Research Article: Myelopoiesis and Myeloid Leukaemogenesis in the Zebrafish

Date Published: July 19, 2012

Publisher: Hindawi Publishing Corporation

Author(s): A. Michael Forrester, Jason N. Berman, Elspeth M. Payne.


Over the past ten years, studies using the zebrafish model have contributed to our understanding of vertebrate haematopoiesis, myelopoiesis, and myeloid leukaemogenesis. Novel insights into the conservation of haematopoietic lineages and improvements in our capacity to identify, isolate, and culture such haematopoietic cells continue to enhance our ability to use this simple organism to address disease biology. Coupled with the strengths of the zebrafish embryo to dissect developmental myelopoiesis and the continually expanding repertoire of models of myeloid malignancies, this versatile organism has established its niche as a valuable tool to address key questions in the field of myelopoiesis and myeloid leukaemogenesis. In this paper, we address the recent advances and future directions in the field of myelopoiesis and leukaemogenesis using the zebrafish system.

Partial Text

The zebrafish is emerging as a powerful model system in which to study haematopoiesis and leukaemogenesis. In addition to the benefits afforded by scale and simplicity of this versatile genetic model system for studying developmental aspects of haematopoiesis, the last decade has seen an explosion of molecular methods and models to facilitate studies informing on haematopoietic disease biology, particularly leukaemogenesis and cancer. At its inception as a cancer model, proliferation and angiogenesis were proposed as phenotypic attributes as readouts relevant to cancer pathogenesis [1]. However, it was the generation of a transgenic zebrafish expressing the C-myc oncogene under the control of the rag2 promoter that went on to develop T-cell acute lymphoblastic leukaemia (ALL), which really revolutionized the view of the scientific world on this small organism as a cancer disease model [2]. In the ensuing 10 years, many models of oncogene induced cancer have been generated in zebrafish along with mutagenesis strategies to identify novel tumour suppressor genes or chromosome instability loci [3–5]. The utility of such models to answer key biological questions continues to grow. In this paper, we focus on developments in the field of myelopoiesis in the zebrafish, cancer models affecting the myeloid lineages, and how these have instructed our knowledge on the biology of these diseases.

Zebrafish haematopoiesis occurs in two waves in the developing embryo, termed primitive and definitive [6]. In contrast to human and murine haematopoiesis (where primitive haematopoiesis initiates with the development of primitive erythroid cells in the blood islands of the yolk sac), in zebrafish, primitive wave erythroid cells develop from caudal lateral plate mesoderm in bilateral stripes that migrate towards the midline forming a structure termed the intermediate cell mass (ICM). A population of primitive macrophages also emerges from a distinct anatomical location in the anterior lateral plate mesoderm (ALPM) between 12 and 24 hours after fertilization (hpf) [7, 8]. Definitive haematopoiesis initiates at around 24 hpf in the posterior blood island (PBI), with the emergence of bipotent erythromyeloid progenitors (EMPs). These cells are marked in their undifferentiated state by combined expression of gata1 and lmo2 or by expression of cd41 [6]. These cells have both proliferative and differentiation potential and increase in number, peaking at 30–36 hpf. This wave of haematopoiesis gives rise to further erythrocytes and myeloid cells and recently has been shown to give rise to early mast cells in developing embryos [9]. Multipotent definitive haematopoietic stem cells (HSCs) expressing cd41, c-myb, and runx1 arise directly from kdrl-expressing haemogenic endothelium in the ventral wall of the aorta starting around 26–28 hpf [10, 11]. These cells then migrate to the caudal haematopoietic tissue (CHT) where they seed and divide giving rise to all lineages of adult blood cells. These cells go on to populate the adult organs of haematopoiesis in the zebrafish, the kidney and the thymus. The precise timing of the move from primitive wave haematopoiesis to definitive wave haematopoiesis has yet to be fully established, but evidence from globin gene expression and mutants with normal primitive wave blood production suggests that the major contribution of haematopoiesis comes from definitive HSC derived cells by around 5 days post fertilization [12–14].

Cross-reactive antibodies to zebrafish proteins are lacking, arguably more so in the haematopoietic system than in others. This limitation means that the detailed lineage and differentiation status analysis of haematopoiesis, so elegantly understood in the murine system, is currently challenging to undertake in the zebrafish. Thus a major endeavour in recent years has been the generation of new tools for such analysis in the haematopoietic system. Along with the development of these tools has also come a much broader understanding of myeloid lineage development in zebrafish. The first transgenics developed to mark myeloid cells expressed enhanced green fluorescent protein (eGFP) from the major myeloid transcription factor pu.1. Tg(spi1/pu.1:eGFP) animals express eGFP in primitive wave myeloid cells but by 2 days postfertilization (dpf), expression of eGFP in myeloid cells is markedly reduced as pu.1 expression is downregulated [15, 16]. To visualize neutrophil granulocytes later in development, several transgenic lines have been generated by various laboratories. These include the Tg(lysc:dsRed) and Tg(lysc:eGFP) lines [17] as well as Tg(mpx:eGFP) [18, 19] and Tg(myd88:eGFP) [20]. While all of these lines label predominantly neutrophil granulocytes, it is notable that the overlap in expression of the endogenous transcripts (by in situ hybridization) or protein (by antibody) as well as the reporter gene expression between transgenic lines is not fully concordant, suggesting that subtly different populations are labelled by each transgene depending on the developmental time point of evaluation [17, 21]. Some of these subtleties in gene and protein expression have been addressed. L-plastin specifically has in some early studies been suggested to mark monocyte/macrophage lineage cells but there is a clear evidence that this protein is expressed (as in mammals) in all leucocytes [21]. The Tg(lysc:eGFP) expresses GFP from 22 hpf, initially in primitive macrophages arising from the ALPM. Expression of eGFP increases and is notable in the CHT (likely labelling and differentiating definitive myeloid cells) and the developing brain and retina (more likely to represent the on-going expression in a proportion of macrophages). To clarify precisely which cells express the eGFP from the Tg(lysc:eGFP) transgene, Hall et al. performed anti-GFP staining along with fluorescent in situ hybridization for mpx, l-plastin, and fms. Dual staining was observed for eGFP with each of these myeloid transcripts; however, there were some eGFP (lysc) expressing cells that did not express mpx, some fms expressing cells that did not express eGFP (lysc), and some l-plastin expressing cells that did not express eGFP. Thus, the Tg(lysc:eGFP) marks primitive macrophages and a majority of developing granulocytes but does not label all mpx positive granulocytes or all fms expressing macrophages [17]. It is conceivable that these subtleties may in time come to give us more detailed information about subpopulation of myeloid cells, such as their stage of differentiation. More recently transgenic lines using the mpeg1 or fms (csf1r) promoter [22, 23] have been used to distinguish macrophage populations from granulocytic myeloid cells, further enhancing studies of innate immune system. However, fms reporter animals exhibit expression in neural crest-derived xanthophores as well as macrophages, which may result in some limitations in the use of this system. By contrast, the mpeg1 promoter appears exclusive to macrophages, but expression in adult fish is maintained only in zebrafish lines generated using direct transgenic approaches, and not detectable in those lines in which mpeg1 is linked to a GAL4/UAS expression system. To further delineate the expression pattern of macrophages and other mononuclear phagocytes in adult zebrafish, a promoter fragment of the MHC class II beta gene, mhc2dab, was isolated. By virtue of the combined transgene expression, the Tg(mhc2dab:eGFP) transgenic line in combination with Tg(CD45:dsRed) (which labels all leukocytes except B cells) has now allowed identification of macrophages and dendritic cells as well as B lymphocytes in adult zebrafish tissues [24].

Many aspects of myelopoiesis have been interrogated using the zebrafish embryo. Foremost, forward genetic screens have been employed to identify novel genes required for primitive or definitive myelopoiesis. The critical role of transcription factors and developmental microenvironment in determining haematopoietic lineage fate choice has also been elegantly addressed using this model, using reverse genetics and transplantation techniques. More recently transient heterologous overexpression of mutated human oncogenes has provided some mechanistic insight into the potential pathogenetic effects of such genes on normal developmental haematopoiesis and malignant transformation. In addition functional studies have also addressed aspects of the innate immune system using the zebrafish (also reviewed elsewhere in this issue of AIH). What follows is a summary of a selection of studies in zebrafish that highlight its diverse and unique capacity to answer a range of biological questions pertaining to myelopoiesis.

Aged wild-type zebrafish (24+ months) are susceptible to the development of a spectrum of neoplasms with an incidence rate around 11% [3], however the incidence of haematopoietic malignancies is rare. Studies of transgenic zebrafish, with tissue specific or ubiquitous promoters driving human or murine oncogenes, have however resulted in faithful models of myeloid leukaemias with features of their human disease counterparts. Below is a summary of the existing models of myeloid leukaemia, the novel findings such models have contributed to our understanding of human myeloid malignancies and a critique of existing and emerging technologies within this field.

Overall, compared to the lymphoid tumours, models of myeloid leukaemia are relatively less penetrant with leukaemia rates ranging from 25% [51] to <1% [52]. The generation of novel promoters may facilitate more faithful models of human myeloid disease in zebrafish. In particular, dissection of the zebrafish runx1 promoters has provided new insights into the regulation of this gene in zebrafish but may also prove to be a better driver of oncogene-induced malignant myeloid disease [87]. One potential complimentary strategy is the recent interest in developing methodologies for xenotransplantation of human or mouse cancer cells into zebrafish and applying this approach to myeloid disease [88]. Tissue culture assays and animal models have been instrumental in determining key molecular pathway in cancer and novel drug development. However, in vitro assays lack the critical context of the tumour microenvironment, while mouse xenografts are cost-prohibitive and require extensive engraftment time. By contrast, the use of zebrafish facilitates scalability, where large numbers of rapidly developing, externally fertilized transparent embryos can be used to screen compounds in a high-throughput manner. By using embryos at 48 hpf, xenograft rejection is minimized, by virtue of their lack of an adaptive immune system during the first week of life [89]. The zebrafish embryo has contributed significantly to our understanding of the developmental biology of haematopoiesis and myelopoiesis over the past decade. The exponential rise in our ability to dissect the biology of myeloid cells in this small vertebrate will no doubt fuel further insights and broaden the scope for current models of myeloid leukaemias. The advent of TALENs and zinc finger nucleases as well as the zebrafish mutation project at the Sanger Centre ( promises to deliver us knockouts for all genes in the zebrafish genome that will greatly enhance future studies, particularly of tumour suppressor genes in myeloid disease.   Source:


Leave a Reply

Your email address will not be published.