The antiproliferative activity of kinase inhibitors in chronic myeloid leukemia cells is mediated by FOXO transcription factors

Abstract

Chronic myeloid leukemia (CML) is initiated and maintained by the tyrosine kinase BCR-ABL which activates a number of signal transduction pathways, including PI3K/AKT signaling and consequently inactivates FOXO transcription factors. ABL-specific tyrosine kinase inhibitors (TKIs) induce minimal apoptosis in CML progenitor cells, yet exert potent antiproliferative effects, through as yet poorly understood mechanisms. Here, we demonstrate that in CD34+ CML cells, FOXO1 and 3a are inactivated and relocalized to the cytoplasm by BCR-ABL activity. TKIs caused a decrease in phosphorylation of FOXOs, leading to their relocalization from cytoplasm (inactive) to nucleus (active), where they modulated the expression of key FOXO target genes, such as Cyclin D1, ATM, CDKN1C, and BCL6 and induced G1 arrest. Activation of FOXO1 and 3a and a decreased expression of their target gene Cyclin D1 were also observed after 6 days of in vivo treatment with dasatinib in a CML transgenic mouse model. The over-expression of FOXO3a in CML cells combined with TKIs to reduce proliferation, with similar results seen for inhibitors of PI3K/AKT/mTOR signaling. While stable expression of an active FOXO3a mutant induced a similar level of quiescence to TKIs alone, shRNA-mediated knockdown of FOXO3a drove CML cells into cell cycle and potentiated TKI-induced apoptosis. These data demonstrate that TKI-induced G1 arrest in CML cells is mediated through inhibition of the PI3K/AKT pathway and reactivation of FOXOs. This enhanced understanding of TKI activity and induced progenitor cell quiescence suggests that new therapeutic strategies for CML should focus on manipulation of this signaling network.

Keywords: BCR-ABL; CD34+ progenitor cells; Chronic myeloid leukemia; FOXO transcription factors; Quiescence; Tyrosine kinase inhibitors.

Figure 1

BCR-ABL affects phosphorylation and localization of FOXO isoforms in CML CD34⁺ cells. (A, left): Immuno-fluorescence was used to show the localization of FOXO3a in CML (a-d) and normal CD34⁺ cells (e–h) (green: FOXO3a; blue: nuclear DAPI; ×100 magnification). Three-dimensional images for single cells are shown (d and h). (A, right) This was confirmed by quantification of fluorescence intensity in cytoplasm versus nucleus (n = 11 and n = 37 cells for CML and normal, respectively). (B): Levels of FOXO1, 3a, and 4 in CML and normal CD34⁺ cells were measured by Q-PCR (n = 3) as shown by the difference in the Ct values (DCT). (C): Phosphorylation of FOXO1, 3a, and 4 was measured by flow cytometry in CML and normal CD34⁺ cells (upper panel). Isotype histogram is shown for the three forms of FOXO since the same secondary anti-Rabbit FITC was used to detect all three antibodies. Bar graphs for three independent experiments showing ratio of sample over isotype for total (lower panel, Right) and phosphorylated FOXO1, 3a, and 4 (lower panel, Left) are given (n = 3, ***, p < .001; **, p < .01). Abbreviations: CML, chronic myeloid leukemia; DAPI, 4′6-diamidino-2-phenylindole; GM, geometric mean.

Figure 2

FOXO1, 3a, and 4 phosphorylation are affected by TKI treatment in CD34⁺ chronic myeloid leukemia (CML) cells. (A): CD34⁺ CML cells were treated with 150 nM DAS for 24 hours and cell cycle was analyzed by flow cytometry after PI staining (representative plot, n = 4) (B) Costaining with Ki67/7-AAD was carried out to determine the proportion of quiescent, Ki67-negative cells (representative plot, n = 3). (C): Levels of pFOXO1, 3a, and 4 (left, lower panel) (n = 4, *, p ≤ .05), and total FOXO1, 3a, and 4 (right, lower panel) (n = 4) following treatment of CD34⁺ CML cells with 150 nM DAS for 24 hours, as measured by flow cytometry. Representative flow cytometry histograms are shown in the upper panel. (D): CD34⁺ CML cells were either left untreated (a, c, e, g, and i) or treated with 150 nM DAS for 24 hours (k, m, o, q, and s). Localization of FOXO1, 3a, and 4 and of p-FOXO1 and 4 was analyzed by immuno-fluorescence (overlay image is shown. Green: FOXOs; blue: nuclear DAPI; ×100 magnification). Three-dimensional images of the fluorescence intensity are shown (b, d, f, h, j and l, n, p, r, t). (E): Quantification of cytoplasmic versus nuclear fluorescence for total FOXO1, 3a, and 4 (n = 6–27 cells, **, p < .01 and ***, p < .001, respectively). Abbreviations: DAS, dasatinib; NDC, no drug control.

Figure 3

Expression of FOXO target genes are altered following TKI treatment. (A): Analysis of an Affymetrix U133A microarray chips showing downstream targets of FOXO1, 3a, and 4 which were significantly modulated following 16 hours exposure to 150 nM DAS in CD34⁺ cells. CCN1D/Cyclin D1, BCL6, p57/CDKN1C, and ATM are highlighted. Interactions were taken from MetaCore knowledge base (22/08/13). (B): CD34⁺ CML cells were treated with TKIs (pool of samples treated with imatinib 5 µM, dasatinib 150 nM, and nilotinib 5 µM) for 24 hours and mRNA levels of ATM, BCL6, and p57/CDKN1C were measured by Q-PCR (three independent samples were analyzed in triplicate, ***, p < .001; *, p ≤ .05). (C): Levels of CCN1D/Cyclin D1 mRNA and protein were analyzed after 24 hours of 150 nM DAS treatment. A representative flow cytometry histogram is shown for Cyclin D1 protein (n = 3, ***, p < .001; *, p < .05). Abbreviations: DAS, dasatinib; NDC, no drug control; Q-PCR, quantitative PCR; TKI, tyrosine kinase inhibitor.

Figure 4

Dasatinib treatment in SCLtTA/BCR-ABL mice induces FOXO1 and 3a activation in vivo. (A): Experimental layout for the leukemia induction in SCLtTA/BCR-ABL mice and plan of investigation. (B): Percentage of Gr1 and Mac1 in recipient mice after induction of BCR-ABL expression. (C): Percentage of Gr1 and Mac1 in recipient mice after 6 days of treatment with 50 mg/kg of dasatinib. (D): Level of phosphorylation of FOXO1, 3a, and 4 in c-Kit⁺ cells derived from dasatinib and untreated control mice. (E): Cell cycle analysis by PI staining carried out in c-Kit⁺ cells derived from dasatinib and untreated control mice. (F): Localization of FOXO3a analyzed by immuno-fluorescence (overlay image is shown. Green: FOXO3a; blue: nuclear DAPI; ×100 magnification). Three-dimensional images of the fluorescence intensity are shown. (G): mRNA levels of CCN1D/Cyclin D1 were measured by quantitative PCR in c-Kit⁺ cells derived from dasatinib and untreated control mice (n = 3/arm) (*, p < .05; **, p < .01). Abbreviations: DAS, dasatinib; DAPI, 4′6-diamidino-2-phenylindole; NDC, no drug control.

Figure 5

Over-expression of FOXO3a induces strong inhibition of proliferation in the K562 cell line. (A): FOXO3a WT, FOXO3a TM (a constitutively active mutant), and vector control were transiently over-expressed in K562 cells and cytoplasmic versus nuclear fractionation carried out. Following isolation, 10 µg of nuclear and cytoplasmic lysates were separated by SDS PAGE and the distribution of FOXO3a determined by Western blotting. GAPDH was used as a cytoplasmic marker. (B): Cells transfected with FOXO3a WT or TM were treated or not with dasatinib (DAS, 10 nM) for 24 hours and BrdU used to assess proliferation (n = 3, **, p < .01). (C): Levels of apoptosis were determined by Annexin V/7-AAD staining (n = 3, *, p ≤ .05). (D): CML CD34⁺ cells were treated for 24 hours with rapamycin (10 nM), LY294002 (25 µM), DAS (150 nM), or the combination of DAS with either LY294002 or rapamycin, and FOXO1, 3a, and 4 phosphorylation were measured by flow cytometry (n = 3). Statistical analysis for each treatment is reported in Supporting Information Figure S2. Abbreviations: DAS, dasatinib; NDC, no drug control.

Figure 6

Stable transfection of a FOXO3a active mutant into K562 cells increases quiescence. K562 cells were stably transfected with FOXO3a WT, FOXO3a TM, or an empty vector control. (A): Cells were lysed in RIPA buffer and 20 µg of each were separated by SDS-PAGE, Western blotting was used to show levels of FOXO3a and Actin. Fold change relative to Actin is shown. (B): Localization of FOXO3a WT and TM was determined by IF (green: FOXO3a, blue: nuclear DAPI, ×100 magnification) and (C) quantification of nuclear to total fluorescence (n = 31–45 cells, **, p < .01, *, p ≤ .05). (D): BrdU incorporation was used to measure proliferation in stable cell lines (n = 3, *, p ≤ .05), while (E) shows the percentage of Ki67-negative cells as determined by Ki67/7-AAD staining (n = 3, *, p ≤ .05). To compare this to quiescence induced by TKIs, cell lines were treated for 72 hours with IM (1 µM), DAS (10 nM), or NIL (50 nM) before being stained for Ki67/7-AAD. (F): Shows the number of Ki67-negative cells following treatment (n = 3, **, p < .01; *, p ≤ .05 relative to vector only NDC). (G): Levels of CCND1/Cyclin D1 (left, n = 3 *, p < .001) and p57/CDKN1C (right, n = 3) mRNA were determined by quantitative PCR. Abbreviations: DAS, dasatinib; DAPI, 4′6-diamidino-2-phenylindole; IM, imatinib; NDC, no drug control; NIL, nilotinib.

Figure 7

FOXO3a knockdown decreases quiescence and increases sensitivity to tyrosine kinase inhibitors (TKIs). (A): K562 cells were transfected with a FOXO3a shRNA plasmid containing a green fluorescent protein (GFP) tag (sh-FOXO3a) or a scrambled control (sh-Scramb). Twenty-four hours after transfection, GFP-positive cells were selected and knockdown of FOXO3a was confirmed by quantitative PCR (n = 3, *, p ≤ .05). (B): To determine the effect on sensitivity to TKIs, GFP-sorted cells were treated with dasatinib (DAS, 10 nM) for 72 hours and the level of apoptosis measured by Annexin V/DAPI staining (n = 3, *, p ≤ .05). (C): Staining with Ki67 and DAPI was used to determine the number Ki67 negative, quiescent cells following knockdown (left, n = 3, *, p ≤ .05). Representative dot-plots are shown (right). Abbreviations: DAS, dasatinib; DAPI, 4′6-diamidino-2-phenylindole; NDC, no drug control.