A novel zebrafish jak2aV581F model shared features of human ...

Experimental Hematology 2009;37:1379–1386

A novel zebrafish jak2aV581F model shared features of human JAK2V617F polycythemia vera Alvin C.H. Maa, August Fana, Alister C. Wardb, Clifford Liongueb, Rowena S. Lewisb, Suk H. Chengc, P.K. Chanc, Sze-Fai Yipd, Raymond Lianga, and Anskar Y.H. Leunga c

a Department of Medicine, The University of Hong Kong, Hong Kong; bSchool of Medicine, Deakin University, Waurn Ponds, Victoria, Australia; Department of Biology and Chemistry, City University of Hong Kong, Hong Kong; dDepartment of Pathology, The University of Hong Kong, Hong Kong

(Received 27 June 2009; revised 28 August 2009; accepted 31 August 2009)

Objective. The Janus kinase 2 (JAK2) is important for embryonic primitive hematopoiesis. A gain-of-function JAK2 (JAK2V617F) mutation in human is pathogenetically linked to polycythemia vera (PV). In this study, we generated a zebrafish ortholog of human JAK2V617F (referred herewith jak2aV581F) by site-directed mutagenesis and examined its relevance as a model of human PV. Materials and Methods. Zebrafish embryos at one-cell stage were injected with jak2aV581F mRNA (200pg/embryo). In some experiments, the embryos were treated with a specific JAK2 inhibitor, TG101209. The effects of jak2a stimulation on hematopoiesis, jak/stat signaling, and erythropoietin signaling were evaluated at 18-somites. Results. Injection with jak2aV581F mRNA significantly increased erythropoiesis, as enumerated by flow cytometry based on gfp+ population in dissociated Tg(gata1:gfp) embryos. The response was reduced by stat5.1 morpholino coinjection (control: 4.37% ± 0.08%; jak2aV581F injected: 5.71% ± 0.07%, coinjecting jak2aV581F mRNA and stat5.1 morpholino: 4.66% ± 0.13%; p ! 0.01). jak2aV581F mRNA also upregulated gata1 (1.83 ± 0.08 fold; p [ 0.005), embryonic a-hemoglobin (1.61 ± 0.12 fold; p [ 0.049), and b-hemoglobin gene expression (1.65 ± 0.13– fold; p [ 0.026) and increased stat5 phosphorylation. These responses were also ameliorated by stat5.1 morpholino coinjection or treatment with a specific JAK2 inhibitor, TG101209. jak2aV581F mRNA significantly reduced erythropoietin gene (0.24 ± 0.03 fold; p [ 0.006) and protein expression (control: 0.633 ± 0.11; jak2aV581F mRNA: 0.222 ± 0.07 mIU/mL; p [ 0.019). Conclusion. The zebrafish jak2aV581F model shared many features with human PV and might provide us with mechanistic insights of this disease. Ó 2009 ISEH - Society for Hematology and Stem Cells. Published by Elsevier Inc.

Embryonic hematopoiesis occurs in two successive waves, known as primitive and definitive hematopoiesis. In mammals, these two processes are initiated successively in the extraembryonic yolk sac and the aorto-gonado-mesonephro [1]. Zebrafish (Danio rerio) has emerged as a model for the study of embryonic hematopoiesis. In this organism, primitive hematopoiesis is evident at 10 hours postfertilization (hpf) (bud stage) by the embryonic expression of the stem cell leukemia (scl), lim-only domain protein 2 Offprint requests to: Anskar Y.H. Leung, M.D., Ph.D., Department of Medicine, Queen Mary Hospital, The University of Hong Kong, Pok Fu Lam Road, Room K418, K Block, Hong Kong; E-mail: [email protected] Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.exphem.2009.08.008.

(lmo2), and gata2 genes in the posterior lateral plate mesoderm [2,3]. This mainly generates cells of the erythroid lineage, which are evident as early as 12 hpf (6-somites), by expression of gata1 first in the posterior lateral plate mesoderm and thereafter in the intermediate cell mass at 18 hpf (18-somites) [2,3]. On the other hand, embryonic macrophage production is initiated within the anterior lateral plate mesoderm (or rostral blood island), characterized first by expression of spi1 at 12 hpf and later l-plastin at 18 hpf in this region [4–6]. Definitive hematopoiesis begins at 24 hpf along the ventral wall of dorsal aorta, the equivalent of mammalian aorto-gonado-mesonephro, and is characterized by expression of runx1 and c-myb [2,7,8]. Janus kinase (JAK)/signal transducer and activation of transcription (stat) pathway is an important

0301-472X/09 $–see front matter. Copyright Ó 2009 ISEH - Society for Hematology and Stem Cells. Published by Elsevier Inc. doi: 10.1016/j.exphem.2009.08.008

1380

A.C.H. Ma et al./ Experimental Hematology 2009;37:1379–1386

Figure 1. Zebrafish jak2aV581F and its transient expression in zebrafish embryos. (A) Protein alignment of jak2a in zebrafish (ZF), JAK2 in human (HM), and jak2 in mouse (MO) demonstrating the close similarity of these proteins in this region. The valine 617 in human JAK2 and mouse jak2 were equivalent to valine 581 in zebrafish jak2a (*). (B–E) Injection of jak2aV581F mRNA up to 200 pg did not induce any morphological changes in zebrafish embryos at 18somites (B, C) and 48 hours postfertilization (hpf) (D, E). (F, G) Morphology of blood cells extracted at 48 hpf from uninjected (F) and jak2aV581F mRNA injected (G) embryos (Nikon SMZ800, magnification 50, colors corrected after acquisition with Adobe Photoshop). (H) Flow cytometric analysis at 18somites of Tg(gata1:gfp) embryos. (I) Average results of (H) presented in mean 6 standard error of mean from four to six experiments. Comparison between four groups of data was evaluated using the Kruskal-Wallis Test.


1381

Figure 2. Western blot analysis showing increased signal transducer and activation of transcription (stat) 5 signaling in embryos injected with jak2aV581F mRNA. The total amount of protein in each sample was normalized with respect to b-actin. Compared with uninjected embryos (control, lane 1), injection of jak2aV581F mRNA significantly increased stat5 phosphorylation, but not the total amount of stat5 protein (lane 2). Knockdown of stat5.1 significantly reduced both the stat5 phosphorylation and the total amount of stat5 protein (lane 3). Treatment of embryos with a nonselective JAK (AG490) and a specific JAK2 inhibitor (TG101209) significantly reduced stat5 phosphorylation without affecting the total amount of stat5 protein (lane 4, 5). Injection of stat5.1 morpholino as well as treatment with TG101209 ameliorated the enhanced stat5 phosphorylation induced by jak2aV581F mRNA (lane 6, 7). Results were representative of three separate experiments.

intracellular mediator of cytokine receptor function. There are four members in the mammalian JAK family (JAK1, JAK2, JAK3, and TYK2), of which JAK2 plays an important role in the regulation of hematopoiesis [9,10]. In zebrafish, the jak2 gene has undergone duplication into jak2a and jak2b, with distinctive expression patterns in the embryos: jak2a in the intermediate cell mass and jak2b in the developing eyes and pronephric ducts [11]. Morpholinomediated knockdown of jak2a downregulated genes associated with hematopoietic stem cells (HSC) (scl and lmo2), as well as erythroid (gata1 and embryonic hemoglobin) and myeloid (spi1 and myeloperoxidase [mpo]) lineages in a stat5-dependent fashion [12]. In addition to its role in normal hematopoiesis, deregulation of JAK2 has been implicated in human hematological diseases. In particular, a gain-of-function V617F mutation in the JH2 autoregulatory domain of JAK2 (JAK2V617F) was identified in O90% patients with polycythemia vera (PV), underscoring the pathogenetic role of a deregulated JAK2 tyrosine kinase [13–15]. Recently, transgenic jak2V617F murine models [16–18] and transplantation of jak2V617F-transduced donor mouse bone marrow (BM) cells into irradiated recipients [19–21] have been shown to induce PV and other myeloproliferative disease (MPD)-like features. Although activation of jak2a based on a Drosophila mutation could induce expansion of primitive hematopoiesis in zebrafish embryos [12], a zebrafish PV model based on orthologous JAK2V617F mutation is lacking. In the present study, we generated a zebrafish ortholog of human JAK2V617F by site-directed mutagenesis and examined its impact on primitive hematopoiesis, jak/stat signaling, and erythropoietin expression with particular reference to the pathogenetic link to human PV. Materials and methods Zebrafish Wild-type zebrafish were obtained from a local aquarium store and were raised and maintained under standard conditions at 28 C.

Embryos were obtained from natural spawning and were staged according to Kimmel et al. [22]. Transgenic zebrafish lines Tg(gata1:gfp) [23] was obtained from Tsinghua University (a gift from Dr. Anming Meng). Protocols for whole-mount in situ hybridization and Western blot have been described previously [12,24]. Generation of jak2aV581F mutant Wild-type zebrafish jak2a clone in the mammalian expression vector pCS2þ has been described previously [11]. Mutagenesis to generate a zebrafish jak2a mutant equivalent to the human PV JAK2V617F mutation was performed using a QuickChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA), following manufacturer’s recommendations (Supplementary Table E1, online only, available at www.exphem.org). Sequence analysis confirmed that the change in zebrafish jak2a, corresponding to amino acid residue 581 (jak2aV581F), had been correctly introduced in the absence of extraneous mutations. The pCS2þ clone containing the zebrafish jak2aV581F mutant insert was linearized with NotI and in vitro transcribed from the SP6 promoter using a mMessage mMachine Kit (Ambion, Austin, TX, USA). The transcribed mRNA was polyadenylated with a Poly(A) Tailing Kit (Ambion) to increase stability. Modulation of jak2a and stat5.1 In the present study in which stat5.1 signaling was evaluated, morpholino (3 ng) targeting at the splice-junction (SS) of zebrafish stat5.1 (Gene-Tools, LLC, Philomath, OR, USA) [25] was injected into one- to four-cell stage embryos. To modulate jak2a activity, we treated the embryos with a nonselective JAK inhibitor, AG490 (Calbiochem, San Diego, CA, USA) at 50 mmol/L and a specific JAK2 inhibitor, TG101209, at 0.5 mmol/L (Symansis Cell Signaling Science, Timaru, New Zealand). Controls were set up from the same clutches of embryos. The embryos were maintained at 28 C until analyzed. Flow cytometric analysis of erythropoiesis Quantitative analysis of erythropoiesis was performed using transgenic Tg(gata1:gfp) embryos in which the erythroid cells were green fluorescent protein–positive and could be evaluated by flow cytometry. The 18-somite stage embryos were dechorionated and digested with 0.05% (w/v) trypsin/ethylenediamine tetraacetic acid solution (Invitrogen, Carlsbad, CA, USA) for 10 minutes at 28 C, followed by complete dissociation to single-cell suspension


1382

Table 1. Effects of jak2aV581F mRNA injection (200 pg) on lin-eagespecific gene expression based on real-time quantitative polymerase chain reaction of embryos at 18-somite stage, with reference to that of uninjected embryos Gene scl lmo2 gata1 a-eHb b-eHb spi1 mpo l-plastin fli1

a

Control 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

V581F

jak2a

1.05 6 0.03 1.01 6 0.05 1.83 6 0.08 1.61 6 0.12 1.65 6 0.13 1.26 6 0.07 1.12 6 0.06 1.14 6 0.05 1.02 6 0.01

USA). The size and integrity of the synthesized riboprobe were confirmed by RNA formaldehyde gel electrophoresis.

b

p Value 0.275 0.879 0.005 0.049 0.026 0.063 0.124 0.167 0.242

Data are mean 6 standard error of mean of three to four separate experiments. a Control was arbitrarily 1.00 in all cases. b Comparison between uninjected control and injected embryos using paired t-test.

by pipetting. Trypsin digestion was terminated by the addition of CaCl2 (2 mmol/L). Cells were washed and harvested in phosphatebuffered saline with 2% (v/v) fetal bovine serum before examination by flow cytometry. Reverse transcriptase polymerase chain reaction and real-time quantitative polymerase chain reaction (Q-PCR) Total mRNA was obtained from 18-somite stage embryos and complementary DNA (cDNA) generated as described previously [12]. Real-time Q-PCR was performed to examine the relative expression of genes specific for HSC (scl, lmo2), erythroid (gata1, a-embryonic hemoglobin, b-embryonic hemoglobin), early myeloid (spi1), heterophilic granulocyte (mpo), and macrophage (l-plastin) lineages using the ABI Prism 7700 Sequence Detector (Applied Biosystems Inc., Foster City, CA, USA) as described previously [12]. Q-PCR for erythropoietin was similarly performed. Sequences of primers used for Q-PCR were shown in Supplementary Table E2 (online only, available at www.exphem.org). Blood sampling from zebrafish embryos and cytospin preparation Embryos at 48 hpf were anesthetized by 0.016% Tricaine (3-amino benzoic acid ethylester) in 0.9 phosphate-buffered saline and blood cells were collected by aspiration with a pipette after making a stab wound at the Duct of Curvier. Blood cells from 30 embryos were pooled and centrifuged by 7620 Cytopro Cytocentrifuge (Wescor Inc., South Logan, UT, USA) at 800 rpm for 5 minutes. Blood cells were subjected to Wright staining and observed by light microscopy. Cloning and synthesis of riboprobes for erythropoietin (epo) A 1032 bp epo sequence containing the full coding sequence of epo-S [26] were cloned (see Supplementary Materials, Supplementary Figures E1, E2, and Tables E1 and E3, for detail of cloning). The epo riboprobe, which recognized all three epo variants (L1, L2, and S), was synthesized from linearized vector containing the insert. Digoxigenin-labeled anti-sense and sense probes were synthesized by SP6 and T7 RNA polymerase according to manufacturer’s protocols (Roche Applied Science, Indianapolis, IN,

Cloning and synthesis of riboprobes for erythropoietin receptor (epor) A 581 bp segment of the zebrafish epor gene near the 50 UTR was amplified by PCR (Supplementary Table E1, online only, available at www.exphem.org) from cDNA of 18-somites embryos and subcloned into pGEM-T vector (pGEM-T Vector Systems, Promega, Madison, WI, USA). The epor riboprobe was synthesized from linearized vector containing the insert. Digoxigeninlabeled anti-sense and sense probes were synthesized by SP6 and T7 RNA polymerase according to manufacturer’s protocols (Roche Applied Science). The size and integrity of the synthesized riboprobe were confirmed by RNA formaldehyde gel electrophoresis. Chemiluminescent immunoassay for epo level We further examined the effects of jak2aV581F injection on erythropoietin protein expression. Forty embryos at 18 hpf were homogenized in 0.5 mL denaturing protein extraction buffer (6.3 mM Tris.HCl, pH 6.8, 10% glycerol, 5% 2-mercaptoenthanol, 3.5% sodium dodecyl sulfate) in each experiment. Chemiluminescent immunoassay for epo level was performed using the Access2 epo assay System (Beckman Coulter, Inc., Fullerton, CA, USA) following manufacturer’s recommendations. Amount of input proteins were normalized based on b-actin level as shown by Western blot and quantified using software ImageJ (version 1.37, National Institutes of Health, Bethesda, MD, USA). Statistical analysis Data are expressed as mean 6 standard error of mean. Comparisons between paired and nonpaired numerical data were performed using paired Student’s t-tests and Mann-Whitney U-test. Comparisons between groups of data were made using KruskalWallis test. A p value !0.05 was considered statistically significant.

Results Zebrafish ortholog of human JAK2V617F We have previously demonstrated sequence and functional similarities between zebrafish jak2a, human JAK2, and mouse jak2 genes [11,12] and the role of jak2a/stat5.1 pathway during zebrafish primitive erythropoiesis. In the present study, we investigated the PV-associated JAK2 mutation in the zebrafish model. Amino acid sequence alignment in these species showed that they are remarkably conserved at the JH2 domain, with the valine residue at 617 of the human sequence corresponding to that at position 581 in zebrafish (Fig. 1A). To reproduce the PV-associated JAK2V617F mutation, this residue was mutated to phenylalanine (referred hereafter as jak2aV581F). Effects of jak2aV581F mRNA injection on erythropoiesis We then examined the effects of jak2aV581F mRNA injection at one-cell stage. Initial dose-finding study showed that injection of up to 200 pg mRNA was compatible with


1383

Figure 3. Effects of jak2aV581F mRNA injection on erythropoietic genes expression. (A) In situ hybridization for gata1, a-embryonic hemoglobin (a-eHb) and b-embryonic hemoglobin (b-eHb) were done in both uninjected embryos and embryos injected with jak2aV581F mRNA at 18 hours 18-somites. Both lateral view and flat-mount images are shown. Results were representative of three separate experiments. (B) The increase in erythropoietic gene expression upon jak2aV581F mRNA injection, as shown by real-time quantitative polymerase chain reaction (Q-PCR), was ameliorated with concomitant treatment with TG101209. Results were representative of three separate experiments. Comparison between four groups of data was evaluated using the Kruskal-Wallis Test.

normal development and higher dosage was associated with high embryonic mortality rate, with embryos exhibiting common nonspecific toxic phenotypes, including small head, pericardium edema, severely deformed somites, and

curved tail (data not shown). Therefore, 200 pg jak2aV581F mRNA was injected throughout the study. There was no morphological abnormality in the injected embryos up to 48 hpf (Fig. 1B–E). Furthermore, erythrocytes

1384


Figure 4. Effects of jak2aV581F mRNA injection on erythropoietin signaling. In situ hybridization showing expression of erythropoietin (epo) (A, B) and erythropoietin receptor (epor) (C, D) at 18-somites in control (A, C) and jak2aV581F mRNA-injected embryos (B, D). The images were taken from a lateral view with anterior to the left and dorsal to the top. (Nikon SMZ800, magnification 63, colors corrected after acquisition with Adobe Photoshop).

obtained from larvae at 48 hpf showed no cytological difference between injected and uninjected embryos (Fig. 1F and G). Both early and late erythroid progenitors could be observed in both groups of embryos with no noticeable difference in erythroid differentiation. At 18somite stage, jak2aV581F mRNA induced a significant increase in green fluorescent protein–positive cells in Tg(gata1:gfp) embryos. The increase in erythropoiesis could be ameliorated by concomitant knockdown of zebrafish stat5.1 using MO (uninjected embryos: 4.37% 6 0.08%; embryos injected with jak2aV581F mRNA: 5.71% 6 0.07%; embryos injected with jak2aV581F mRNA and stat5.1 MO: 4.66% 6 0.13%, n 5 4 – 6 experiments using O400 embryos in total; p ! 0.01) (Fig. 1H and I). However, injection of zebrafish stat5.1 MO per se had no effect on the basal level of erythropoiesis (4.15% 6 0.03%, n 5 4 experiments using O200 embryos in total; p 5 0.153 as compared with uninjected embryos). Effects of jak2aV581F mRNA injection on stat5 phosphorylation We further examined the effects of jak2aV581F mRNA on jak/ stat pathway activation. The total amounts of protein in each sample were normalized based on b-actin expression in

Western blot (Fig. 2). Injection of jak2aV581F mRNA significantly increased stat5 phosphorylation, but not the total amount of stat5 protein (lane 2). Knockdown of stat5.1 significantly reduced both the stat5 phosphorylation and the total amount of stat5 protein (lane 3). Treatment of embryos with a nonselective JAK (AG490) and a specific JAK2 inhibitor (TG101209) significantly reduced stat5 phosphorylation without affecting the total amount of stat5 protein (lane 4, 5). Importantly, stat5.1 knockdown with morpholino as well as treatment with TG101209 ameliorated the enhanced stat5 phosphorylation induced by jak2aV581F mRNA (lane 6, 7). Collectively, these results indicate that stat5 activation is an important downstream mediator in the pathogenic jak2aV581F signaling. Effects of jak2aV581F mRNA injection on lineage-specific gene expression The increase in erythropoiesis upon activation of zebrafish jak2a led us to examine its effects on lineage-specific gene expression by real-time Q-PCR. Genes encoding for erythropoiesis (gata1, embryonic a- and b-hemoglobin) were significantly upregulated, whereas those encoding for HSC (lmo2, scl), as well as early (spi1) and late (mpo: granulocytic; l-plastin: macrophage) myeloid lineages, showed

A.C.H. Ma et al./ Experimental Hematology 2009;37:1379–1386 Table 2. Effects of jak2aV581F mRNA injection on erythropoietin (epo) and erythropoietin receptor (epor) expression based on real-time quantitative polymerase chain reaction of embryos at 18-somite stage, with reference to that of uninjected embryos Gene

Controla

jak2aV581F

p Valueb

epo epor

1.00 1.00

0.24 6 0.03 1.75 6 0.07

0.006 0.004

Data are mean 6 standard error of mean of three to four separate experiments. a Control was arbitrarily 1.00 in all cases. b Comparison between uninjected control and injected embryos using paired t-test.

no significant changes (Table 1). The upregulation of genes associated with erythropoiesis was further confirmed by in situ hybridization (Fig. 3A). Importantly, the increase in erythropoiesis upon jak2aV581F mRNA injection was ameliorated with concomitant treatment with TG101209 (Fig. 3B). Effects of jak2aV581F mRNA injection on erythropoietin signaling Both zebrafish epo and its receptor epor have been shown to mediate normal hematopoiesis and an increase in epo was associated with polycythemia induced by hypoxia [26,27]. In human PV, the level of serum epo is often reduced. Therefore, we examined the effects of jak2a activation on the expression of these genes. Injection of jak2aV581F mRNA resulted in significant downregulation of epo and upregulation of epor (Fig. 4). The results were confirmed quantitatively using real-time Q-PCR (Table 2). To further substantiate our findings that jak2aV581F mRNA downregulated erythropoietin expression, we performed immunoassay for epo in zebrafish. The test was based on a combination of polyclonal and monoclonal antibodies against full-length human epo. It has been shown to correlate with those obtained with an antibody targeting a conserved region of epo protein in fish [27]. Indeed, injection of jak2aV581F mRNA significantly reduced zebrafish epo protein level (control: 0.633 6 0.11; jak2aV581F mRNA: 0.222 6 0.07 mIU/mL; p 5 0.019).

Discussion We have previously demonstrated that jak2a played an important role in initiating primitive hematopoiesis in zebrafish embryos based on morpholino injection as well as treatment with a nonselective JAK inhibitor, AG490 [12]. In this study, we generated a zebrafish jak2aV581F mutant, an ortholog of human JAK2V617F mutation. Expression of jak2aV581F mRNA induced a significant increase in erythropoiesis and downregulation of erythropoietin expression. Furthermore, there were a number of interesting findings that might provide important insights to erythropoiesis during normal development as well as neoplastic transformation.

1385

First, transient expression of jak2aV581F in zebrafish embryos shared features of human PV. Erythropoiesis was significantly increased and erythropoietin reduced at both mRNA and protein levels. These phenotypes recapitulated those seen in patients with PV as well as in murine models carrying the jak2V617F mutation [16,20,21]. The mechanistic link between the increase in erythropoiesis and reduction in epo level was not examined. Further investigation should determine if the decrease in epo level is a negative feedback control involving hypoxia-inducible factor. Specifically, the increase in erythropoiesis approximated 30%, a magnitude commonly encountered in PV patients. Therefore, our data highlighted a conserved phenomenon whereby Jak2 hyperactivation induces hematopoietic expansion in Drosophila, zebrafish, mice, and humans [28,29]. Second, jak2aV581F expression activated stat5 phosphorylation and the response could be modulated by specific stat5.1 knockdown and jak2 inhibition. These findings provided a mechanistic link between jak2a hyperactivation and increased erythropoiesis. Third, stat5.1 morpholino, at the present dose tested, induced differential effects on basal and activated erythropoiesis. In particular, despite its ability to ameliorate activated erythropoiesis upon jak2aV581F expression, the morpholino had no effect on basal erythropoiesis. We surmised that basal erythropoiesis required only low levels of baseline stat5.1 activity. The gene dosage effects of stat5.1 on erythropoiesis would have to be further evaluated. A number of observations remain unexplained. For instance, jak2aV581F expression induced an increase in erythroid but not myeloid development in this model. This is in contrast to the murine PV models, in which not only the erythroid but also the granulocytic and megakaryocytic lineages underwent hyperplasia. It was possible that expansion of the latter lineages was limited by the restricted differentiation repertoire of primitive hematopoiesis, which is primarily erythroid in lineage. Furthermore, although a subset of lmo2þ and sclþ cells are associated with primitive erythropoiesis [3], jak2aV581F mRNA injection had no effect on their expression. We speculated that an increase in erythropoiesis might not manifest as upregulation of lmo2 and scl gene expression at whole-embryo level because these genes are also expressed in nonerythroid cells. Our findings have also provided us with important leads for further research. For example, the effects of jak2aV581F expression in adult hematopoiesis should be further investigated. A transgenic jak2aV581F zebrafish model is being generated in our laboratory and such a model will enable us to further investigate if this mutant gene in zebrafish phenocopies human PV beyond primitive hematopoiesis. This could not be addressed with the transient expression system reported in the present study. It would also be interesting to see if the hematopoietic cells from adult zebrafish could form spontaneous erythroid colonies. In addition, the activity of jak2aV581F reported in this study was based

1386


primarily on in vivo overexpression in zebrafish embryos. It remains to be demonstrated if the jak2aV581F mutant could activate JAK/stat pathway in mammalian cell lines. Results of the present study have a clinical relevance. The unique pathogenic link between JAK2V617F mutation and human PV has led to the use of JAK2 inhibitor in the treatment of this disorder [30–32]. The zebrafish model of PV based on jak2aV581F expression is unique in its simplicity and cost-effectiveness and might provide a robust model for screening therapeutic agents targeting at JAK2 signaling. Acknowledgment The project was supported by the General Research Fund (HKU 7520/06 M and HKU 770308 M) and a grant from the strategy research theme of cancer stem cells in the University of Hong Kong.

Conflict of Interest Disclosure The authors in the manuscript declared no conflict of interest.

References 1. Lensch MW, Daley GQ. Origins of mammalian hematopoiesis: in vivo paradigms and in vitro models. Curr Top Dev Biol. 2004;60:127–196. 2. Carradice D, Lieschke GJ. Zebrafish in hematology: sushi or science? Blood. 2008;111:3331–3342. 3. Davidson AJ, Zon LI. The ‘definitive’ (and ‘primitive’) guide to zebrafish hematopoiesis. Oncogene. 2004;23:7233–7246. 4. Lieschke GJ, Oates AC, Paw BH, et al. Zebrafish SPI-1 (PU.1) marks a site of myeloid development independent of primitive erythropoiesis: implications for axial patterning. Dev Biol. 2002;246:274–295. 5. Onnebo SM, Yoong SH, Ward AC. Harnessing zebrafish for the study of white blood cell development and its perturbation. Exp Hematol. 2004;32:789–796. 6. Herbomel P, Thisse B, Thisse C. Ontogeny and behaviour of early macrophages in the zebrafish embryo. Development. 1999;126:3735–3745. 7. Burns CE, DeBlasio T, Zhou Y, Zhang J, Zon L, Nimer SD. Isolation and characterization of runxa and runxb, zebrafish members of the runt family of transcriptional regulators. Exp Hematol. 2002;30:1381–1389. 8. Hsia N, Zon LI. Transcriptional regulation of hematopoietic stem cell development in zebrafish. Exp Hematol. 2005;33:1007–1014. 9. Rane SG, Reffy EP. JAKs, STATs and Src kinases in hematopoiesis. Oncogene. 2002;21:3334–3358. 10. Ward AC, Touw I, Yoshimura A. The Jak-Stat pathway in normal and perturbed hematopoiesis. Blood. 2000;95:19–29. 11. Oates AC, Brownlie A, Pratt SJ, et al. Gene Duplication of zebrafish JAK2 homologs is accompanied by divergent embryonic expression patterns: only jak2a is expressed during erythropoiesis. Blood. 1999; 94:2622–2636. 12. Ma AC, Ward AC, Liang R, Leung AY. The role of jak2a in zebrafish hematopoiesis. Blood. 2007;110:1824–1830. 13. James C, Ugo V, Le Couedic JP, et al. A unique clonal JAK2 mutation leading to constitutive signaling causes polycythaemia vera. Nature. 2005;434(7037):1144–1148.

14. Kralovics R, Passamonti F, Buser AS, et al. A gain-of-function mutation of JAK2 in myeloproliferative disorders. N Engl J Med. 2005;352: 1779–1790. 15. Baxter EJ, Scott LM, Campbell PJ, et al. Cancer Genome Project. Acquired mutation of the tyrosine kinase JAK2 in human myeloproliferative disorders. Lancet. 2005;365(9464):1054–1061. 16. Xing S, Wanting TH, Zhao W, et al. Transgenic expression of JAK2V617F causes myeloproliferative disorders in mice. Blood. 2008;111:5109–5117. 17. Shide K, Shimoda HK, Kumano T, et al. Development of ET, primary myelofibrosis and PV in mice expressing JAK2 V617F. Leukemia. 2008;22:87–95. 18. Tiedt R, Hao-Shen H, Sobas MA, et al. Ratio of mutant JAK2-V617F to wild-type Jak2 determines the MPD phenotypes in transgenic mice. Blood. 2008;111:3931–3940. 19. Lacout C, Pisani DF, Tulliez M, Gachelin FM, Vainchenker W, Villeval JL. JAK2V617F expression in murine hematopoietic cells leads to MPD mimicking human PV with secondary myelofibrosis. Blood. 2006;108:1652–1660. 20. Wernig G, Mercher T, Okabe R, Levine RL, Lee BH, Gilliland DG. Expression of Jak2V617F causes a polycythemia vera-like disease with associated myelofibrosis in a murine bone marrow transplant model. Blood. 2006;107:4274–4281. 21. Bumm TG, Elsea C, Corbin AS, et al. Characterization of murine JAK2V617F-positive myeloproliferative disease. Cancer Res. 2006; 66:11156–11165. 22. Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF. Stages of embryonic development of the zebrafish. Dev Dyn. 1995; 203:253–310. 23. Long Q, Meng A, Wang H, Jessen JR, Farrell MJ, Lin S. GATA-1 expression pattern can be recapitulated in living transgenic zebrafish using GFP reporter gene. Development. 1997;124:4105–4111. 24. Leung AY, Mendenhall EM, Kwan TT, et al. Characterization of expanded intermediate cell mass in zebrafish chordin morphant embryos. Dev Biol. 2005;277:235–254. 25. Lewis RS, Stephenson SE, Ward AC. Constitutive activation of zebrafish Stat5 expands hematopoietic cell populations in vivo. Exp Hematol. 2006;34:179–187. 26. Chu CY, Cheng CH, Chen GD, et al. The zebrafish erythropoietin: functional identification and biochemical identification. FEBS Lett. 2007;581:4265–4271. 27. Lai JC, Kakuta I, Mok HO, Rummer JL, Randall D. Effects of moderate and substantial hypoxia on erythropoietin levels in rainbow trout kidney and spleen. J Exp Biol. 2006;209:2734–2738. 28. Paffett-Lugassy N, Hsia N, Fraenkel PG, et al. Functional conservation of erythropoietin signaling in zebrafish. Blood. 2007;110:2718–2726. 29. Luo H, Dearolf CR. The JAK/STAT pathway and Drosophila development. Bioessays. 2001;23:1138–1147. 30. Pardanani A, Hood J, Lasho T, et al. TG101209, a small molecule JAK2-selective kinase inhibitor potently inhibits myeloproliferative disorder-associated JAK2V617F and MPLW515L/K mutations. Leukemia. 2007;21:1658–1668. 31. Manshouri T, Quintas-Cardama A, Nussenzveig RH, et al. The JAK kinase inhibitor CP-690,550 suppresses the growth of human polycythemia vera cell carrying the JAK2V617F mutation. Cancer Sci. 2008; 99:1265–1273. 32. Werning G, Kharas MG, Okabe R, et al. Efficacy of TG101348, a selective JAK2 inhibitor, in treatment of a murine model of JAK2V617Finduced polycythemia vera. Cancer Cell. 2008;13:311–320.


1386.e1

Supplementary materials and methods Cloning and synthesis of riboprobes for epo The amino acid sequence of human epo was used for BLAST search zebrafish genome Zv4 Ensembl database (http://www .ensembl.org/Danio_rerio), and a predicted novel transcript was identified. RT-PCR was performed to amplify a 500-bp sequence from cDNA obtained from 24 hpf embryos (Supplementary Table E1) and the PCR product was subcloned into pCR2.1-TOPO vector (Invitrogen) and sequenced. To obtain full-length sequence, gene specific primers for 50 - and 3 0 -RACE were designed based on the nucleotide sequence of cloned RT-PCR product. 5 0 -RACE and 3 0 -RACE were performed using the SMART cDNA amplification kit (Clontech Laboratories, Palo Alto, CA, USA) according to the manufactur’s manual. The RACE products were subcloned into pBlueScript II KSþ vector (Stratagene, La Jolla, CA, USA) and sequenced. After confirming the 50 - and 30 -UTR, RT-PCR was carried out using primers designed from sequence of the cloned RACE products. The RT-PCR product of 1032 bp was subcloned into pBlueScript II KSþ vector for sequencing. Digoxigenin-labeled anti-sense and sense epo riboprobe were synthesized from linearized vector by SP6 and T7 RNA polymerase according to manufacturer’s protocols (Roche Applied Science, Indianapolis, IN, USA). The size and integrity of the synthesized riboprobe were confirmed by RNA formaldehyde gel electrophoresis. Mapping and cross-species alignment of the cloned epo NCBI BLAST program was used for gene analysis to obtain genes similar to the cloned product. Chromosome location of the cloned gene was determined by BLAST search the sequence of cloned gene against the zebrafish genome Zv4 Ensembl database (http://www.ensembl.org/Danio_rerio). Alignment of mRNA-to-genomic DNA was used to determine the genomic organization of the cloned gene. The alignment was performed with the BioEdit biological sequence alignment editor (Tom Hall, Iris Pharmaceuticals Inc.). Gaps were inserted into cDNA sequence of the cloned gene and manually adjusted in order to maintain exact matches in exons and to have introns with canonical dinucleotides ‘‘GT’’ at the 50 end and ‘‘AG’’ at the 30 end of the introns. Multiple sequence alignment of zebrafish epo with orthologs of other organisms was performed using ClustalW (http:// www.ebi.ac.uk/clustalw). Orthologs peptide sequences were downloaded from GenBank. The accession number of the orthologs were: human (Homo sapiens) NP_000790, mouse (Mus musculus) NP_031968, rat (Rattus norvegicus) NP_058697, dog (Canis familiaris) NP_00100647, cat (Felix catus) NP_001009269, cow (Bos taurus) NP_776334, pig (Sus scrofa) NP_999299, sheep (Ovis aries) MP_001019908, horse (Equus cabalus) BAC55239, common carp (Cyprinus carpio) ABB83930, spotted green pufferfish (Tetraodon nigroviridus) AAR25698 and orange spotted grouper (Epinephelus coioides) AAW29029. Phylogenetic tree was prepared using PHYLIP program (version 3.6b by J. Felsenstein, 2004; http://evolution .genetics.washington.edu/phylip.html) with Bootstrap value of 1000 replicates. Tree was drawn by TreeExplorer version 2.12.

Supplementary results Cloning and mapping of zebrafish erythropoietin The full-length zebrafish epo transcript was 1117-bp long, containing the 50 -UTR, the coding sequence and 30 -UTR of 234, 516, and 367 bp, respectively (Supplementary Fig. E1). The deduced peptide sequence consisted of 171 amino acid residues which is shorter than previously reported (GenBank accession number NP_001033098). The amino acid sequence of zebrafish epo was aligned with orthologs from common carp, spotted green pufferfish, orange spotted grouper, human, mouse, rat, dog, cat, cow, pig, sheep, and horse. Zebrafish epo protein shared 89%, 62%, and 57% sequence similarity with common carp, grouper and pufferfish orthologs, respectively (Supplementary Table E3), but !40% with mammalian genes. According to the entry of human erythropoietin in PROSITE, there is a signature sequence (P – x(4) – C – D – x – R –[LIVM](2) – x – [KR] – x(14) – C; underlined in Supplementary Fig. E2A) for erythropoietin/thrompoietin. The signature sequence contains two cysteine residues, Cys34 and Cys56, which are involved in the formation of two disulfide bonds (Cys34 – Cys188 and Cys56–Cys60; highlighted in Supplementary Fig. E2A). Both signature and cysteine residues were highly conserved in all 13 epo proteins, including zebrafish. A phylogenetic tree was showed in Supplementary Figure E2B. Zebrafish epo with the other teleost orthologs formed a clade that was separated from the mammalian counterparts.

1386.e2


Supplementary Figure E1. Nucleotide sequence and deduced amino acid sequences of zebrafish erythropoietin (epo). Only the open reading frame and 30 untranslated sequence are shown.


1386.e3

A

B

Supplementary Figure E2. Zebrafish epo gene is highly conserved. (A) Multiple sequence alignment of erythropoietin (EPO) protein in different species. EPO proteins from human, mouse, rat, cat, cow, pig, dog, sheep, horse, common carp, spotted green pufferfish, orange spotted grouper, and zebrafish were aligned using ClustalW program. Identical residues are presented as dots. Conserved cysteine residues for disulfide bonding were highlighted. EPO/thrombopoietin (TPO) signature sequence was underlined. (B) Phylogenetic tree of EPO gene. The clade of teleost EPOs was indicated by brace.


1386.e4

Supplementary Table E1. Sequences of morpholino and gene specific primers used Oligo

Sequence

Morpholinos stat5.1SS morpholino Primers for site-directed mutagenesis of zebrafish jak2a Anti-sense primer Sense primer Primers for the RT-PCR of zebrafish epo ZF epoRTf ZF epoRTr Primers for the 50 RACE of zebrafish epo ZF epo50 RACE-1stRound ZF epo50 RACE-2ndRound Primers for the 3’RACE of zebrafish epo ZF epo30 RACE-1stRound ZF epo30 RACE-2ndRound Primers for the cloning of the full length zebrafish epo ZF epoRTf ZF epoRTr Primers for the cloning of epor probe ZF eporProbef ZF eporProber

GTG AAC TTG TGA CTT ACC AGA GTT G CCA TGA TGT GTT CAT CCG CGC AAA AAC AGA TGC CGT AGT TCA GCT GAA CTA CGG CAT CTG TTT TTG CGC GGA TGA ACA CAT CAT CAG GAC TCT TTG CCT TAC TGC T TGT CGA CAG ACA GGT GCA TTG TGT AAG CCT GAC TGG ACC TCC TGA GC CTC TGC ATC CCA TGC CTC CTT AAT GA ACG AGG CCA TCG GCT CAT TAC AGA TA TCA GAC AAG TGC TGC GAA GTC TCA GC CTA GCC TCA ACG TCC CGA TCT AC CAG GGA TTC CCG CTC TAG GAA ATA CCG AGA TGA AAC CGA GGA TA TCC TCC ACA GCG TAT CTG AC

Supplementary Table E2. Primers for real-time quantitative reverse transcription polymerase chain reaction Oligo ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF ZF

b-actinf b-actinr sclf sclr lmo2f lmo2r gata1f gata1r a-eHbf a-eHbr b-eHbf b-eHbr fli1f fli1r spi1f spi1r l-plastinf l-plastinr mpof mpor jak2af jak2ar epof epor eporf eporr

Sequence TTC CTT CCT GGG TAT GGA ATC GCA CTG TGT TGG CAT ACA GG CTA TTA ACC GTG GTT TTG CTG G CCA TCG TTG ATT TCA ACC TCA T GGA CGC AGG CTT TAC TAC AAA C CCG GAT CCT CTT TTC ACA GGA A AAG ATG GGA CAG GCC ACT AC TGC TGA CAA TCA GCC TCT TTT TGC TCT CTC CAG GAT GTT GA TCA CAG TCT TGC CGT GTT TC AGG CTC TGG CAA GGT GTC TCA CAT TGG GTT TCC CAG GAT CAA CGG ATC CAG AGA GTC G CCA TGT AGC CAG TAT AGT TCA TCT G GGG CAG TTT TAA CCA AAG ATC A CCC AAG AGT GAT CGT TCT GAC GAA GCT CTG ATC GCT CTG CT GCT TCT TTT CAT CCG TCA GG GGG GCA GAA GAA GAA AGT CC CCC TTG CTA AAC TCT CAT CTC G GGT GAG GAT AAA CAG GGT CAG A ATG GAT CTC TGC TGG ACA CC CAG ACA AGT GCT GCG AAG TC GCT TTT CCC CGA AGA AAG TT TCA CAG GAC TGG TCC AAG AA GTC AGA TAC GCT GTG GAG GA

Based on data from reference 12.

Supplementary Table E3. Percentage of similarity of erythropoietin in various species Organism Human Mouse Rat Cow Sheep Pig Cat Dog Horse Common carp Spotted green pufferfish Orange spotted grouper

Similarity (%) 35 39 36 35 35 36 35 33 35 89 62 57