxycycline to the culture medium (n = 3). (B) K562^S and K562^R cells stably expressing the doxycycline-inducible constructs were analyzed by MTS assay 72 hours after addition of doxycycline (n = 3). Error bars represent SEM; *P < .05.

Figure 4

Cytoplasmic RAN contributes to BCR-ABL1 kinase-independent TKI resistance. (A) Whole cell, nuclear, and cytoplasmic lysates of K562^S, K562^R, AR230^S, and AR230^R cells (n = 3) either untreated or treated with 1 μM imatinib were separated by SDS-PAGE and analyzed for RAN expression and subcellular localization by immunoblot analyses. (B) qRT-PCR quantitation of RAN mRNA in shRAN-expressing cells (n = 3) either untreated or treated with doxycycline (72 hours, 0.1 μg/mL). (C) shRAN-induced apoptosis was assessed by staining with annexin V followed by flow cytometric analysis (n = 6). (D) Parental and TKI-resistant K562 and AR230 cells were incubated with and without doxycycline at graded imatinib concentrations, followed by quantification of viable cells by MTS assay at 72 hours (n = 3). (E) Table shows imatinib IC₅₀ (nM) of K562^R and AR230^R cells expressing shRAN in the presence or absence of 0.1 μg/mL doxycycline, as measured by MTS assay following treatment of 72 hours. Error bars represent SEM; *P < .05.

Figure 5

Ectopic expression of RanGAP confers resistance to imatinib. (A-B) K562^S cells were nucleofected with the pDsRed1-N1 RanGAP plasmid for expression of RanGAP or control vector, followed by culture in graded concentrations of imatinib. RanGAP expression was confirmed by qRT-PCR (A). At 120 hours, the viable cells were quantified by flow cytometric analyses (B).

Figure 6

Enhanced RAN/XPO1 shuttling activity is associated with enhanced levels of cytoplasmic SET in TKI-resistant CML cell lines. (A) Whole cell, nuclear, and cytoplasmic lysates of K562^S, K562^R, AR230^S, and AR230^R cells (n = 3) either untreated or treated with 1 μM imatinib were separated by SDS-PAGE and analyzed for SET expression and subcellular localization by immunoblot analyses. The α-tubulin and lamin B fractionation blots overlap with that of Figure 4A. (B) Whole cell, nuclear, and cytoplasmic lysates of K562^S, K562^R, AR230^S, and AR230^R cells (n = 3) expressing shRAN in the presence or absence of doxycycline were separated by SDS-PAGE and analyzed for SET expression and subcellular localization.

Figure 7

Inhibition of RAN/XPO1 impairs survival of CML but not normal CD34⁺ cord blood cells. (A) CD34⁺ cells from normal cord blood (n = 2) were either infected with doxycycline-inducible shRAN and plated in semisolid medium with and without 2.5 µM imatinib (i) or plated in semisolid medium in the presence or absence of graded concentrations of KPT-330 (ii-iii). Colonies were counted after 14 days. RAN inhibition had no effect on survival of normal CD34⁺ cord blood cells. (B) CD34⁺ cells from newly diagnosed CML patients were either infected with shRAN (i) or treated with KPT-330 (ii) and analyzed for colony formation in the indicated conditions. Inhibition of XPO1 by treatment with KPT-330 also resulted in enhanced levels of nuclear RAN, SET, and p53 in CD34⁺ cells from newly diagnosed CML patients. Lamin B was analyzed to control for nuclear fractionation and α-tubulin for cytoplasmic fractionation. (C) CD34⁺ cells from patients with clinical TKI resistance were either infected with shRAN or treated with KPT-330 and analyzed for colony-forming ability. shRAN and KPT-330 significantly reduced survival of CML CD34⁺ cells from TKI-resistant patients with wild-type BCR-ABL1 (i), but not from a patient with BCR-ABL1^T315I (ii). Error bars represent SEM; *P < .05. Because only 1 BCR-ABL1^T315I patient sample was analyzed, standard errors are not provided.