hnRNP A1 nucleocytoplasmic shuttling activity is required for normal myelopoiesis and BCR/ABL leukemogenesis

Abstract

hnRNP A1 is a nucleocytoplasmic shuttling heterogeneous nuclear ribonucleoprotein that accompanies eukaryotic mRNAs from the active site of transcription to that of translation. Although the importance of hnRNP A1 as a regulator of nuclear pre-mRNA and mRNA processing and export is well established, it is unknown whether this is relevant for the control of proliferation, survival, and differentiation of normal and transformed cells. We show here that hnRNP A1 levels are increased in myeloid progenitor cells expressing the p210(BCR/ABL) oncoprotein, in mononuclear cells from chronic myelogenous leukemia (CML) blast crisis patients, and during disease progression. In addition, in myeloid progenitor 32Dcl3 cells, BCR/ABL stabilizes hnRNP A1 by preventing its ubiquitin/proteasome-dependent degradation. To assess the potential role of hnRNP A1 nucleocytoplasmic shuttling activity in normal and leukemic myelopoiesis, a mutant defective in nuclear export was ectopically expressed in parental and BCR/ABL-transformed myeloid precursor 32Dcl3 cells, in normal murine marrow cells, and in mononuclear cells from a CML patient in accelerated phase. In normal cells, expression of this mutant enhanced the susceptibility to apoptosis induced by interleukin-3 deprivation, suppressed granulocytic differentiation, and induced massive cell death of granulocyte colony-stimulating factor-treated cultures. In BCR/ABL-transformed cells, its expression was associated with suppression of colony formation and reduced tumorigenic potential in vivo. Moreover, interference with hnRNP A1 shuttling activity resulted in downmodulation of C/EBPalpha, the major regulator of granulocytic differentiation, and Bcl-X(L), an important survival factor for hematopoietic cells. Together, these results suggest that the shuttling activity of hnRNP A1 is important for the nucleocytoplasmic trafficking of mRNAs that encode proteins influencing the phenotype of normal and BCR/ABL-transformed myeloid progenitors.

FIG. 1.

hnRNP A1 expression in normal and BCR/ABL-transformed cells. (A) Northern (top panel) and Western blot (bottom panel) analysis of hnRNP A1 expression in parental and BCR/ABL-expressing 32Dcl3 cells in the presence of IL-3 (lanes 1 and 3) or after IL-3 deprivation (12 h) (lanes 3 and 4). rRNA and HSP90 levels were used as controls for RNA and protein loading, respectively. (B) Western blot showing expression of hnRNP A1, BCR/ABL, and GRB2 in the CD34⁺ and CD34⁻ fractions of mononuclear cells from a CML-AP patient (lanes 1 and 2) and in samples of mononuclear marrow cells from four CML-CP and four CML-BC patients (lanes 3 to 10).

FIG. 2.

Role of BCR/ABL in the regulation of hnRNP A1 levels. (A) Left panel, stability of HA-tagged wild-type hnRNP A1 in exponentially growing parental and BCR/ABL-expressing 32Dcl3 cells. The half-life (t1/2) of hnRNP A1 was assessed by pulse-chase assay and quantitated by densitometry. Each point on the graph represents the relative amount of hnRNP A1 during the chase period; half-lives were calculated using the formula given in Materials and Methods. Right, levels of HA-tagged wild-type hnRNP A1 in parental and BCR/ABL-expressing cells treated with cycloheximide (CHX). (B) Effect of the proteasome inhibitor ALLN (lanes 1 to 6) and the ABL tyrosine kinase inhibitor STI571 (lanes 7 and 8) on endogenous hnRNP A1 levels in IL-3-deprived (8 h) parental and BCR/ABL-expressing cells. hnRNP A1 was detected with the 9H10 monoclonal antibody (38). HSP90 levels were monitored as a control for equal loading. Data are representative of those from three different experiments.

FIG. 3.

Generation and expression of a nucleus-localized shuttling-deficient hnRNP A1 mutant. (A) Schematic representation of wild-type (WT-A1-HA) and mutant (A1-G274A-HA and NLS-A1-HA) hnRNP A1 constructs. Amino acid sequences of the hnRNP A1 M9 domain and of hnRNP K bipartite-basic NLS are boxed. NI, nuclear import; NE, nuclear export. (B) Anti-HA immunofluorescence shows the subcellular localization of WT-A1-HA, A1-G274A-HA, and NLS-A1-HA in transiently transfected 293T cells untreated or treated with actinomycin D (Act. D). (C) Effect of WT-A1-HA and NLS-A1-HA expression on nuclear (Nucl.) and cytoplasmic (Cytopl.) levels of HA-tagged FUS. Western blots show expression of HA-tagged FUS, HA-tagged wild-type (WT-A1-HA) and mutant (NLS-A1-HA) hnRNP A1, hnRNP C1/2, and HSP90 in nuclear and cytoplasmic fractions of 293T cells transiently transfected with the indicated plasmids. Expression of hnRNP C1/2 was used as a nuclear marker, while HSP90 was used as a cytoplasmic marker. Data are representative of those from three independent experiments. (D) Expression of wild-type and mutant hnRNP A1 in two clones of parental (lanes 1 to 5) and BCR/ABL-expressing (lanes 5 to 8) 32Dcl3 cells infected with the WT-A1-HA or the NLS-A1-HA retrovirus. The inset shows levels of NLS-A1-HA hnRNP A1 mutant in total lysates (lane T) and in nuclear (lane N) and cytoplasmic (lane C) fractions of parental and BCR/ABL-expressing 32Dcl3 cells.

FIG. 4.

Requirement of hnRNP A1 shuttling activity for survival and colony formation of myeloid precursor 32Dcl3 cells and primary murine marrow cells. (A) Effect of IL-3 deprivation (left) and G-CSF treatment (right) on the viability of parental and derivative cell lines ectopically expressing wild-type hnRNP A1 (32D-WT-A1-HA) or the nucleus-localized, shuttling-deficient hnRNP A1 mutant (32D-NLS-A1-HA) or coexpressing wild-type and shuttling-deficient hnRNP A1 (32D-NLS-A1-HA/WT-A1-HA). Each point represents the mean and standard deviation from three independent experiments. Cell death percentage was determined by trypan blue exclusion. (B) Methylcellulose colony formation, in the absence or in the presence of different concentrations of WEHI-3B conditioned medium used as a source of IL-3, from 32Dcl3 and 32D-NLS-A1-HA cells (10³ cells/plate). Values are means and standard deviations for duplicate cultures from two independent experiments. (C) Clonogenic efficiency in the absence of growth factors or in the presence of increasing concentrations of WEHI conditioned medium or recombinant human G-CSF of murine mononuclear marrow cells (BMC) transduced with the empty LXSP or with the NLS-A1-HA retrovirus. After infection, cells (10⁵ cells/plate) were plated in semisolid medium in the presence of 1.25 μg of puromycin per ml. The results are representative of those from two experiments performed in duplicate.

FIG. 5.

Requirement of hnRNP A1 shuttling activity for granulocytic differentiation of 32Dcl3 cells. (A) Representative microphotographs of May-Grunwald-Giemsa-stained cytospins of G-CSF-treated parental and 32Dcl3-derived cell lines. (B) Effect of WT-A1-HA and NLS-A1-HA expression on protein levels (left panels) of Bcl-2, Bcl-X_L, C/EBPα, G-CSFR, and FUS and on mRNA levels (right panel) of Bcl-X_L and c/ebpα. Bcl-X_L cytoplasmic mRNA levels were detected by RT-PCR (see Materials and Methods); actin levels are shown as a control for equal loading. c/ebpα cytoplasmic mRNA levels were detected by Northern blotting using the murine 3′ untranslated region as a probe. rRNA levels are shown as a control for equal loading. The results are representative of those from three different experiments. (C) Western blot show expression of HA-tagged wild-type hnRNP A1 (lane 3) or C/EBPα (lane 2) in 32D-NLS-A1-HA cells. (D) G-CSF-stimulated granulocytic differentiation of 32D-NLS-A1-HA cells coexpressing WT-A1-HA or C/EBPα. Representative microphotographs of May-Grunwald-Giemsa-stained cytospins are shown.

FIG. 6.

Requirement of hnRNP A1 shuttling activity for colony formation and tumorigenesis of BCR/ABL-transformed cells. (A) Methylcellulose colony formation, in the absence or in the presence of different concentration of WEHI-3B conditioned medium used as a source of IL-3, from 32D-BCR/ABL and 32D-BCR/ABL-NLS-A1-HA cells (10⁴ cells/plate). Values are means and standard deviations for duplicate cultures from two independent experiments. (B) Clonogenic efficiency in the absence of growth factors or in the presence of increasing concentrations of recombinant human IL-3 or G-CSF of primary CML-AP^CD34+ cells transduced with the empty LXSP or with the NLS-A1-HA retrovirus. After infection, cells (5 × 10⁴ cells/plate) were plated in semisolid medium in the presence of 1.25 μg of puromycin per ml. Inset, Western blots show expression of NLS-A1-HA, Bcl-X_L, and GRB2 in vector- and NLS-A1-HA-transduced CML-AP^CD34+ cells. (C) Expression of Bcl-X_L protein (first panel) and mRNA (fourth panel) in 32Dcl3, 32D-NLS-A1-HA, 32D-BCR/ABL, and 32D-BCR/ABL-NLS-A1-HA cells. Levels of p210 BCR/ABL, HSP90, and actin were monitored as controls. Bcl-X_L cytoplasmic mRNA levels were detected by RT-PCR (see Materials and Methods). (D) Subcutaneous tumors in SCID mice injected with 32D-BCR/ABL and 32D-NLS-A1-HA cells. The latency time (days) and tumor weight (means and standard deviations) were calculated; P < 0.01. The results are representative of those from two independent experiments.