Mutant ASXL1 cooperates with BAP1 to promote myeloid leukaemogenesis

Abstract

ASXL1 mutations occur frequently in myeloid neoplasms and are associated with poor prognosis. However, the mechanisms by which mutant ASXL1 induces leukaemogenesis remain unclear. In this study, we report mutually reinforcing effects between a C-terminally truncated form of mutant ASXL1 (ASXL1-MT) and BAP1 in promoting myeloid leukaemogenesis. BAP1 expression results in increased monoubiquitination of ASXL1-MT, which in turn increases the catalytic function of BAP1. This hyperactive ASXL1-MT/BAP1 complex promotes aberrant myeloid differentiation of haematopoietic progenitor cells and accelerates RUNX1-ETO-driven leukaemogenesis. Mechanistically, this complex induces upregulation of posterior HOXA genes and IRF8 through removal of H2AK119 ubiquitination. Importantly, BAP1 depletion inhibits posterior HOXA gene expression and leukaemogenicity of ASXL1-MT-expressing myeloid leukemia cells. Furthermore, BAP1 is also required for the growth of MLL-fusion leukemia cells with posterior HOXA gene dysregulation. These data indicate that BAP1, which has long been considered a tumor suppressor, in fact plays tumor-promoting roles in myeloid neoplasms.

Fig. 1

Expression of BAP1 stabilizes ASXL1-MT and induces its monoubiquitination at lysine 351. a Schematic presentation of ASXL1 and BAP1 constructs. ASXH Asx homology domain, PHD PH domain, UCH ubiquitin C-terminal hydrolase, ULD Uch37-like domain. b 293T cells were transfected with FLAG-ASXL1-MT (MT), FLAG-ASXL1-MT-K351R (KR), BAP1, and BAP1-C91S (CS). Cell lysates were subjected to western blotting analysis. c, d 293T cells were transfected with FLAG-ASXL1-MT (c) or FLAG-ASXL1-WT (d) together with vector or HA-BAP1. Forty-eight hours later, cells were treated with 50 μg/ml cycloheximide (CHX) for the indicated times, and cell extracts were analyzed with anti-FLAG and anti-alpha-tubulin antibodies. The band intensities of FLAG relative to alpha-tubulin are shown. The value FLAG/alpha-tubulin without CHX treatment was set to 1. e 293T cells were transfected with FLAG-ASXL1-MT, FLAG-ASXL1-MT-K351R, HA-BAP1, and Myc-ubiquitin. Total cell lysates were immunoprecipitated with anti-FLAG M2 antibody, and the ubiquitinated ASXL1-MT was detected with anti-Myc. f 293T cells were retrovirally transduced with 3xFLAG-tagged ASXL1-MT (coexpressing GFP), and then GFP-high or -low fraction (expressing relatively high or low level of ASXL1-MT) was sorted. Shown is a FACS plot of gating strategy for GFP-low and GFP-high fractions. g Cell lysates extracted from parent, GFP-low (expressing low level of ASXL1-MT) and GFP-high (expressing high level of ASXL1-MT) 293T cells were subjected to western blotting analysis. (*): monoubiquitinated ASXL1-MT, (#): non-ubiquitinated ASXL1-MT. h GFP-low and GFP-high 293T cells were transduced with Cas9 together with vector or sgRNAs targeting BAP1. Cell lysates were subjected to western blotting analysis. (*): monoubiquitinated ASXL1-MT, (#): non-ubiquitinated ASXL1-MT. Experiments were independently repeated at least three times with similar results. Shown are representative results

Fig. 2

UBE2O is an E3 ubiquitin that induces polyubiquitination of ASXL1-MT. a Schematic experimental procedures of nano-LC-MS/MS analysis. b Venn diagram shows the overlap of proteins bound to ASXL1-MT in the conditions with or without coexpression of BAP1. Eighty-eight proteins, including UBE2O, bound to ASXL1-MT only with the coexpression of BAP1. See also Supplementary Data 3 and 4. c 293T cells were transfected FLAG-ASXL1-MT, MYC-UBE2O (WT), and MYC-UBE2O-C885S (CS). Total cell lysates were immunoprecipitated with anti-FLAG M2 antibody, and the ubiquitinated ASXL1-MT was detected with anti-ubiquitin antibody. d 293T cells were transduced with Cas9 and a vector control or three independent sgRNAs targeting UBE2O. Cells were then transfected with FLAG-ASXL1-MT and HA-BAP1. Cell lysates were subjected to western blotting analysis. The band intensities of FLAG relative to alpha-tubulin are shown. The value of FLAG/alpha-tubulin for ASXL1-MT alone without UBE2O depletion was set to 1. e 293T cells were transfected with FLAG-ASXL1-MT, HA-BAP1 (WT), HA-BAP1-C91S (CS), and Myc-ubiquitin. Total cell lysates were immunoprecipitated with anti-FLAG M2 antibody, and the ubiquitinated ASXL1-MT was detected with anti-Myc. Experiments shown in (c–e) were independently repeated at least three times with similar results. Shown are representative results

Fig. 3

Monoubiquitinated ASXL1-MT enhances catalytic function of BAP1. a 32Dcl3 cells stably expressing HA-BAP1 together with vector, FLAG-ASXL1-MT, or FLAG-ASXL1-WT were treated with 50 μg/ml CHX for the indicated times, and cell extracts were analyzed with anti-HA, anti-FLAG, and anti-GAPDH antibodies. The band intensities of BAP1 relative to GAPDH are shown. The value of BAP1/GAPDH without CHX treatment was set to 1. ASXL1-MT, but not ASXL1-WT, stabilized BAP1. b 293T cells were transfected with FLAG-ASXL1-WT, FLAG-ASXL1-MT, HA-BAP1, and Myc-ubiquitin. Total cell lysates were immunoprecipitated with anti-HA antibody, and the ubiquitinated BAP1 was detected with anti-Myc. ASXL1-MT, but not ASXL1-WT, reduced ubiquitination of BAP1. c 293T cells were transfected with HA-BAP1 together with vector, FLAG-ASXL1-MT (MT), or FLAG-ASXL1-MT-K351R (KR), and were stained with anti-FLAG (rabbit) or anti-HA (mouse) antibody followed by secondary anti-rabbit Alexa 568 (red) or anti-mouse Alexa 488 (green) staining. Nuclei were visualized with DAPI (Blue). Confocal laser scanning microscopy (Nikon A1) was used to observe localization of ASXL1-MT and BAP1 (left). Scale bars: 10 μm. Subcellular localization of BAP1 was quantified by counting 400 cells exhibiting nuclear localization (nucleus) or diffuse distribution in both nucleus and cytoplasm (both). d–f 32Dcl3 cells were transduced with the combinations of wild-type ASXL1 (ASXL1-WT), ASXL1-MT (MT), ASXL1-MT-K351R (KR), BAP1 (BAP1-WT), and BAP1-C91S (CS). Cell lysates extracted from them were subjected to immunoblotting with anti-H2AK119ub, anti-H3K27me3, anti-total H3, and anti-GAPDH antibodies. Expression of ASXL1-MT (but not ASXL1-WT) and wild-type BAP1 (but not BAP1-C91S) strongly reduced H2AK119ub (d, e). ASXL1-MT-K351R showed attenuated activity to enhance BAP1-induced deubiquitination of H2AK119 (f). Experiments were independently repeated at least three times with similar results. Shown are representative results

Fig. 4

ASXL1-MT/BAP1 complex impairs multilineage differentiation of haematopoietic progenitors except for differentiation towards monocytes. a, b Murine c-Kit⁺ bone marrow cells were transduced with vector, ASXL1-MT (MT), or ASXL1-MT-K351R (KR) (coexpressing GFP) together with vector or BAP1 (coexpressing puromycin-resistant gene). After the puromycin selection for 48 h, cells were cultured with cytokines to induce mast cell or monocyte/macrophage differentiation. Mast cell maturation was assessed by the ratio of FcεIRa and c-Kit double-positive cells (enriched with mature mast cells) after 9 days of culture (a). Monocyte (CD115⁺F4/80⁻) and macrophage (CD115⁺F4/80⁺) differentiation was assessed after 7 days of culture (b). (n = 3 each) (c) Murine c-Kit⁺ bone marrow cells were transduced with vector, ASXL1-MT (MT), or ASXL1-MT-K351R (KR) (coexpressing blasticidin-resistant gene) together with vector or BAP1 (coexpressing puromycin-resistant gene). After the selection with puromycin and blasticidin for 3 days, cells were cultured with cytokines to induce differentiation toward both granulocytes and monocytes. Granulocyte (Gr-1⁺CD11b⁺) and monocyte (Gr-1⁻CD11b⁺) differentiation was assessed by FACS analyses (left) and Wright–Giemsa staining (right, scale bars: 20 μm) after 5 days of culture. (n = 3) (d–f) Human CB CD34⁺ cells were transduced with vector, ASXL1-MT (MT), or ASXL1-MT-K351R (KR) (coexpressing GFP) together with vector or BAP1 (coexpressing puromycin-resistant gene). After the puromycin selection for 48 h, cells were cultured in myeloid skewing (d), erythroid-skewing (e), and stem cell-maintaining (f) mediums, respectively. Myeloid maturation of CB cells was assessed by the ratio of CD33⁺CD66b⁺ cells after 7 days of culture (d). Erythroid maturation of CB cells was assessed by the ratio of CD79⁺CD235a⁺ cells after 5 days of culture (e). HSC maintenance of CB cells was assessed by the ratio of CD34-high cells after 7 days of culture (f). (n = 4 each) Data are shown as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons test

Fig. 5

ASXL1-MT/BAP1 complex increases colony-replating activity of haematopoietic progenitors and promotes RUNX1-ETO-induced leukaemogenesis. a Bone marrow c-Kit⁺ cells were transduced with vector/ASXL1-MT (MT)/ASXL1-MT-K351R (KR) (coexpressing blasticidin resistance gene) together with vector/BAP1 (coexpressing puromycin resistance gene). After the drug selection for 3 days, the cells were serially replated. Shown are weekly colony counts per 10⁴ replated cells (mean ± s.e.m.) from duplicate plates (left, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons test), and representative photos of each colony at second round (right, Scale bars: 200 μm). b Fetal liver c-Kit⁺ cells were transduced with RUNX1-ETO9a (RE9a) in combination with vector/MT/KR and were transplanted into recipient mice. Shown are the survival curves (vector: n = 7, MT: n = 8, KR: n = 9). *P < 0.05, **P < 0.01, log-rank test. c Fetal liver c-Kit⁺ cells were transduced with RE9a, vector/MT/KR, and vector/BAP1, and were transplanted into recipient mice. Shown are the survival curves (n = 10 each). **P < 0.01, log-rank test. d RUNX1-ETO transduced human CB CD34⁺ cells were transduced with vector/MT/KR (coexpressing GFP). Shown are the changes of GFP⁺ (vector/MT/KR transduced) cell frequency. e, f Cell cycle status and apoptosis of the cells described in (d) on day 11. e The frequency of S/G2/M phase in GFP⁺ cells was normalized to that in GFP^- cells (n = 6). f Shown are the frequency of Annexin V⁺ cells (n = 4). Data are shown as the mean ± s.e.m. **P < 0.01, ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons test. g, h RUNX1-ETO-expressing CB cells were transduced with vector/MT/KR (coexpressing GFP) in combination with vector/BAP1 (coexpressing NGFR) (g), with vector/MT (coexpressing GFP) in combination with vector/BAP/BAP1-C91S (CS) (coexpressing NGFR) (h). Shown are the changes of GFP/NGFR double-positive (DP) cell frequency. i Cell lysates extracted from RUNX1-ETO-expressing CB cells transduced with vector/MT/KR in combination with vector/BAP1 were subjected to immunoblotting

Fig. 6

ASXL1-MT/BAP1 complex induces upregulation of posterior HOXA genes and IRF8 through removal of H2AK119ub. a Murine bone marrow c-Kit⁺ cells were transduced with ASXL1-MT (MT) or ASXL1-MT-K351R (KR) (coexpressing blasticidin-resistant gene) together with vector or BAP1 (coexpressing puromycin-resistant gene), and were cultured in M3234 containing 20 ng/ml SCF, 10 ng/ml IL-3, and 10 ng/ml IL-6. After the selection with blasticidin and puromycin for 3 days, colony-forming cells were collected to extract RNA for RNA-seq analysis. A heatmap of the top 500 differentially expressed genes is shown. b Relative expression of Hoxa5, Hoxa7, Hoxa9, Irf8, Pten, and Sf3b1 in the indicated cells. Expression of each gene in cells expressing BAP1 alone was set as 1. Data are shown as mean ± s.e.m. of three biologically independent experiments. *P < 0.05, **P < 0.01, ****P < 0.0001, one-way ANOVA with Tukey’s multiple comparisons test. c Gene set enrichment analysis (GSEA) revealed that Hoxa9 target genes (upper) and genes related to monocyte/macrophage (lower) were highly expressed in cells coexpressing ASXL1-MT and BAP1 compared with those in cells expressing ASXL1-MT or BAP1 alone. d, e Murine bone marrow c-Kit⁺ cells were transduced with BAP1 together with vector or HA-tagged ASXL1-MT, and were cultured in semisolid medium. Genomic DNA fragments from these cells were immunoprecipitated with anti-HA (d) and anti-H2AK119ub (e) antibodies. Enrichments of H2AK119ub and ASXL1-MT (HA) at Hoxa5, Hoxa7, Hoxa9, Irf8, Pten, and Sf3b1 promotor loci were measured by qPCR. Data are shown as mean ± s.e.m. of three biological independent experiments. *P < 0.05, Student’s t-test. f RUNX1-ETO-expressing CB cells were transduced with vector/MT/KR (coexpressing GFP) in combination with vector/BAP1. GFP/NGFR double-positive cells were sorted 48 h after transduction. Relative mRNA levels of HOXA5, HOXA7, HOXA9, IRF8, PTEN, and SF3B1 were analyzed by qRT-PCR. Results were normalized to GAPDH, with the relative mRNA level in vector-transduced cells set at 1. Data are shown as mean ± s.e.m. of four biological independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, one-way ANOVA with Dunnett’s multiple comparisons test

Fig. 7

Depletion of BAP1 abrogates ASXL1-MT-induced differentiation block and inhibits leukaemogenesis. a 32Dcl3 cells were first transduced with Cas9, and were then transduced with a vector control or two independent sgRNAs targeting Bap1. Cell lysates were extracted from these cells and parent 32Dcl3 cells (pa), followed by immunoblotting analysis. b 32Dcl3 cells described in (a) were transduced with ASXL1-MT (coexpressing GFP), and were cultured with 50 ng/ml G-CSF. CD11b expression in untransduced (GFP⁻, red line) and transduced (GFP⁺, blue line) cells was assessed on day 6. c Morphology of the cells described in (b) was assessed by Wright–Giemsa staining. Original magnification, ×1000; Scale bars: 20 μm (left). Two independent experiments were performed, and proportions of segmented cells are shown as mean ± s.e.m. (right). *P < 0.05, one-way ANOVA with Dunnett’s multiple comparisons test. d Cell lysates extracted from 32Dcl3 cells described in (b) were subjected to immunoblotting. e Experimental scheme used in (f–h). Primary leukemia cells transformed by ASXL1-MT and SETBP1-D868N (cSAM cells) were transduced with Cas9 and Bap1-targeting sgRNAs, and were cultured in semisolid medium or transplanted into recipient mice. f Bap1-depleted cSAM cells produced significantly reduced numbers of colonies compared with control cSAM cells. Expression of human BAP1 reversed the reduced colony formation of Bap1-depleted cSAM cells. Colony numbers were counted on day 7 (n = 4). Data are shown as mean ± s.e.m. ****P < 0.0001, two-way ANOVA with Tukey’s multiple comparison test (left). Representative photos of colonies are also shown (right). Scale bars: 200 μm. g Survival curves of recipient mice transplanted with vector- or sgBAP1-transduced cSAM cells (n = 7 each). **P = 0.0031, ***P = 0.0008, log-rank test. h Relative expression of Hoxa5, Hoxa7, Hoxa9, Irf8, Pten, and Sf3b1 were assessed in vector- or sgBAP1-transduced cSAM cells using qRT-PCR. Results were normalized to Gapdh, with the relative mRNA level in vector-transduced cells set at 1. Data are shown as mean ± s.e.m. of four biological independent experiments. *P < 0.05, ^#P < 0.01, one-way ANOVA with Dunnett’s multiple comparisons test

Fig. 8

Hoxa genes and Irf8 contribute to leukaemogenesis and monopoiesis induced by ASXL1-MT and BAP1. a Cas9-expressing cSAM cells were transduced with vector or two independent sgRNAs targeting mouse Hoxa5, Hoxa7, Hoxa9, or Irf8. Colony-forming activity was assessed using these cells. Shown are the colony counts per 10⁴ plated cells from quadricate plates. Hoxa5, Hoxa7, or Hoxa9 depletion reduced the colony-forming activity of cSAM cells. One-way ANOVA with Dunnett’s multiple comparisons test. b Mouse bone marrow cells expressing RUNX1-ETO9a, ASXL1-MT, BAP1, and Cas9 were transduced with vector or two independent sgRNAs targeting mouse Hoxa5, Hoxa7, or Hoxa9. Colony-forming activity was assessed using these cells. Shown are the colony counts per 10⁴ plated cells from triplicate plates. Hoxa5, Hoxa7, or Hoxa9 depletion reduced the colony-forming activity of RUNX1-ETO9a, ASXL1-MT, and BAP1-expressing cells. One-way ANOVA with Dunnett’s multiple comparisons test. c Schematic presentation of experimental procedures for experiments shown in Fig. 8d. Cas9-expressing cSAM cells were transduced together with vector, HOXA7, or HOXA9 (coexpressing neomycin resistance gene). After G418 (1 mg/ml) selection for 7 days, cells were transduced with vector or two independent sgRNAs targeting mouse Bap1 (sgBap1 #1, #2). d Ectopic expression of HOXA7 or HOXA9 partially reversed the reduced colony-forming activity of Bap1-depleted cSAM cells. Colony numbers were counted on day 7 (n = 4). Two-way ANOVA with Tukey’s multiple comparison test. e Schematic presentation of experimental procedures for experiments shown in Fig. 8f. c-Kit⁺ bone marrow cells were purified from Rosa-LSL-Cas9 mice and were then retrovirally transduced with control vectors or ASXL1-MT (coexpressing blastcidin resistance gene) and BAP1 (coexpressing NGFR)-expressing vectors, and were lentivirally transduced with control vector or sgRNAs (coexpressing puromycin resistance gene) targeting mouse Irf8. After puromycin and blastcidin selection, cells were cultured with cytokines to induce differentiation toward both granulocytes and monocytes. f Granulocyte (Gr-1⁺CD11b⁺) and monocyte (Gr-1⁻CD11b⁺) differentiation was assessed by FACS analyses (left) and Wright–Giemsa staining (right, scale bars: 50 μm) after 5 days of culture (n = 3). One-way ANOVA with Tukey’s multiple comparisons test. Data are shown as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, n.s. not significant

Fig. 9

BAP1 depletion inhibits the growth of myeloid leukemia cells with ASXL1 mutations or MLL-fusions. a–c Murine bone marrow c-Kit⁺ cells were transformed by MLL-AF9. The MLL-AF9 leukemia cells were transduced with Cas9 and Bap1-targeting sgRNAs, and were cultured in semisolid medium or transplanted into recipient mice. a Vector or human BAP1-transduced (tdTomato⁺) cells were sorted, and were then transduced with vector or two independent sgRNAs targeting mouse Bap1 (sgBap1 #1, #2). Colony-forming activity was assessed using these cells. Shown are the colony counts per 10⁴ plated cells (mean ± s.e.m.) from quadricate plates. Expression of human BAP reversed the reduced colony formation of Bap1-depleted MLL-AF9 cells. ****P < 0.0001, two-way ANOVA with Tukey’s multiple comparison test. b Survival curves of recipient mice transplanted with vector- or sgBAP1-transduced MLL-AF9 cells (n = 7 for each group). **P = 0.0010, ^$$P = 0.0026, log-rank test. c Relative expression of Hoxa5, Hoxa7, Hoxa9, Irf8, Pten, and Sf3b1 were assessed in vector- or sgBAP1-transduced MLL-AF9 cells using the qPCR analyses. Results were normalized to Gapdh, with the relative mRNA level in vector-transduced cells set at 1. Data are shown as mean ± s.e.m. of four biological independent experiments. *P < 0.05, ^#P < 0.01, one-way ANOVA with Dunnett’s multiple comparisons test. d Human AML cell lines harboring an ASXL1 mutation (Kasumi-1 and MEG-01), a CML cell line with an ASXL1 mutation (TS9;22), AML cell lines with MLL-AF9 (MOLM-13 and THP-1), a T cell leukemia cell line (Jurkat) and a B cell lymphoma cell line (Raji) were transduced with Cas9 together with a vector or two independent sgRNAs targeting human BAP1 (sgBAP1 #3 and #4). Shown are cell numbers in each culture counted on days 2, 4, and 7. BAP1-targeting sgRNAs induced efficient depletion of BAP1 in all these cells. Two independent experiments were performed, and data are shown as mean ± s.e.m. n.s. not significant, **P < 0.01, ***P < 0.001; vector versus sgBAP1#3, ^$$$P < 0.001; vector versus sgBAP1#4, one-way ANOVA with Dunnett’s multiple comparisons test