NPM1-fusion proteins promote myeloid leukemogenesis through XPO1-dependent HOX activation

Abstract

Nucleophosmin (NPM1) is a nucleolar protein and one of the most frequently mutated genes in acute myeloid leukemia (AML). In addition to the commonly detected frameshift mutations in exon12 (NPM1c), previous studies have identified NPM1 gene rearrangements leading to the expression of NPM1-fusion proteins in pediatric AML. However, whether the NPM1-fusions are indeed oncogenic and how the NPM1-fusions cause AML have been largely unknown. In this study, we investigated the subcellular localization and leukemogenic potential of two rare NPM1-fusion proteins, NPM1::MLF1 and NPM1::CCDC28A. NPM1::MLF1 is present in both the nucleus and cytoplasm and occasionally induces AML in the mouse transplantation assay. NPM1::CCDC28A is more localized to the cytoplasm, immortalizes mouse bone marrow cells in vitro and efficiently induces AML in vivo. Mechanistically, both NPM1-fusions bind to the HOX gene cluster and, like NPM1c, cause aberrant upregulation of HOX genes in cooperation with XPO1. The XPO1 inhibitor selinexor suppressed HOX activation and colony formation driven by the NPM1-fusions. NPM1::CCDC28A cells were also sensitive to menin inhibition. Thus, our study provides experimental evidence that both NPM1::MLF1 and NPM1::CCDC28A are oncogenes with functions similar to NPM1c. Inhibition of XPO1 and menin may be a promising strategy for the NPM1-rearranged AML.

Fig. 1. Cytoplasmic localization of the NPM1-fusion proteins.

NES nuclear export signal. AS acidic domains. NLS nuclear localization signal. A Schematic presentation of wild-type NPM1, NPM1::MLF1, NPM1::CCDC28A. Numbers indicate positions of amino acid residues from the N-terminus. B HEK293T cells were transduced with the flag-tagged NPM1::MLF1 and NPM1::CCDC28A. The cell lysates were stained with anti-Flag antibody. C K562 cells were transduced with vector, flag-tagged wild-type NPM1, NPM1c, NPM1::MLF1 or NPM1::CCDC28A, and were then stained with anti-Flag. Cell Nuclei were visualized with DAPI. Scale bar, 20 μm. WT wild-type.

Fig. 2. NPM1::MLF1 and NPM1::CCDC28A show leukemogenic activity.

A Colony numbers of vector, NPM1::MLF1 or NPM1::CCDC28A-transduced mouse bone marrow c-Kit⁺ cells. Weekly colony count per 1 × 10⁴ replated cells are shown. Data are shown as mean ± SEM. B Wright-Giemsa staining of cells at the third round. All vector or NPM1::MLF1-transduced cells were differentiated macrophages or mastocytes, whereas NPM1::CCDC28A-transduced colonies contained myeloblasts. Scale bar, 20 μm. C Frequency of GFP⁺ cells in peripheral blood of all recipient mice bearing cells expressing vector (n = 6), NPM1::MLF1 (n = 12) or NPM1::CCDC28A (n = 6). PB peripheral blood. D White blood cells, hemoglobin, platelet and mean corpuscular volume levels of the recipient mice 6, 10 or 14 weeks after transplantation are shown. Data from all recipient mice with vector-transduced cells, and mice with NPM1::MLF1 or NPM1::CCDC28A-expressing cells in which the GFP⁺ cells increased after transplantation were used (vector, n = 6, NPM1::MLF1, n = 2, NPM1::CCDC28A, n = 6). Data are shown as mean ± SEM. One-way ANOVA was used for multiple comparisons. WBC white blood cells, Hb hemoglobin, Plt platelet, MCV mean corpuscular volume. E Kaplan-Meier survival curve of mice transplanted with bone marrow progenitor cells transduced with NPM1::MLF1 (n = 12), NPM1::CCDC28A (n = 6), or the vector (n = 6).

Fig. 3. NPM1::MLF1 and NPM1::CCDC28A induce the development of transplantable AML in vivo.

A–C Bone marrow cells were collected from the moribund mice bearing NPM1::MLF1 or NPM1::CCDC28A-expressing cells. BM bone marrow. A Representative flow cytometry profiles (right) and a bar graph showing the frequency of GFP⁺CD11b⁺ cells (left). Data are shown as means ± SEM. (NPM1::MLF1: n = 2, NPM1::CCDC28A: n = 6). B Representative flow cytometry profiles (right) and a bar graph showing the frequency of GFP⁺c-kit⁺ cells and GFP⁺ granulocytic-monocytic progenitors are shown. Data are shown as means ± SEM. Numbers indicate the frequency of each population. GMP granulocytic-monocytic progenitors. CMP common myeloid progenitors. MEP megakaryocyte/erythrocyte progenitors. C Wright-Giemsa staining of bone marrow cells collected from the recipient mice bearing NPM1::MLF1 or NPM1::CCDC28A-expressing cells. Scale bar, 20 μm. D Kaplan-Meier survival curve of secondary recipient mice transplanted with the NPM1::MLF1 or NPM1::CCDC28A-expressing cells (NPM1::MLF1, n = 6, NPM1::CCDC28A, n = 6).

Fig. 4. NPM1::MLF1 and NPM1::CCDC28A induced specific and shared gene expression signatures.

A Mouse bone marrow c-kit⁺ cells were transduced with vector, NPM1::MLF1 or NPM1::CCDC28A and were transplanted into recipient mice. GFP⁺ bone marrow cells were harvested two months after transplantation to evaluate their gene expression profiles. Principal component analysis of GFP⁺ cells expressing vector, NPM1::MLF1 or NPM1::CCDC28A is shown. FC fold change. B Venn diagram showing upregulated (upper panel) or downregulated (lower panel) genes in cells expressing NPM1::MLF1 or NPM1::CCDC28A. C A volcano plot showing of up- or down-regulated genes in NPM1::MLF1 or NPM1::CCDC28A-expressing cells compared to vector-transduced cells. Each dot represents one gene. D Venn diagram showing the overlapping genes between CUT & RUN signal (FDR > 1) and upregulated genes from RNA-seq in cells expressing NPM1::CCDC28A. E Integrative Genomic Viewer screenshots of FLAG-NPM1::CCDC28A and IgG control CUT&RUN signals around HOX gene clusters.

Fig. 5. Selinexor inhibits colony formation and HOX gene activation induced by the NPM1-fusions.

A Colony numbers of NPM1::MLF1-induced leukemia treated with either 100 nM selinexor or DMSO for 7 days. Error bars indicate the standard error (SE). Two-tailed unpaired t test was used for pairwise comparisons. B Colony numbers of NPM1::CCDC28A-induced leukemia treated with either 100 nM selinexor or DMSO for 7 days. Error bars indicate the standard error (SE). Two-tailed unpaired t test was used for pairwise comparisons. C, D HOXA gene levels in cells expressing NPM1::MLF1 (C) or NPM1::CCDC28A (D) treated with/without 100 nM selinexor for 7 days. Data are shown as means ± SEM. Two-tailed unpaired t test was used for pairwise comparisons.

Fig. 6. Selinexor inhibits cytoplasmic localization and chromatin binding of NPM1-fusions.

293 T cells were transfected with Flag-tagged NPM1::MLF1 (A) or Flag-tagged NPM1::CCDC28A (B) and were treated with vehicle (DMSO) or Selinexor (100 nM) for 24 h. Cells were stained with anti-Flag antibody. Cell nuclei were visualized with DAPI. Representative images (left) are shown. Scale bar, 10 μm. Subcellular localization of NPM1::MLF1 and NPM1::CCDC28A was evaluated in 100 cells (right). Mouse AML cells expressing NPM1::MLF1 (C) or NPM1::CCDC28A (D) were treated with 100 nM selinexor for 7 days. Binding of NPM1::MLF1 (C) and NPM1::CCDC28A (D) on the HOX gene cluster was assessed by ChIP-qPCR. Data are shown as mean ± SEM from triplicate wells. Two-tailed unpaired t test was used for pairwise comparisons. E 293 T cells were cotransfected with vector, MLL::ENL, NPM1::MLF1, NPM1::CCDC28A, wild-type NPM1 or NPM1c together with pGL4.74 vector and HoxA9 reporter. Selinexor was added 24 h after transfection (right), and the HOXA9 activation was measured 48 h after transfection using the Dual-Luciferase Reporter Assay System. Data are shown as means ± SEM from triplicate wells. One-way ANOVA was used for multiple comparisons.

Fig. 7. Dose-dependent effects of XPO1 and menin inhibitors on NPM1::CCDC28A cells.

A NPM1::CCDC28A cells and mouse cKit⁺ bone marrow cells were incubated with selinexor or VTP50469 at the indicated concentration for 72 h. Cell viability was assessed using the Cell Counting Kit-8. Data are shown as means ± standard deviation (SD) from three technical replicates. NPM1::CCDC28A cells and mouse cKit⁺ bone marrow cells (B) or human cord blood CD34⁺ cells (C) were incubated with selinexor + VTP50469 at the indicated concentrations for 72 h. Cell viability was assessed using the Cell Counting Kit-8. Data are shown as means ± standard deviation (SD) from three technical replicates.