High levels of the BCR/ABL oncoprotein are required for the MAPK-hnRNP-E2 dependent suppression of C/EBPalpha-driven myeloid differentiation

Abstract

The inability of myeloid chronic myelogenous leukemia blast crisis (CML-BC) progenitors to undergo neutrophil differentiation depends on suppression of C/EBPalpha expression through the translation inhibitory activity of the RNA-binding protein hnRNP-E2. Here we show that "oncogene dosage" is a determinant factor for suppression of differentiation in CML-BC. In fact, high levels of p210-BCR/ABL are required for enhanced hnRNP-E2 expression, which depends on phosphorylation of hnRNP-E2 serines 173, 189, and 272 and threonine 213 by the BCR/ABL-activated MAPK(ERK1/2). Serine/threonine to alanine substitution abolishes hnRNP-E2 phosphorylation and markedly decreases its stability in BCR/ABL-expressing myeloid precursors. Similarly, pharmacologic inhibition of MAPK(ERK1/2) activity decreases hnRNP-E2 binding to the 5'UTR of C/EBPalpha mRNA by impairing hnRNP-E2 phosphorylation and stability. This, in turn, restores in vitro and/or in vivo C/EBPalpha expression and G-CSF-driven neutrophilic maturation of differentiation-arrested BCR/ABL(+) cell lines, primary CML-BC(CD34+) patient cells and lineage-negative mouse bone marrow cells expressing high levels of p210-BCR/ABL. Thus, increased BCR/ABL oncogenic tyrosine kinase activity is essential for suppression of myeloid differentiation of CML-BC progenitors as it is required for sustained activation of the MAPK(ERK1/2)-hnRNP-E2-C/EBPalpha differentiation-inhibitory pathway. Furthermore, these findings suggest the inclusion of clinically relevant MAPK inhibitors in the therapy of CML-BC.

Figure 1

BCR/ABL posttranslationally enhances hnRNP-E2 expression by increasing its protein stability. (A) Western blot (left panel) and RT-PCR (right panel) analyses of hnRNP-E2 expression in parental and BCR/ABL-transduced 32Dcl3 cells cultured in the presence of IL-3 or IL-3 deprived for 8 hours. (B) Effect of ALLN (25 μM) and ALLM (25 μM) on hnRNP-E2 levels. Note that 100 μg and 25 μg lysates from 32Dcl3 and 32D-BCR/ABL cells were loaded for each lane, respectively. (C) Effect of the protein synthesis inhibitor cycloheximide (CHX) on hnRNP-E2 and hnRNP-E1 levels in parental and BCR/ABL-transduced 32Dcl3 cells.

Figure 2

The BCR/ABL-activated MAPK^ERK1/2 is responsible for enhanced hnRNP-E2 expression. (A) Effect of different chemical kinase inhibitors on hnRNP-E2 protein levels in Ph1(+) EM-3, K562, 32D-BCR/ABL, and primary CML-BC^CD34+ progenitors. Graph shows mean plus and minus standard deviation (SD) of hnRNP-E2 levels after normalization with HSP90 in untreated and U0126-treated primary CML-BC^CD34+ progenitors (n = 3). Effect of wild-type and dominant-negative K71R ERK1 and K52R ERK2 expression on hnRNP-E2 levels in 32D-BCR/ABL cells. (B) Schematic diagram shows ERK phosphorylation sites (S/T-P) in hnRNP-E2. (C) In vivo MAPK-dependent phosphorylation of hnRNP-E2. Autoradiography (top) shows phosphorylated HA-hnRNP-E2 in ³²P-labeled 32D-BCR/ABL cells ectopically expressing HA-hnRNP-E2 (untreated and U0126-treated) and HA-hnRNP-E2 ^{S173A, S189A, T213A, S272A}. Anti-HA Western blot (bottom) shows expression of HA-hnRNP-E2 and HA-hnRNP-E2 ^{S173A, S189A, T213A, S272A}. Asterisks indicate IgG chains. (D) In vitro MAPK-dependent phosphorylation of hnRNP-E2. (top panel). Bacterially expressed and purified wild-type MBP-hnRNP-E2 was subjected to an in vitro kinase assay and Western blot analysis. Autoradiography (top) shows phosphorylated MBP-hnRNP-E2 by the anti-ERK1/2 immunoprecipitates. Western blot (bottom) shows expression of MBP-hnRNP-E2 and MBP. Effect of serine/threonine to alanine mutations on hnRNP-E2 phosphorylation (bottom panel). Wild-type MBP-hnRNP-E2 and its serine/threonine to alanine mutants were subjected to a kinase assay with recombinant ERK1 (rERK1) and ERK2 (rERK2), respectively. Autoradiography show phosphorylated MBP-hnRNP-E2 by ERK1 and ERK2, respectively. Coomassie blue–stained gels were used as controls for equal loading.

Figure 3

ERK-dependent phosphorylation controls hnRNP-E2 protein stability. (A) hnRNP-E2 expression in parental and BCR/ABL-transduced 32Dcl3 cells expressing an empty vector, HA-hnRNP-E2, HA-hnRNP-E2^{S173A,S189A,T213A,S272A} (Ser/Thr→Ala), and HA-hnRNP-E2^{S173D,S189D,T213E,S272D} (Ser/Thr→Asp/Glu). Cells were IL-3 deprived for 8 hours. (B) Stability of newly synthesized HA-tagged hnRNP-E2, hnRNP-E2^{S173A,S189A,T213A,S272A}, and hnRNP-E2^{S173D,S189D,T213E,S272D} in ³⁵S-labeled 32D-BCR/ABL cells. The half-life of hnRNP-E2 was assessed by pulse-chase assay and quantified by densitometry. Each point on the graph represents the relative normalized amounts of hnRNP-E2 during the chase period. (C) Stability of HA-tagged hnRNP-E2, hnRNP-E2^{S173A,S189A,T213A,S272A}, and hnRNP-E2^{S173D,S189D,T213E,S272D} in 32D-BCR/ABL cells treated with cycloheximide (CHX). HSP90 levels were used as control for equal loading.

Figure 4

Effect of ERK-dependent hnRNP-E2 down-regulation on C/EBPα expression and neutrophilic differentiation of 32D-BCR/ABL^high cells.(A) Protein levels of C/EBPα and hnRNP-E2 in newly established 32Dcl3 cell clones expressing different levels of p210-BCR/ABL (top panel). Levels of C/EBPα, hnRNP-E2, phospho-ERK1/2, total ERK1/2, and BCR/ABL in 32Dcl3 cells and untreated, imatinib-treated, and U0126-treated 32D-BCR/ABL cells expressing high levels of the p210-BCR/ABL oncoprotein (32D-BCR/ABL^high) (bottom panel). Cells were cultured in the presence of IL-3 or G-CSF. (B) REMSA with ³²P-labeled WT-uORF/spacer CEBPA oligoribonucleotide probe and cytosolic lysates from 32Dcl3 cells and untreated, imatinib-treated, and U0126-treated 32D-BCR/ABL^high cells (left panel). Expression of Gr-1 on 32Dcl3 and untreated, U0126-treated, and CI-1040–treated 32D-BCR/ABL^high cells upon G-CSF stimulation for 4 days (right panel). (C) May-Grunwald/Giemsa staining of 32Dcl3 and 32D-BCR/ABL^high cells cultured with IL-3 or G-CSF for 7 days in the presence or absence of imatinib, U0126, or CI-1040.

Figure 5

Effect of ERK-dependent suppression of hnRNP-E2 on C/EBPα expression and neutrophilic differentiation of Lin⁻ BMC-BCR/ABL^high cells and patient-derived CML-BC^CD34+ progenitors. (A) Levels of hnRNP-E2, C/EBPα, G-CSFR, phospho-ERK1/2, and total ERK1/2 in lineage-negative (Lin⁻) mouse bone marrow cells (BMCs) transduced with the empty vector (Lin⁻BMC-MigRI) or with a p210-BCR/ABL retrovirus and sorted for low and high GFP expression (Lin⁻BMC-BCR/ABL^low and -BCR/ABL^high) (left panel). Effect of MAPK inhibition by U0126 on hnRNP-E2 and C/EBPα expression and/or G-CSF–driven differentiation in Lin⁻ BMC-BCR/ABL^high cells (middle and right panels). May-Grumwald/Giemsa–stained representative microphotographs of Lin⁻ BMC-MigRI, BMC-BCR/ABL^low, untreated, and U0126-treated BMC-BCR/ABL^high cells (right panel). (B) Effect of MAPK inhibition by U0126 and CI-1040 on hnRNP-E2 and C/EBPα expression in primary CD34⁺ human normal bone marrow cells (NBM^CD34+) and CML-BC^CD34+ cells (left panel). May-Grumwald/Giemsa staining of NBM^CD34+ and untreated, U0126-treated, and CI-1040-treated CML-BC^CD34+ cells (right panel).

Figure 6

In vivo effect of pharmacologic inhibition of ERK-dependent hnRNP-E2 expression on neutrophilic maturation of differentiation-arrested myeloid 32D-BCR/ABL cells. H&E (top panels) and Leder (bottom panels) staining of the spleen sections of untreated and CI-1040–treated control and 6.15-MigRI and 6.15-WT-uORF cell–injected mice. In addition, all mice were cotreated with G-CSF. Microphotographs are representative of data obtained by analyzing multiples sections of spleens from 3 mice per group.

Figure 7

Effect of the p53 genetic background on BCR/ABL-dependent regulation of C/EBPα expression. C/EBPα levels at a different time course in nontransduced (none) and BCR/ABL-transduced wild-type and p53^−/− mouse bone marrow cells (BMCs) cultured in the presence of G-CSF for the indicated time.