A MAPK/HNRPK pathway controls BCR/ABL oncogenic potential by regulating MYC mRNA translation

Abstract

Altered mRNA translation is one of the effects exerted by the BCR/ABL oncoprotein in the blast crisis phase of chronic myelogenous leukemia (CML). Here, we report that in BCR/ABL+ cell lines and in patient-derived CML blast crisis mononuclear and CD34+ cells, p210(BCR/ABL) increases expression and activity of the transcriptional-inducer and translational-regulator heterogeneous nuclear ribonucleoprotein K (hnRNP K or HNRPK) in a dose- and kinase-dependent manner through the activation of the MAPK(ERK1/2) pathway. Furthermore, HNRPK down-regulation and interference with HNRPK translation-but not transcription-regulatory activity impairs cytokine-independent proliferation, clonogenic potential, and in vivo leukemogenic activity of BCR/ABL-expressing myeloid 32Dcl3 and/or primary CD34+ CML-BC patient cells. Mechanistically, we demonstrate that decreased internal ribosome entry site (IRES)-dependent Myc mRNA translation accounts for the phenotypic changes induced by inhibition of the BCR/ABL-ERK-dependent HNRPK translation-regulatory function. Accordingly, MYC protein but not mRNA levels are increased in the CD34+ fraction of patients with CML in accelerated and blastic phase but not in chronic phase CML patients and in the CD34+ fraction of marrow cells from healthy donors. Thus, BCR/ABL-dependent enhancement of HNRPK translation-regulation is important for BCR/ABL leukemogenesis and, perhaps, it might contribute to blast crisis transformation.

Figure 1.

Effect of BCR/ABL on HNRPK mRNA and protein expression. (Ai) Northern blot shows expression of Hnrpk mRNA in parental, untreated, and imatinib-treated BCR/ABL-expressing myeloid progenitor 32Dcl3 cells. (Aii) RT-PCR on nuclear RNA shows levels of Hnrpk pre-mRNA in parental and BCR/ABL-expressing myeloid progenitor 32Dcl3 cells (+RT). The amplified PCR product is comprehensive of part of exon 2, intron 2, and part of exon 3 of the mouse Hnrpk gene.-RT indicates that PCR was performed on non-reverse-transcribed nuclear RNA. GRB2 mRNA levels were measured as control. (Aiii) Effect of actinomycin D on Hnrpk mRNA levels in parental and BCR/ABL-expressing 32Dcl3 cells. Cells were incubated for the indicated time with actinomycin D used at a concentration (5 μg/mL) that reportedly inhibits RNA polII-dependent transcription. (Bi) Western blots show HNRPK protein levels in parental 32Dcl3 cells (lane 1) and in untreated (lane 2) and imatinib-treated (lane 3) 32D-BCR/ABL cells, in BCR/ABL-induced (doxycycline 1 mg/mL, 24 hours) TonB210.1 lymphoid progenitors (lanes 4-5), and in untreated and imatinib-treated Ph1 K562 cells (lanes 6-7). (C) HNRPK protein stability was determined by anti-HNRPK Western blots in untreated and imatinib-treated 32D-BCR/ABL cells exposed to the protein synthesis inhibitor cycloheximide (CHX). GRB2 levels were monitored as control for equal loading. (Di) Plot shows HNRPK protein levels expressed as log of arbitrary densitometric units after normalization with GRB2 levels in mononuclear cells from CML patients in the chronic (CML-CP) and blast crisis (CML-BC) phases. (P = .001; Wicoxon rank sum test). (Dii) Graphs shows GRB2-normalized HNRPK protein levels expressed as densitometric units in the CD34⁺ fraction of bone marrow from 2 healthy donors (NBM), and 2 CML-CP, 1 CML-AP, and 2 CML-BC patients. The Western blot in the inset shows total tyrosine phosphorylation, and the arrow indicates levels of active BCR/ABL. (Diii) HNRPK protein expression in untreated and imatinib-treated CD34⁺ CML-BC patient marrow cells. GRB2 was used as control for equal loading.

Figure 2.

Role of MAPK in the regulation of HNRPK expression in BCR/ABL⁺ cells. (A) Effect of different chemical inhibitors of known BCR/ABL-activated pathways on HNRPK protein levels in 32D-BCR/ABL and K562 cells. (B) Effect of the MEK1/MAPK inhibitor PD098059 (lane 2) on Hnrpk mRNA (first row) and protein (third row) expression. Hnrpk mRNA (first row) and protein (third row) levels in parental 32Dcl3 cells (lane 3), vector-transduced 32D-BCR/ABL (lane 4), and 32D-BCR/ABL cells ectopically expressing wild-type MAPK ERK1 (lane 5), the K71R dominant-negative MAPK ERK1 mutant (lane 6), wild-type ERK2 (lane 7), and dominant-negative K52R ERK2 mutant (lane 8). rRNA and GRB2 are shown as controls for equal loading. (C) Effect of BCR/ABL tyrosine kinase on MAPK/ERK activity. pERK1/2 indicates phosphorylated (active) ERK1 and ERK2 mitogen-activated protein kinases. Western blots show levels of HNRPK, phospho-ERK1/2, total ERK1/2, BCR/ABL, and GRB2 in 32Dcl3 myeloid precursors transduced with a GFP/BCR/ABL bicistronic retrovirus and sorted for low, medium, and high GFP levels. (D) Levels of HNRPK, phospho-ERK1/2, BCR/ABL, and GRB2 in mononuclear marrow cells from paired chronic phase (CP) and blast crisis (BC) CML patient samples. (E) Levels of pERK1/2, ERK1/2, and GRB2 in NBM^CD34+, CML-CP^CD34+, CML-AP^CD34+, and CML-BC^CD34+samples.

Figure 3.

In vitro and in vivo requirement of HNRPK for BCR/ABL oncogenic potential. (Ai) HNRPK levels in parental 32Dcl3, 32D-BCR/ABL, and 32D-BCR/ABL cells transduced with an antisense (lane 3) or an shRNA (lane 5) HNRPK retroviral construct. (Aii) Effect of HNRPK down-regulation by shRNA (lane 3) on the levels of the HNRPK- and BCR/ABL-regulated Myc protein. (B) Effect of HNRPK down-regulation on the growth factor-independent proliferation of 32D-BCR/ABL cells. (C) Effect of Hnrpk antisense and shRNA on the growth factor-independent clonogenic efficiency of 32D-BCR/ABL cells. Bars represent mean ± SE of data from 3 independent experiments. (D) 32D-BCR/ABL and GFP⁺ 32D-BCR/ABL cells retrovirally transduced with pSuper-retroneo-Hnrpk shRNA were induced to differentiate with G-CSF. Cells were subjected to cytospin and were May-Grünwald/Giemsa stained. 32Dcl3 parental cells were used as a control for granulocyte colony-stimulating factor (G-CSF)-induced granulocytic differentiation. Images were taken with a Zeiss Axioskope 2 Plus and a 40×/0.75 NA objective, with a Canon Powershot A70 (Canon, Lake Success, NY) and Canon Capture software.

Figure 4.

HNRPK transcription-regulatory activity is dispensable for BCR/ABL oncogenic potential. (A) Western and Northern blots show levels of Myc and/or HA-tagged wild-type and mutant C299-HNRPK in 32D-BCR/ABL cells transduced with the empty MigR1 vector (lane 1), wild-type HNRPK (lane 2), and C299-HNRPK dominant-negative mutant (lane 3). (B) Confocal microscopy of anti-HA-stained wild-type and C299-HNRPK-expressing 32D-BCR/ABL cells. (C) Effect of wild-type and C299-HNRPK on growth factor-independent clonogenic efficiency of 32D-BCR/ABL cells.

Figure 5.

HNRPK translation-regulatory activity is essential for BCR/ABL oncogenic potential. (Ai) Confocal microscopy on anti-HNRPK-stained 32D-BCR/ABL cells untreated and treated with the MAPK inhibitor PD098059 or with imatinib. (Aii) Effect of imatinib on the BCR/ABL-induced phosphorylation of MAPK ERK1/2. (Bi) Confocal microscopy on anti-HA-stained 32D-BCR/ABL cells ectopically expressing HA-tagged wild-type, dominant-active S284/353D, and dominant-negative S284/353A mutant HNRPK proteins. (Bii) Levels of ectopic wild-type and mutant HNRPK proteins. (C) Effect of wild-type, S284/353D, and S284/353A mutant HNRPK on growth factor-independent proliferation (i) and cytokine-dependent and independent clonogenic efficiency (ii) of 32D-BCR/ABL cells. Bars represent the mean ± SE of values from 3 different experiments. (Di) Effect of wild-type and HNRPK mutants on the tumorigenic potential of 32D-BCR/ABL cells. Cells (10⁶) were injected subcutaneously into SCID mice; tumor appearance and growth were monitored daily and mice were killed 15 days after injection. Bars represent the mean ± SE from the weight of 5 tumors for each group. (Dii) Effect of wild-type and S284/353A-HNRPK on the leukemogenic process induced in SCID mice by intravenous injection of BCR/ABL-expressing cells. Kaplan-Meier plot shows survival of mice injected with 32D-BCR/ABL cells expressing wild-type or S284/353A-HNRPK. The log-rank test evaluated the differences among survival distributions (P < .001). Inset shows spleens of 3 mice of each group killed 4 weeks after cell injection.

Figure 6.

HNRPK binding to Myc mRNA and effect of ectopic Myc expression on the leukemic phenotype of S284/353A-HNRPK-expressing 32D-BCR/ABL cells. (Ai) RT-PCR shows presence of Myc mRNA in anti-HNRPK and anti-HA immunoprecipitates from lysates of parental (lane 2) and S284/353(A/D) (lanes 3-4) HNRPK-expressing 32D-BCR/ABL cells, respectively. Lane 1 shows levels of Myc in total mRNA, whereas lane 5 represents RT-PCR of mRNA immunoprecipitated with a nonrelated (anti-Flag) antibody; (Aii) REMSA (top) and UV cross-linking (bottom) with cytoplasmic extracts of 32Dcl3, 32D-BCR/ABL, and K562 cells show binding of HNRPK to the HNRPK binding site contained in the MYC IRES element. (B) Graphs shows firefly luciferase activity after normalization with Renilla luciferase activity in 32D-BCR/ABL cells transduced with the empty pRF and with the pRMF(IRES-MYC) plasmid containing the MYC IRES element cloned in front of the firefly luciferase cDNA. (C) Effect of ectopic MYC expression on the growth factor-independent colony formation (i) and tumorigenic potential (ii) of S284/353A-HNRPK-expressing 32D-BCR/ABL cells. Inset in panel Ci shows total Myc protein levels in vector-transduced, S284/353A-HNRPK/MYC-transduced, and S284/353A-HNRPK/MYC-transduced 32D-BCR/ABL cells. (D) Effect of MYC overexpression on the growth factor-independent clonogenic potential of CD34⁺ bone marrow CML-BC patient cells ectopically expressing the S284/353A-HNRPK mutant. Primary CD34⁺ CML-BC marrow cells were transduced with the MigR1-S284/353A-HNRPK-HA construct, selected for GFP⁺, and infected again with the MSCV-puro-MYC retrovirus. In panels B-D, bars represent mean ± SE of values from 3 different experiments.

Figure 7.

BCR/ABL- and MAPK-dependent phosphorylation of HNRPK regulates Myc mRNA translation in BCR/ABL-transformed myeloid cells. (A) Western blots show Myc levels in parental 32Dcl3 cells (lane 1), untreated 32D-BCR/ABL cells (lane 6); or cells treated with imatinib (lane 7) or with the MEK1 inhibitor PD098059 (lane 8); or cells transduced with the empty retrovirus (lane 2) or with wild-type and mutant HNRPK (lanes 3-5) or with MAPK ERK1 (lanes 9-10) and ERK2 (lanes 11-12) cDNAs. All cell lines were cultured in the presence of IL-3. (B) Northern blots show Myc mRNA levels in the indicated cell lines either maintained in IL-3 or IL-3 deprived for 12 hours in the presence or absence of imatinib. (C) Autoradiography and anti-MYC Western blot show levels of ³⁵S-labeled newly synthesized and total immunoprecipitated Myc protein in parental 32Dcl3, untreated, and imatinib-treated 32D-BCR/ABL and S284/353A expressing 32D-BCR/ABL cells. (D) Levels of Myc protein, expressed as arbitrary densitometric units after normalization with GRB2 levels, in cycloheximide-treated parental 32Dcl3, 32D-BCR/ABL, and its derivative cell lines expressing wild-type or mutant HNRPK proteins. Half-life was calculated using the algorithm t½ = 0.693t_n/ln(C₀/C_n), where n represents the last time point. (E) Histograms show levels of MYC mRNA (red bars) and protein (blue bars) after normalization with GAPDH mRNA and GRB2 protein levels, respectively. Values in samples 2 to 7 represent arbitrary densitometric units expressed as percentage of those of sample 1. Western blot and RT-PCR were performed with lysate and total RNA, respectively, of CD34⁺marrow cells from healthy individuals (lanes 1-2) and from CML-CP (lanes 3-4), CML-AP (lane 5), and CML-BC (lanes 6-7) patients.