The granulocyte colony-stimulating factor receptor (G-CSFR) is a key regulator of neutrophil production, or granulopoiesis, by impacting on the proliferation and differentiation of precursors as well as enhancing neutrophil survival (1, 2). The G-CSFR is activated by the cytokine G-CSF, which is produced developmentally (3), but also in response to injury and infection to stimulate so-called ‘emergency’ granulopoiesis (2). This cytokine is also used clinically in situations where neutrophil numbers are low (4).

Acquired somatic mutations in CSF3R, the gene encoding G-CSFR, have been found in a variety of leukemias (5). These include nonsense and frameshift mutations within exon 17 that serve to truncate the G-CSFR intracellular domain that are observed in acute myeloid leukemia (AML), particularly subsequent to severe chronic neutropenia (SCN) (6, 7), but also de novo and relapsed forms of the disease (8–10). Similar mutations have also been identified in chronic neutrophilic leukemia (CNL), atypical chronic myelogenous leukemia (aCML) and chronic myelomonocytic leukemia (CMML) patients (9, 11–14), although these are usually co-incident with alternative constitutively-activating G-CSFR mutations (15). The locations of the receptor truncations are quite variable, being found from position 738 through to 819 (5, 7–14). They lead to hyperresponsiveness to G-CSF, enhanced signaling particular of STAT5, and increased proliferation (16–18), with the ability to mediate leukemic transformation in vitro (10). Various mouse models of G-CSFR truncation mutations have been generated by different gene targeting approaches (19–21). While all exhibited enhanced responsiveness to G-CSF and altered receptor internalization, they were variable with respect to their effects on steady-state neutrophil levels.

Zebrafish has proven to be a robust experimental platform for the study of blood and immune cell development and its perturbation in disease (22). Zebrafish granulopoiesis occurs in multiple waves like mammals (23). This includes a primitive wave generating myeloid precursors in the rostral blood island that differentiate into neutrophils as they migrate across the yolk sac and a definitive wave that ultimately produces neutrophils from the kidney marrow, the equivalent of mammalian bone marrow (24). Zebrafish has been shown to have a structurally and functionally conserved csf3r gene (3), which contributes to neutrophil production throughout the life-course (25, 26). The zebrafish csf3r gene was targeted with CRISPR-Cas9 to generate truncating hyperresponsive G-CSFR mutations based on those observed in leukemia. These csf3r mutant fish possessed enhanced numbers of neutrophils during primitive and definitive hematopoiesis, but in adulthood neutrophil numbers were normal but maturation was reduced. Together these data identify on-going impacts of G-CSFR truncations on neutrophil production, and describe a useful in vivo model for further studies.

The G-CSFR has a pivotal function in the regulation of neutrophil production and function, as well as hematopoietic stem cell mobilization, particularly in ‘emergency’ situations (1, 2). Mutations that render the G-CSFR non-functional have been shown to result in profound neutropenia (45–47), which has been verified in mice (19, 48). In contrast, mutations that enhance G-CSFR signaling are associated with leukemias and other myeloproliferative disorders (11, 12). These include acquired truncation mutations first identified in SCN patients with a strong predisposition to the development of AML (5, 49), but also described in de novo and recurrent AML, as well as CNL, aCML and CMML patients (8–14). These mutations serve to truncate the G-CSFR intracellular domain, resulting in hyperresponsiveness to G-CSF (5, 18, 49). Previous studies have shown zebrafish possess a conserved G-CSFR, including key residues critical for intracellular signaling (3), with inactivating mutations of zebrafish G-CSFR able to reproduce the sustained neutrophil deficiency observed in SCN (25, 26). This study aimed to create truncating G-CSFR mutations in zebrafish using CRISPR-Cas9-mediated genome editing and to characterize their impact on primitive, definitive and emergency hematopoiesis.

Three mouse models of truncated ‘hyperresponsive’ G-CSFR mutants have been generated. One in which the endogenous mouse gene was mutated (19) and another in which a truncated human G-CSFR was expressed transgenically (21) showed decreased peripheral neutrophils, although the bone marrow contained normal neutrophil numbers and an elevated number of immature myeloid cells. In contrast, an alternate mouse model with targeted mutation of the endogenous gene showed normal circulating neutrophil numbers, but elevation of band and segmented neutrophils as well as progenitors in the bone marrow (20). The results in zebrafish adults are consistent with this last mouse model, with the number of peripheral neutrophils not significantly different between genotypes, but those in the kidney marrow exhibiting reduced maturation. This zebrafish study has allowed for the first time to assess the effect of hyperresponsive truncating G-CSFR mutations on primitive and early definitive hematopoiesis. Strikingly, during primitive hematopoiesis, neutrophil numbers were elevated. While this is clearly different to what is observed in adult mice and zebrafish, it should be noted that G-CSF/G-CSFR signaling in zebrafish has been shown to be particularly important for primitive neutrophil production, with ablation of either leading to an almost complete loss of this population (3, 26). This developmental stage is therefore perhaps more comparable to the situation when adult mice are injected with G-CSF, when significant increases in neutrophils are observed. In contrast, during early definitive zebrafish hematopoiesis the effects of G-CSFR truncation were less marked, consistent with previous studies showing the G-CSFR ablation has a more modest impact on zebrafish neutrophils during this phase (3, 26). All mouse models noted hyperresponsive to exogenous G-CSF resulting in increased production of cells along the neutrophil lineage (19–21). This collectively illustrates the complex regulation of neutrophil production mediated by G-CSFR.

Truncated G-CSFR mutants have been described extensively as exerting a ‘dominant-negative’ effect on wild-type receptors (5). However, the data presented here indicated that the truncating csf3r allele was not strictly dominant over the wild-type allele. Notably, heterozygous mutants were similar to homozygous mutants with respect to primitive hematopoiesis, but more like wild-type fish with respect to early definitive hematopoiesis and variable in adults. Moreover, close examination reveals that this is also the case in other studies. For example, while heterozygous and homozygous mutant mice showed equivalent initial responses to G-CSF injection, these were more sustained in homozygote animals (50). In other studies, cells expressing both normal and truncated G-CSFR exhibited intermediate phenotypes or responses, including with respect to sustained signaling and impaired internalization (17, 50). Thus, truncating G-CSFR mutations should be more accurately described as ‘co-dominant’. Indeed, there are good biochemical rationale as to why this may be. Firstly, since functional receptor complexes consist of multimers, a relevant cell in a heterozygous individual will express a mixture of receptor forms on their surface, a proportion of which will contain only wild-type receptors able to signal normally. Secondly, those receptor complexes that contain at least one wild-type receptor chain will still interact with the normal regulatory machinery. Thirdly, the enhanced signaling in heterozygote cells is likely associated with increased expression of SOCS3, a key negative regulator of the G-CSFR (51), enhancing its ability to act on those receptor complexes carrying a wild-type G-CSFR chain, since it retains the requisite SOCS3 binding site (52).

Zebrafish expressing truncated G-CSFRs did not exhibit overt leukemia, which is also true of mouse models (19–21), confirming that co-operating genes are absolutely required. In patients who have acquired AML subsequent to SCN, a number of common genetic lesions have been identified, including partial or complete loss of chromosome 7 and activating RAS mutations (53, 54). Inactivating mutations in RUNX1 (55, 56) and CEBPA (57) as well as the PML-RARa oncoprotein fusion (58) have been reported to cooperate with truncated hyperresponsive G-CSFR mutations. In addition, STAT5, which shows particularly strong sustained activation by truncated G-CSFRs (17), has been implicated in mediating the effects of truncated G-CSFRs (59). The zebrafish model that was generated in this study will be highly valuable in analyzing additional co-operating genes.

A range of other G-CSFR mutations have been identified that are associated a variety of leukemias and myeloproliferative disorders (5). These include activating mutations such as T618I in CNL and related disorders (11, 55, 60), and a E785K polymorphism associated with MDS (61). These exert distinctive pathological effects, activate different (and sometimes overlapping) pathways (11, 60), and have been demonstrated to respond differentially to pharmacological agents. For example, the JAK2 inhibitor ruxolitinib has been shown to decrease proliferative signaling and neutrophil count in patients with G-CSFR-T6181 mutations whereas the SRC inhibitor dasatinib (but not ruxolitinib) was able to selectively inhibit proliferation of cells expressing truncated G-CSFRs (11). Analyzing the array of pathogenic G-CSFR mutants in zebrafish provides an opportunity to explore the specificity of their action, downstream pathways and therapeutic sensitivity in the context of a whole animal, which is an exciting prospect. Finally, the truncating mutations often occur concurrently with these (and other) G-CSFR mutations (11, 55, 62), which could also be investigated in zebrafish.

This study has successfully created and characterized a zebrafish model of hyperresponsive G-CSFR truncations. Significantly enhanced neutrophil production was observed throughout successive waves of embryonic hematopoiesis, but a defect in neutrophil maturation was also evident in adult kidney marrow. The G-CSFR truncation also displayed partial dominance. Collectively, this has shed new light on the impact of such mutations across the lifespan and generated a bone fide zebrafish model to facilitate further clinically-relevant studies to explore the role of co-operating mutations and genes, as well as downstream effector pathways that may inform new approaches to therapy, which could be readily trialed in this model.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The animal study was reviewed and approved by Deakin University Animal Ethics Committee.

AW conceived the project. FB and VB generated the zebrafish G-CSFR mutants and VB and MG performed most of the in vivo and in vitro characterization. AW and CL co-supervised VB, MG and FB and contributed to on-going experimental design. All authors contributed to the article and approved the submitted version.