Asxl1 C-terminal mutation perturbs neutrophil differentiation in zebrafish

ASXL1 is one of the most frequently mutated genes in malignant myeloid diseases. In patients with myeloid malignancies, ASXL1 mutations are usually heterozygous frameshift or nonsense mutations leading to C-terminal truncation. Current disease models have predominantly total loss of ASXL1 or overexpressed C-terminal truncations. These models cannot fully recapitulate leukemogenesis and disease progression. We generated an endogenous C-terminal-truncated Asxl1 mutant in zebrafish that mimics human myeloid malignancies. At the embryonic stage, neutrophil differentiation was explicitly blocked. At 6 months, mutants initially exhibited a myelodysplastic syndrome-like phenotype with neutrophilic dysplasia. At 1 year, about 13% of mutants further acquired the phenotype of monocytosis, which mimics chronic myelomonocytic leukemia, or increased progenitors, which mimics acute myeloid leukemia. These features are comparable to myeloid malignancy progression in humans. Furthermore, transcriptome analysis, inhibitor treatment, and rescue assays indicated that asxl1-induced neutrophilic dysplasia was associated with reduced expression of bmi1a, a subunit of polycomb repressive complex 1 and a reported myeloid leukemia-associated gene. Our model demonstrated that neutrophilic dysplasia caused by asxl1 mutation is a foundation for the progression of myeloid malignancies, and illustrated a possible effect of the Asxl1-Bmi1a axis on regulating neutrophil development.


Introduction
Myeloid malignancies are characterized by the proliferation and defective differentiation of myeloid progenitors [1,2]. The main types of these malignancies are myelodysplastic syndrome (MDS), myeloproliferative neoplasms (MPNs), and acute myeloid leukemia (AML). Chronic myelomonocytic leukemia (CMML) exhibits features of both MPN and MDS [3]. MDS, MPNs, and CMML have an inherent risk of transforming to AML. Knowledge about the leukemogenic mechanisms of myeloid malignancies can be a basis for developing therapeutic options.
Patients with multiple myeloid malignancies [4] including MDS, CMML, and AML frequently harbor somatic mutations in addition of sex combs-like 1 (ASXL1). In patients, ASXL1 mutations are usually heterozygous frameshift or nonsense mutations leading to C-terminal truncation [4] and loss of the ASXM2 domain and a PHD finger. Truncated ASXL1 mutations independently predict poor outcomes and shorter overall survival for patients with CMML [5,6].
To understand how ASXL1 participates in normal hematopoiesis and leukemogenesis, numerous models of aberrant ASXL1 expression have been described [7][8][9][10][11]. Mouse Asxl1 knockout (KO) models exhibit multilineage dysplasia [7]. Whereas loss of asxl1 in zebrafish is described as leading to the apoptosis of hematopoietic stem cells (HSCs), two 17-month adult asxl1 mutants developed an AML-like phenotype [12]. Overexpression of C-terminally truncated ASXL1 in mice inhibits myeloid differentiation and induces MDS-like diseases [8,10,11]. To obtain a model more closely resembling human disease, heterozygous Asxl1 G643fs mutant knock-in mice were generated [13,14]. These models have no obvious hematopoiesis disorders within 18 months, but show MDS/MPN-like disease by 18-24 months [13,14]. Even with these models and their different phenotypes, the role of ASXL1 in hematopoiesis is still largely unknown.
Mechanistically, the ASXL1 protein is thought to mediate the balance between polycomb and trithorax functions by regulating histone modifications such as monoubiquitination of histone H2A at lysine 119 (H2AK119ub) [15,16] and methylation of histone H3 at lysine 3 (H3K4me3) [10]. C-terminally truncated ASXL1 aberrantly enhances BAP1 function in deubiquitination of H2AK119ub [16][17][18], which is catalyzed by polycomb repressive complex 1 (PRC1) [19]. This hyperactive mutant ASXL1/BAP1 complex is thought to promote impairment of myeloid differentiation through inhibition of H2AK119ub at posterior HOXA genes and IRF8, encoding an essential transcription factor in myeloid lineages [18]. The ASXL1 G643fs mutation disrupts interactions with BMI1, a subunit of PRC1, and downregulated BMI1-driven H2AK119ub is observed in the derepressed p16Ink4a, leading in an MDS-like phenotype [14]. Hence, precisely how ASXL1 regulates H2AK119ub remains unclear. Furthermore, no consistent evidence indicates which factor is more important in mutant ASXL1-induced leukemogenesis. We need a model that closely resembles disease development in patients to recapitulate potential leukemogenesis and disease progression with ASXL1 C-terminal mutations.
An increasing number of leukemia models have been developed in zebrafish (Danio rerio) [20]. With properties such as external fertilization and embryonic development, zebrafish are an excellent model to study hematopoiesis, especially in early embryogenesis. To elucidate the function of Asxl1 in hematopoiesis and the leukemogenic mechanisms of Asxl1-truncated mutants, we established an endogenous truncated mutant in zebrafish that is comparable to human disease. Our C-terminally truncated zebrafish asxl1 mutant, unlike mouse models, has explicitly impaired neutrophil development. At 1 year, in addition to neutrophilic dysplasia, around 13% of the mutants had the phenotype of monocytosis, which resembles CMML, or increased myeloblasts, similar to AML. This leukemogenesis induced by the truncated mutant Asxl1 is transplantable. Furthermore, we found that expression of bmi1a, a subunit of PRC1, was decreased in asxl1 −/− fish and its reduction was associated with neutrophil dysplasia in our mutants. These results were demonstrated with transcriptome analysis, inhibitor treatment, and rescue assays. Our work showed that impaired neutrophil differentiation is fundamental for the progression of myeloid malignancies and illustrated the possible function of the Asxl1-Bmi1 axis in regulating neutrophil development.

Materials and methods
All information on oligos and target sequences is in Supplementary Table 1.

Zebrafish, embryo collection, and treatment
All experiments involving zebrafish were done in accordance with guidelines from the Institutional Animal Care and Use Committee of South China University of Technology. Wild-type AB strain zebrafish were raised and maintained under standard conditions [21]. Embryos were collected and staged according to Kimmel et al. [22].
Generation of asxl1-mutant lines and validation CRISPR/Cas9 was used to create asxl1 mutants. Cas9 mRNA was synthesized using mMESSAGE mMACHINE mRNA transcription-synthesis kits (Ambion). Cas9 mRNA (0.1 ng/embryo) and guide RNA (0.05 ng/embryo) were injected into single-cell wild-type embryos. Mutant embryos were identified through genotyping (see Supplementary Methods) and allowed to develop to F0. We identified F0 fish with inheritable mutations. F1 progeny from F0 outcrosses was identified by genotyping and sequencing. Identified F1 and progeny were used for experiments. asxl1 +/+ and asxl1 −/− embryos were generated and genotyped from heterozygous intercrosses.

RNA extraction and qRT-PCR
RNA was extracted using TRIzol (Invitrogen) and reverse transcribed by M-MLV Reverse Transcriptase (Promega). qPCR was with FastStart SYBR Green Master (Roche) and fold-changes determined by ΔΔCt method. All reactions were normalized against ef1α. Melting-curve analysis confirmed the presence of single PCR products. Significance was determined using Student's t tests with a significance threshold of p < 0.05.

RNA-sequencing (RNA-seq) and bioinformatics
At 3 days postfertilization (dpf), asxl1 +/+ and asxl1 −/− embryos from asxl1 +/− intercrosses were collected, and tails including caudal hematopoietic tissue were used for total RNA extraction. Each sample library was from around 20 tails. Sequencing libraries were prepared from poly-A selected RNA using NGS RNA Library Prep kits (Novogene). Sequencing was by an Illumina NovaSeq sequencing platform to obtain 2 × 150-bp pair-end reads. Sequencing quality was assessed using FastQC [23]. Reads were mapped onto the GRCz11 zebrafish reference genome using STAR (version 2.5.1b) with default setting [24]. Raw read counts for each Ensembl-annotated gene were calculated by Featurecounts [25]. EdgeR (version 3.26.5, false discovery rate (FDR) < 0.05; exact test) was used to call differentially expressed genes [26,27]. Gene ontology (GO) for biological processes was enriched using the Metascape online tool [28].

Results
Zebrafish asxl1 is expressed in hematopoietic cells, including myeloid cells The asxl1 and asxl2 genes are zebrafish homologs of the Drosophila Asx gene. Mammalian and zebrafish Asxl proteins are evolutionarily conserved. Phylogenetic analysis clustered zebrafish Asxl1 with mouse and human ASXL1, but separately from groups of ASXL2 or ASXL3 protein sequences ( Supplementary Fig. 1A). The amino acid sequences of zebrafish Asxl1 and Asxl2 share an overall 33 and 42% identity with human ASXL1 and ASXL2 orthologs. They share a conserved common domain architecture, including ASXN, ASXH, ASXM1, ASXM2, and PHD domains ( Supplementary Fig. 1B). These results indicated that asxl1/2 are authentic orthologs of mammalian ASXL1/2. The conserved structural domains suggest functional conservation of ASXL members among species during evolution.
Expression patterns of asxl1 during embryonic development were detected by WISH ( Supplementary Fig. 1C). In embryos, asxl1 was expressed as early as the 1-cell stage, indicating the presence of maternal transcript. At 5.3 hpf, asxl1 was ubiquitously expressed and at 18 hpf, asxl1 was restricted to the developing brain, eye, anterior lateral mesoderm, and posterior lateral mesoderm. At 24 and 36 hpf, asxl1 was specifically expressed in brain and eye. At 2 dpf, asxl1 was mostly in the nervous system (Supplementary Fig. 1C).
To investigate if asxl1 participates in hematopoiesis, we first examined if asxl1 expresses in hematopoietic cells. We collected adult kidney marrow cells from wild type before further sorting myeloid and lymphoid cells by FACS. qRT-PCR confirmed asxl1 expression in adult hematopoietic cells, with higher levels in myeloid cells (Supplementary Fig. 1D). Hence, asxl1 was expressed in hematopoietic cells, including myeloid cells. The expression of asxl1 in myeloid cells supported the hypothesis that asxl1 participates in zebrafish hematopoiesis, including myelopoiesis.

asxl1 C-terminal truncation generates new zebrafish mutants
To examine if asxl1 participates in hematopoiesis, we generated an asxl1-mutant line. CRISPR/Cas9 methods were used at exon 12 of the zebrafish asxl1 gene to produce lines with premature stop codons, similar to mutations in myeloid malignancies (Fig. 1A). We obtained two independent asxl1 alleles with exon 12 deletions of 7 bp in asxl1 e12 (−7) and 22 bp in asxl1 e12 (−22) . Both deletions predicted to lead to stop codons (in amino acids 919 and 914, respectively) before the ASXM2 and PHD domains (Fig. 1A). These two mutants have the same phenotypes so unless specified, asxl1 e12 (−22) data are presented. qRT-PCR for asxl1 expression in mutants showed decreased, which may be caused by nonsense-mediated decay (Fig. 1B). Most surviving asxl1 −/− fish displayed dwarfism and lower body weight (Fig. 1C). These results indicated that C-terminally truncated Asxl1 in zebrafish may lead to developmental abnormalities.

Zebrafish asxl1 mutants have impaired neutrophil differentiation as embryos
To investigate asxl1 in zebrafish hematopoiesis, we detected expression of lineage markers in mutants from intercrosses of asxl1 heterozygotes. First, the effects of asxl1 C-terminal truncated mutation on primitive hematopoiesis (<24 hpf) were examined. Erythropoiesis (gata1) and myeloid progenitor (pu.1) markers were indistinguishable between asxl1 mutants and siblings (Supplementary Fig. 2A). Additionally, a mature neutrophil marker (lyz) and intermediate neutrophilic progenitor marker for myeloperoxidase (mpx) were comparable between asxl1 mutant and siblings (Supplementary Fig. 2A). Finally, we examined expression of a gene associated with macrophages (mfap4) (Supplementary Fig. 2A) and found it was not impaired in asxl1 mutants. These data suggested that primitive hematopoiesis was not disrupted in asxl1 mutants.
We next characterized the biological function of asxl1 in zebrafish definitive hematopoiesis by examining the effects of asxl1 C-terminal truncated mutation on hematopoietic lineages. WISH and qRT-PCR at 3 dpf embryos ( Fig. 2A) showed significantly reduced lyz and mpx transcripts. These observations suggested that asxl1 may disrupt differentiation from myeloid progenitor to neutrophil. Supporting this, Sudan Black staining showed that SB + granule containing mature neutrophils [29] were reduced in asxl1 mutants ( Fig. 2A). Furthermore, FACS-sorted neutrophils from asxl1 +/+ and asxl1 −/− embryos were May-Grünwald-Giemsa stained to observe neutrophils morphology. Compared with wild type, the asxl1 mutant exhibited a lower proportion of mature neutrophils, whose signature is banded and segmented nuclei (Fig. 2B). Expression of HSC marker (c-myb) (Supplementary Fig. 2B) and a lymphocyte marker (rag1) has no obvious change between asxl1 mutants and siblings (Supplementary Fig. 2C). Expression of an erythrocyte (hbbe1) and a macrophage marker (mfap4) was not impaired in asxl1 mutants ( Supplementary Fig. 2C). Collectively, these results suggested that asxl1 C-terminal truncation may specifically impair neutrophil differentiation. To determine if asxl1 mutation caused the impaired neutrophil development, we performed rescue experiments. After injecting wild-type asxl1 mRNA into 1-cell stage embryos, lyz + cells increased in 3 dpf asxl1 mutants and siblings (Fig. 2C). This result suggested a crucial role for asxl1 in neutrophil differentiation.
We therefore tested if adult asxl1 zebrafish mutants developed severe myeloid malignancies. Kidney marrow from 55 1-year-old adult asxl1 −/− mutants was collected for blood cell count analysis ( Supplementary Fig. 3A-F and Supplementary Table 2). These samples showed a significantly decreased proportion of neutrophils, similar to the 6-month mutants ( Supplementary Fig. 3A). We observed a significantly increased proportion of monocytes/macrophages in asxl1 mutants (Supplementary Fig. 3B) and a small increase in the proportion of myeloid progenitors ( Supplementary  Fig. 3C). Eosinophils and lymphocytes had no obvious change in asxl1 mutants ( Supplementary Fig. 3D, E). The significant increase in the proportion of erythroblasts in 6month asxl1 mutants was no longer observed in 1-year stage mutants ( Supplementary Fig. 3F). This result may suggest the increased proportion of erythroblasts may be a temporary phenotype or found only in some samples. These data indicated that the asxl1-truncated mutation induced myeloid malignancies with neutrophilic dysplasia and monocytosis. Among the 55 1-year adult asxl1 mutants, 5 had a severe monocytosis phenotype (defined as >20% of kidney marrow cells as monocytes/macrophages) with neutrophilic dysplasia and without excessive blasts ( Supplementary  Figs. 3A-C and 4A). These phenotypes are similar to CMML patients, who have increased numbers of monocytes with at least one lineage dysplasia, as classified in the World Health Organization (WHO) guidelines [3]. We also found two asxl1 mutants with increased numbers of blasts. In these two mutants, the proportion of myeloid progenitors/ blasts was over 40% of whole kidney marrow (Supplementary Fig. 3C and Fig. 4A), similar to features associated with human AML, as defined by WHO [30]. These data suggested that some of the asxl1-truncated zebrafish mutants further developed CMML-like or AMLlike disease (Fig. 4A).
Histologic analysis of liver sections was performed on zebrafish with myeloid malignancies. CMML-like zebrafish showed perivascular macrophage infiltration (Fig. 4B-c, d). Liver parenchyma appeared abnormal in these mutants, with poorly demarcated cells (Fig. 4B). Cells with vacuolated cytoplasm were frequently observed in asxl1 −/− (Fig. 4B-e), similar to phenotypes as observed in by Gjini et al. [12]. Many gaps in the liver parenchyma were observed (Fig. 4B). Severe fibrosis was also observed in one of the CMML-like zebrafish (Fig. 4B-f). This phenomenon was similar to nonalcoholic fatty liver disease (NAFLD), where macrophages have a central role in progression [31]. The significantly increased proportion of monocytes/macrophages in our asxl1 mutant may be a contributor to NAFLD-like progression in our asxl1 mutants.

C-terminally truncated Asxl1 mutation-induced leukemia cells are transplantable
To confirm if C-terminally truncated asxl1 hematopoietic cells intrinsically induced leukemogenesis, we performed transplantation assays (Fig. 5A). asxl1 +/+ and asxl1 −/− kidney marrow cells were intracardially injected into sublethally irradiated 9-month wild-type recipient fish. At 18 days posttransplantation, 3 asxl1 +/+ and 5 asxl1 −/− recipients survived out of 19 and 35, respectively (Fig. 5A). The peripheral blood and kidney marrow cells from surviving recipients were collected for genotyping. Of five asxl1 −/− surviving recipients, four successfully repopulated ( Supplementary Fig. 4A, B). Kidney marrow from these four successfully repopulated recipients was collected for further blood cell count analysis (Fig. 5B, C). The asxl1 −/− surviving recipients showed a significant decreased in the proportion of neutrophils (Fig. 5C). The neutrophilic dysplasia we observed in asxl1 −/− recipients was similar to the phenotype of 6-month and 1-year asxl1 mutants ( Fig. 3a and Supplementary Fig. 3A). A severe monocytosis phenotype (about 64% of kidney marrow cells were monocytes/macrophages) was observed in one asxl1 −/− recipient, and the morphology of these increased macrophages was aberrant ( Fig. 5B-5). This phenotype is comparable to the CMML-like phenotype. These results demonstrated that our asxl1 mutated cells intrinsically induced leukemogenesis and were transplantable.

Transcriptome analysis shows disruption of neutrophil development
Our data showed that asxl1 mutations induced dysregulation of myeloid differentiation, mainly displayed as similarities to myeloid malignancies with neutrophilic dysplasia. To understand the basis for impaired neutrophil differentiation in zebrafish asxl1 mutants, we performed RNAseq on 3 dpf asxl1 mutants and its littermate controls. Our data showed 294 differentially expressed genes (FDR < 0.05, 146 downregulated genes, 148 upregulated genes, Supplementary Table 3, analysis detailed in "Methods"). Similar to the observed phenotypes, expression of all neutrophil markers (lyz, mpx, npsn, srgn) was downregulated ( Fig. 6A and Supplementary Fig. 5A). Expression of myeloid progenitor (pu.1 and cebpa) and macrophage (mfap4 and mpeg1) markers was not impaired ( Supplementary  Fig. 5B). GO analysis showed that wounding response-related genes were significantly enriched in asxl1 mutants, possibly because impaired neutrophils influenced the response to wounding (Fig. 6B). We also found that genes for an inflammatory cytokine (cxcl18b) and a matrix metalloproteinase (mmp13a) were upregulated in asxl1 mutants (Fig. 6A), which may be from induction by the deficiency of neutrophil granzymes [32,33]. Our transcriptome analysis mainly showed disruption of neutrophil development, also suggesting that neutrophil deficiency may stimulate expression of some inflammatory cytokines and enhance the inflammatory response.

Impaired neutrophil development from asxl1 mutations is rescued by bmi1a mRNA
To determine if impaired neutrophil differentiation was associated with decreased bmi1a/cbx4 expression, we overexpressed bmi1a and cbx4 in asxl1 mutants by injecting bmi1a/cbx4-3xFLAG-P2A-DsRed mRNA ( Supplementary  Fig. 7A). Fluorescence indicated bmi1a/cbx4 expression: DsRed was observed in embryos injected with bmi1a/cbx4-3xFLAG-P2A-DsRed mRNA compared to controls (Supplementary Fig. 7B). Expression of both lyz and SB + neutrophils was significantly increased in asxl1 mutants after injecting bmi1a mRNA ( Fig. 7B and Supplementary  Fig. 7C). Injection with cbx4 mRNA led to almost complete embryonic lethality at 3 dpf and embryos exhibited an expanded heart field ( Supplementary Fig. 7B). Our data showed that expression of bmi1a can rescue the asxl1 mutation-induced phenotype and indicated that decreased bmi1a expression may be a mechanism by which neutrophils were disrupted in asxl1 −/− embryos.
data showed that an endogenous asxl1-truncated mutant mainly disrupted H2AK119ub and this disruption may be associated with reduced expression of genes encoding PRC1 members, bmi1a and cbx4.

Discussion
How might ASXL1 facilitate myeloid malignancies?
Patients with multiple myeloid malignancies frequently harbor an ASXL1 mutation. Existing disease models mainly have total loss of ASXL1 [7,9] or overexpression of a C-terminal truncation [8,10,11]. Loss of ASXL1 in mice is described to exhibit multilineage dysplasia [7,9]. A zebrafish KO model is described to have reduced HSCs at 3 dpf [12]. In addition, C-terminally truncated ASXL1 can be detected in leukemia cells with an ASXL1 mutation [42] and expression of ASXL1 with a C-terminal truncation may confer a gain-of-function by promoting BAP1 activity [18]. These results raised the possibility that a ASXL1 mutation might confer a gain-of-function effect. Therefore, mouse models overexpressing C-terminally truncated ASXL1 were generated and showed dysregulation of multilineage differentiation [10,11]. However, an overexpressing model may not fully recapitulate ASXL1 function because of protein overdose. Therefore, endogenous heterozygous Asxl1 G643fs mutant knock-in mice were generated and had no obvious hematopoiesis disorders within 18 months, but showed MDS/MPN-like disease by 18-24 months [13,14]. In contrast to mouse models, zebrafish have external fertilization and embryonic development, allowing us to observe phenotypes starting from the embryonic stage. Our zebrafish with endogenous truncated asxl1 mutants captured the most frequent C-terminal frameshift mutations in patients. This model had specifically impaired neutrophil differentiation without effects on other lineages including erythrocytes, macrophages, and lymphocytes. Furthermore, transcriptome analysis showed that zebrafish asxl1 mutants exhibited predominantly dysregulation of neutrophilspecific genes. Our work complements a recent study employing single-cell RNA-seq that suggests the involvement of ASXL1 in neutrophil differentiation [44].
studies showing that Asxl1 mutation alone is not sufficient for leukemogenesis [13]. Moreover, we found that neutrophil disruption happened as early as the embryonic stage and persisted throughout the lifetime. At 6 months, mutants with Asxl1 truncations in our study mainly displayed MDS with neutrophilic dysplasia, which presented as refractory neutropenia. At 1 year, in addition to neutrophilic dysplasia, about 13% of mutants acquired other phenotypes of monocytosis that mimic CMML or increased myeloblasts similar to AML (Fig. 4). This result suggests a fundamental role for neutrophil disruption in myeloid malignancies progression. However, further progression to CMML-like or AMLlike phenotypes was observed only in some mutants, suggesting that asxl1 mutations alone are sufficient for MDS-SLD but not for severe myeloid malignancies, with other events responsible for progression. Clonal evolution and the alteration of the hematopoietic niche might contribute to disease progression [1]. In addition, inflammation was recently reported to contribute to clonal hematopoiesis of indeterminate potential [46]. Chronic inflammation may lead to HSC exhaustion and aging, which may promote mutation [47][48][49][50]. Our transcriptome analysis showed a possible inflammatory response in our mutants. Mutant asxl1-induced neutrophilic dysplasia may induce persistent inflammation and increase the risk of myeloid malignancy progression. Further studies need to be done to investigate the mechanism of progression, and our model could provide an opportunity to investigate additional contributors to myeloid malignancy progression.

How might bmi1a be an intermediate in neutrophil disruption?
Our data indicated that disruption of neutrophil development has a fundamental role in myeloid malignancy progression in asxl1 mutants. We explored the potential mechanism. Previous studies show that an ASXL1 G643fs mutant interrupts its interaction with BMI1, leading to an MDS-like phenotype [14]. However, in our model, we observed reduced expression of bmi1a. Hence, ASXL1 mutants may not only block interaction with BMI1 but also disrupt its expression. Absence of Bmi1 affects the proliferative activity of GMP, and Bmi1 −/− GMP has dispersed macrophage colonies [38]. This result indicates that Bmi1 may participate in myeloid differentiation. Using a bmi1a inhibitor, we mimicked a neutrophil dysplasia phenotype (Fig. 7A). Supplementing with bmi1a mRNA rescued asxl1-induced neutrophil dysregulation (Fig. 7B). Hence, our work illustrated an alternative mechanism of Asxl1-Bmi1a axis on regulating neutrophil development.
The reduction in H2AK119ub is important in mutant ASXL1-induced leukemogenesis. Recent studies show that overexpression of C-terminally truncated ASXL1 (ASXL1-MT) and BAP1 leads to a hyperactive complex that subsequently upregulates IRF8, which is an essential transcription factor for myeloid lineage development [51][52][53], through H2AK119 deubiquitination. In addition, disrupted interaction between ASXL1 G643fs and BMI1 perturbs BMI1driven H2AK119ub, further promoting an MDS-like phenotype. The reduced expression of PRC1 members, including BMI1 and CBX4, further contributes to the mutant asxl1-induced dysregulation of H2AK119ub. However, we are not suggesting that reduction of H2AK119ub blocks neutrophil differentiation. This H2AK119ub change could be a consequence of the reduced expression of bmi1a and cbx4. In any case, our data show a novel way for ASXL1 mutations to reduce H2AK119ub and raise the possibility that the Asxl1-Bmi1a axis is involved in neutrophil differentiation.
Taken together, results from this study show that endogenous truncated asxl1 mutations result in impaired neutrophil differentiation. Our model revealed that neutrophilic dysplasia caused by asxl1 mutations is a foundation for the progression of myeloid malignancies, and bmi1a dysregulation is associated with neutrophil dysplasia (Fig. 8).
Zebrafish are an ideal model for large-scale chemical screening. Our endogenous truncated asxl1 zebrafish model In our model, C-terminally truncated with asxl1 mutations specifically blocked neutrophil differentiation. The mutants initially exhibited an MDS-like phenotype with neutrophilic dysplasia, and some mutants progressed to more severe phenotypes, similar to CMML/AML disease. Dysregulation of bmi1a may be a reason for leukemogenesis in asxl1 mutants. Asxl1 wild-type Asxl1 protein, tAsxl1 truncated Asxl1 protein.
provides a new opportunity for a high-throughput drugscreening platform using whole organisms.

Data availability
RNA-seq data are available at GEO under accession number GSE142214. For original data, please contact Xiao Fang: fangxiao@scut.edu.cn.
Author contributions XF, WZ, and KY designed the research; XF generated asxl1 mutants and identified their phenotype; XF and SX performed cell sorting, cell staining, and transplantation assay; XF performed RNA-seq library preparation and analysis; XF generated constructs and performed rescue assays (asxl1 and bmi1a) and drug treatment. XF analyzed results, made figures, and wrote the paper. WZ, KY, YZ, JX, ZH, and WL discussed the results and commented on the paper.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.