Introduction

Hematopoietic stem cells (HSCs) are self-renewing multipotent cells that can generate all types of blood cells over the lifetime of an individual and can be used therapeutically to treat hematopoietic diseases1. In the adult, most HSCs present in bone marrow are quiescent and divide rarely under homeostatic conditions. HSCs produce a heterogeneous pool of hematopoietic progenitor cells (HPCs), which have limited or no self-renewal ability, but rapidly proliferate and differentiate to satisfy the requirements for new mature blood cells2,3. Although the frequency of HSCs is extremely rare in bone marrow, HSC potential can be evaluated by transplantation assays, whereby the relative hematopoietic reconstitution activity of co-transplanted donor and competitor cells are compared in a recipient4. Purification of HSCs from murine and human bone marrow has been facilitated via transplantation assays using combinations of multiple cell-surface markers5,6,7,8,9. Studies in mice revealed that a single CD150+ CD34− c-kit+ Sca-1+ Lineage-marker− cell in the bone marrow showed long-term and multilineage hematopoietic reconstitution following transplantation10,11. Prospective isolation of highly purified HSCs thus elucidated many aspects of HSC biology, including self-renewal, differentiation, and HSC niches.

The zebrafish is an excellent model for the study of HSCs due to its many unique advantages. Many valuable tools and experimental methods have been established for the study of hematopoietic cells in zebrafish (e.g. transgenic/mutant animals, transplantation assays, cell culture assays, etc.)12,13. Moreover, genome-editing technology based on the CRISPR/Cas9 system has further facilitated rapid and scalable reverse genetic approaches in zebrafish14,15,16. Hematopoiesis is highly conserved between mammals and teleost fish at both the cellular and molecular level, while the sites of adult hematopoiesis have shifted during the evolution. The major hematopoietic organ in teleost fish is the kidney where various stages of hematopoietic cells and mature blood cells are observed in the interstitial tissue, the so-called “kidney marrow”17,18,19,20. Although the zebrafish kidney is of great importance to identify evolutionarily conserved regulators of HSCs, little is known regarding which molecules are involved in the maintenance and self-renewal of HSCs in the kidney. This is due in part to the lack of robust methods to purify HSCs in the zebrafish kidney.

The enrichment of HSCs from the kidney has been shown previously in zebrafish. A previous method using the light scatter profile of whole kidney marrow cells (WKMCs) by flow cytometric (FCM) analysis demonstrated that HSCs are present within the forward scatter (FSC)low side scatter (SSC)low “lymphoid” fraction17. The Hoechst dye efflux activity of HSCs is highly conserved amongst vertebrates21, and an enriched HSC fraction has been isolated by sorting of “side population” (SP) cells in the kidney22,23. More recently, two groups demonstrated using a transgenic zebrafish line that HSCs are present in the cd41:GFPlow fraction or runx1:mCherry+ fraction in the kidney24,25. These transgenic animals are also utilized to visualize hematopoietic stem/progenitor cells (HSPCs) in developmental stages25,26,27. The expression of cd41:GFP and runx1:mCherry as well as the SP phenotype is, however, not specific to HSCs, indicating that a combination of multiple markers is required to further purify HSCs in the zebrafish kidney, as has been proven in mammalian bone marrow5,6,7,8,9.

In the present study, we combined two transgenic markers of putative HSCs, gata2a:GFP and runx1:mCherry, and found that zebrafish HSCs are highly enriched in the double-positive fraction (gata2a:GFP+ runx1:mCherry+, hereafter denoted as gata2a+ runx1+) in the kidney. Transcriptome analysis of three distinct hematopoietic cell populations, gata2a+ runx1+, gata2a− runx1+, and gata2a+ runx1− cells, revealed that gata2a+ runx1+ cells displayed typical molecular hallmarks of HSCs. In contrast, gata2a− runx1+ showed expression signatures of erythroid and myeloid cells, whereas gata2a+ runx1− cells showed those of lymphoid and myeloid cells. Transplantation assays demonstrated that gata2a+ runx1+ cells possessed the high level of long-term and multilineage hematopoietic reconstitution activity. Thus, we provide evidence that combined gata2a:GFP and runx1:mCherry expression is a useful method to purify HSCs from the zebrafish kidney.

Results

Isolation of HSPCs using gata2a:GFP; runx1:mCherry double-transgenic zebrafish

In order to purify HSCs from the adult kidney, we utilized gata2a (GATA binding protein 2a) as an HSC marker, of which expression in HSCs and vascular endothelial cells is well-conserved amongst vertebrates23,28,29,30. A zebrafish gata2a:GFP line expresses GFP in a variety of hematopoietic cells and vascular endothelial cells28. We combined the gata2a:GFP line with the runx1:mCherry line, which expresses mCherry under control of the mouse Runx1 (Runt-related transcription factor 1) +23 enhancer25. FCM analysis of kidney marrow cells (KMCs), which contain hematopoietic cells and mature blood cells excluding erythrocytes, showed that the majority of gata2a:GFP+ cells and runx1:mCherry+ cells within the SSClow fraction (non-granulocytic cells) did not overlap, whereas only 0.31% of SSClow cells were found within the gata2a+ runx1+ population (0.16 ± 0.06% in total KMCs, n = 23, ± s.d.) (Fig. 1a,b). HSCs have been shown to be present in the FSClow SSClow “lymphoid” fraction by FCM analysis in the zebrafish kidney17. As shown in Fig. 1c, we subdivided SSClow cells into three distinct fractions based on the intensity of FSC, “low” (in the range of 30 to 70), “mid” (in the range of 70 to 100), and “high” (>100) fraction. Unexpectedly, most gata2a+ runx1+ cells showed the “mid” intensity of FSC. In contrast, gata2a− runx1+ cells showed the “low” to “high” intensity, and gata2a+ runx1− cells mainly showed the “low” intensity of FSC (Fig. 1c).

Figure 1
figure 1

Isolation of HSPCs from the zebrafish kidney. FCM analysis of KMCs was performed in a gata2a:GFP; runx1:mCherry double-transgenic zebrafish. (a) A scatter profile of KMCs. The SSClow non-granulocytic cell fraction is gated. (b) SSClow cells are subdivided into three distinct hematopoietic populations, gata2a+ runx1+ (orange gate), gata2a− runx1+ (red gate), and gata2a+ runx1− (green gate). (c) Contour plot of FSC vs. SSC in KMCs, gata2a+ runx1+, gata2a− runx1+, and gata2a+ runx1− cells. Dotted lines separate the “low”, “mid”, and “high” intensity of FSC.

Full size image

To characterize hematopoietic cell populations in the kidney, RNA-seq analysis was performed on three distinct cell populations within the SSClow fraction, gata2a+ runx1+, gata2a− runx1+, and gata2a+ runx1−. As shown in Fig. 2a, HSC-related genes, such as gata2b (GATA binding protein 2b), gfi1aa (growth factor independent 1A), runx1t1 (runt-related transcription factor 1; translocated to, 1), pbx1b (pre-B-cell leukemia homeobox 1b), fgd5b (FYVE, RhoGEF and PH domain containing 5b), meis1b (Meis homeobox 1b), myca (MYC proto-oncogene, bHLH transcription factor a), apoeb (apolipoprotein Eb), and egr1 (early growth response 1), were highly enriched in gata2a+ runx1+ cells. In contrast, erythroid and thrombocyte marker genes, such as gata1a (GATA binding protein 1a), klf1 (Kruppel-like factor 1), hbaa1 (hemoglobin, alpha adult 1), and itga2b (integrin, alpha 2b, also known as cd41), were strongly expressed in gata2a− runx1+ cells, whereas myeloid marker genes, such as mpx (myeloid-specific peroxidase), lyz (lysozyme), lcp1 (lymphocyte cytosolic protein 1, also known as l-plastin), and spi1b (Spi-1 proto-oncogene b, also known as pu.1), were expressed in both gata2a− runx1+ cells and gata2a+ runx1− cells. Lymphoid marker genes, such as rag1 (recombination activating gene 1), cd8a, cd4-1, lck (Lymphocyte-specific protein-tyrosine kinase), and ighz (immunoglobulin heavy constant zeta), were detected only in the gata2a+ runx1− fraction, whereas a part of lymphoid marker genes, such as tcra (t cell receptor alpha), tcrd (t cell receptor delta), mhc2dab (major histocompatibility complex class II DAB gene), and mhc2dbb (major histocompatibility complex class II DBB gene), were undetected in gata2a+ runx1− cells or within other fractions (Supplementary Table S1). To further characterize these hematopoietic cell populations, gene ontology enrichment analysis was performed. We found that genes involved in “intracellular signal transduction” and “cell communication” were significantly enriched in gata2a+ runx1+ cells, suggesting that these cells actively communicate with environmental niche cells. In contrast, genes involved in “erythrocyte differentiation” and “myeloid cell differentiation” are enriched in gata2a− runx1+ cells, whereas genes involved in “immune response” and “chemotaxis” are enriched in gata2a+ runx1− cells (Fig. 2b). Quantitative PCR (qPCR) analysis also showed that HSC-related genes, gata2b and gfi1aa, were highly expressed in gata2a+ runx1+ cells, whereas the expression of kita (KIT proto-oncogene, receptor tyrosine kinase a), myb (v-myb avian myeloblastosis viral oncogene homolog), and mpl (MPL proto-oncogene, thrombopoietin receptor) was detected in both gata2a+ runx1+ cells and gata2a− runx1+ cells. In contrast, erythroid and thrombocyte marker genes, gata1a and itga2b, were strongly expressed in gata2a− runx1+ cells, and lymphoid marker genes, lck, ighm (immunoglobulin heavy constant mu), and rag1, were expressed only in gata2a+ runx1− cells. A neutrophil marker, mpx, and macrophage marker, lcp1, were predominantly expressed in gata2a− runx1+ cells and gata2a+ runx1− cells, respectively (Fig. 3). These expression patterns in qPCR analysis were consistent with those in RNA-seq analysis, confirming the reliability and validity of our transcriptome analysis. Collectively, these data suggest that gata2a+ runx1+ cells have typical molecular hallmarks of HSCs, whereas gata2a− runx1+ show expression signatures of erythroid and myeloid cells and gata2a+ runx1− cells show those of lymphoid and myeloid cells.

Figure 2
figure 2

Transcriptome analysis of three distinct hematopoietic populations. (a) Hierarchical clustering of selected HSC- (upper), erythroid-myeloid- (middle), and lymphoid-related genes (lower). (b) Gene ontology enrichment analysis of highly expressed genes in gata2a+ runx1+ (upper), gata2a− runx1+ (middle), or gata2a+ runx1− cells (lower).

Full size image
Figure 3
figure 3

Expression of hematopoietic marker genes in HSPC populations. Relative expression levels of hematopoietic marker genes are shown. Orange, red, and green bars denote gata2a+ runx1+, gata2a− runx1+, and gata2a+ runx1− cells, respectively. The expression level in the kidney tissue is shown as 1.0 in each panel. Error bars, s.d.

Full size image

gata2a + runx1 + cells possess long-term and multilineage hematopoietic reconstitution activity

HSC potential can be evaluated via an in vivo competitive repopulation assay, in which contributions of donor- and competitor-derived cells are compared in an irradiated recipient4. To determine if HSCs are enriched in the gata2a+ runx1+ fraction, we performed an in vivo competitive repopulation assay using a triple transgenic zebrafish, gata2a:GFP; runx1:mCherry; bactin2:BFP, which labels whole blood cells with BFP with the exception of mature erythrocytes where bactin2 promoter is not active31. We also utilized a double transgenic animal, kdrl:Cre; bactin2:loxP-STOP-loxP-DsRed (“kdrl-sw”), as a competitor. As has been previously demonstrated31, embryonic expression of the Cre recombinase in the shared vascular precursors of HSCs via the kdrl regulatory elements results in nearly all adult leukocytes becoming labeled with DsRed. One hundred BFP-labeled gata2a+ runx1+, gata2a− runx1+, and gata2a+ runx1− cells were separately sorted and co-transplanted together with 50,000 DsRed-labeled KMCs (competitor) into sublethally irradiated recipients. At 16 weeks post-transplantation (wpt), donor chimerism was determined by the percentage of BFP+ cells within the total BFP+ (donor-derived) and DsRed+ (competitor-derived) cells in each recipient kidney (Fig. 4a). Although competitor-derived DsRed+ cells were detected in all surviving recipients, BFP+ donor-derived cells were detected only in recipients transplanted with gata2a+ runx1+ cells with the exception of one recipient transplanted with gata2a+ runx1− cells. The mean percentage of BFP+ cells within total BFP+ and DsRed+ cells was 51.9 ± 13.5% (n = 9, ±s.e.m) in recipients transplanted with gata2a+ runx1+ cells (Fig. 4b,c). BFP+ cells derived from gata2a+ runx1+ cells were resolved in three distinct blood cell populations, “granulocyte”, “precursor” and “lymphoid”, with the similar ratio to competitor-derived DsRed+ cells (Fig. 5a,b). Moreover, lineage marker genes including mpx (granulocyte marker), lcp1 (macrophage marker), tcra (T cell marker), and ighm (B cell marker) were detected in isolated donor-derived BFP+ cells as well as competitor-derived DsRed+ cells (Fig. 5c), indicating that gata2a+ runx1+ cells possess the ability of long-term and multipotent hematopoietic reconstitution. Based on the contribution of BFP+ cells to DsRed+ cells, the frequency of HSCs in gata2a+ runx1+ cells was estimated to be approximately 540 times higher than that in KMCs, which is close to the frequency of gata2a+ runx1+ cells in KMCs (approximately 1/625). Collectively, these data suggest that most long-term repopulating HSCs are present in the gata2a+ runx1+ fraction in the zebrafish kidney.

Figure 4
figure 4

Transplantation assays of HSPC populations. (a) Experimental procedure of a competitive repopulation assay. One hundred BFP-labeled donor cells were co-transplanted with 50,000 of DsRed-labeled KMCs (competitors) into a sublethally irradiated recipient animal. At 16 wpt, donor chimerism was determined by the percentage of BFP+ cells within the total BFP+ and DsRed+ cells in each recipient. (b) Representative result of FCM analysis in a recipient transplanted with gata2a+ runx1+ cells. (c) Percentage of donor chimerism in each recipient group (mean ± s.e.m).

Full size image
Figure 5
figure 5

Multilineage differentiation of gata2a+ runx1+ cells. (a) Representative result of FCM analysis of KMCs in a recipient co-transplanted with BFP-labeled gata2a+ runx1+ cells and DsRed-labeled competitors. G, granulocyte; P, precursor; L, lymphoid. (b) Percent distribution of gata2a+ runx1+-derived BFP+ cells in “granulocyte”, “precursor”, and “lymphoid”. Data show the mean ± s.e.m. in the recipients (n = 6). (c) Relative expression level of lineage marker genes in isolated gata2a+ runx1+-derived BFP+ cells in a recipient. Red dotted line indicates the expression level of each gene in competitor-derived DsRed+ cells.

Full size image

gata2a − runx1 + cells abundantly contain hematopoietic progenitor cells

HSCs produce a heterogenous pool of HPCs, including multipotent progenitors (MPPs). The potential of HPCs can be evaluated by an in vitro colony-forming assay. To examine the frequency of HPCs in each hematopoietic subset, we performed colony-forming assays, which can determine the percentage of colony-forming unit-erythroid (CFU-E) and -granulocyte (CFU-G). One hundred BFP-labeled gata2a+ runx1+, gata2a− runx1+, and gata2a+ runx1− cells were separately sorted and plated in conditioned media containing Erythropoietin a (Epoa) or Granulocyte colony stimulating factor b (Gcsfb). After 7 days of culture, the percentage of CFU-E and CFU-G in each subset was calculated based on the number of colonies formed. We detected both CFU-E and CFU-G from the wells plated with each hematopoietic subset (Fig. 6a). The percentage of CFU-E in gata2a− runx1+ cells was approximately 4 times higher than that in gata2a+ runx1− cells, whereas there were no significant differences in the percentage of CFU-E and CFU-G between gata2a− runx1+ cells and gata2a+ runx1+ cells (Fig. 6b). We also calculated the absolute number of CFU-E and CFU-G based on the percentage of CFUs and the frequency of each hematopoietic subset in the kidney. Due to the higher frequency of gata2a– runx1+ cells in the kidney (Fig. 1b), the absolute number of CFU-E and CFU-G in gata2a– runx1+ cells was approximately 13.6 and 14.0 times higher than that in gata2a+ runx1+ cells, respectively (Fig. 6c). These results suggest that gata2a– runx1+ cells abundantly contain erythroid and/or myeloid-primed progenitors in the kidney. It should be noted that very large colonies were observed only in the wells plated with gata2a+ runx1+ cells (Fig. 6d), suggesting that actively proliferating MPPs may also be present in the gata2a+ runx1+ fraction.

Figure 6
figure 6

Colony-forming assays of HSPC populations. (a) Representative images of a colony formed in the presence of Epoa (CFU-E) or Gcsfb (CFU-G). (b,c) Percentage (b) and absolute number (c) of CFU-E (red bars) and CFU-G (blue bars) in gata2a+ runx1+, gata2a− runx1+, or gata2a+ runx1− cells in the kidney. Data are shown as mean ± s.e.m. *p < 0.01; **p < 0.001 by one-way ANOVA with Dunnett’s test. (d) Representative image of a large colony observed in the well plated with gata2a+ runx1+ cells in the presence of Gcsfb. The image shows cells expressing BFP. (e,f) Representative images of a colony formed by gata2a+ runx1+ cells in the presence of both Epoa and Gcsfb. Orange, red, green, and black arrowheads in (f) indicate gata2a+ runx1+, gata2a− runx1+, gata2a+ runx1−, and gata2a− runx1− cells, respectively. (g) Percent distribution of individual GFP+ mCherry+ (orange bar), GFP− mCherry+ (red bar), GFP+ mCherry− (green bar), and GFP− mCherry− (black bar) cells at day 0, 3, and 7 of culture. Bars, 20 μm (a), 100 μm (d), 5 μm (e,f).

Full size image

To clarify if gata2a:GFP or runx1:mCherry expression downregulates during erythroid/myeloid differentiation, gata2a+ runx1+ cells were cultured in the presence of both Epoa and Gcsfb and were monitored for GFP and mCherry expression. Although some cells still retained to express both GFP and mCherry (Fig. 6e), there were many mCherry single-positive and double-negative (GFP− mCherry−) cells, but only a few GFP single-positive cells within the colonies at 7 days of culture (Fig. 6f). The percentage of mCherry single-positive cells and double-negative cells increased during the culture (Fig. 6g). These results suggest that erythroid/myeloid differentiation leads to downregulation of gata2a:GFP and/or runx1:mCherry.

Taken together, our transcriptome and functional data suggest that long-term repopulating HSCs are enriched in the gata2a+ runx1+ fraction in the kidney. In addition, it is likely that common myeloid progenitors (CMPs) to erythroid-, thrombocyte-, and myeloid-primed progenitors appear to be present mainly within the gata2a− runx1+ fraction. Moreover, common lymphoid progenitors (CLPs) to T- and B-cell precursors and some myeloid progenitors may be present in the gata2a+ runx1− fraction (Fig. 7).

Figure 7
figure 7

Hematopoietic differentiation in the zebrafish kidney. Schematic diagram of hematopoietic differentiation in the zebrafish kidney is shown. The orange, red, and green area denote the phenotypic transgene expression of gata2a:GFP and runx1:mCherry. Cross-hatched area shows the putative expression of gata2a:GFP and runx1:mCherry in hematopoietic progenitors. Long-term repopulating HSCs show gata2a+ runx1+. In contrast, gata2a− runx1+ cells are mostly erythroid- and thrombocyte-primed progenitors, but also some myeloid-primed progenitors. Additionally, some myeloid-primed progenitors are gata2a+ runx1−. Lymphoid-primed progenitors/precursors may express gata2a:GFP, but not runx1:mCherry.

Full size image

Discussion

In the present study, we have established a method to isolate HSPCs from the zebrafish kidney utilizing gata2a:GFP; runx1:mCherry double transgenic animals. This new method will allow us to further investigate the molecular and cellular mechanisms underlying the regulation of HSPCs in the zebrafish kidney.

In mice and humans, hematopoietic cells and mature blood cells can be isolated by a combination of multiple antibodies against cell-surface markers. Due to the lack of antibodies in zebrafish, fluorescent transgenic lines that label specific blood cell types have instead been developed. It is currently possible to isolate various types of blood cells using these transgenic lines, such as erythrocytes (gata1:DsRed)17, thrombocytes (cd41:GFP)32, neutrophils (mpx:GFP)33, eosinophils/basophils (gata2a:GFP)28, monocytes/macrophages (mpeg1:GFP)34, T cells (lck:GFP)35, and B cells (rag2:GFP)36. To date, however, transgenic lines that label HSPCs are very limited in zebrafish. The mouse Runx1 + 23 enhancer has been shown to be driven in the HSC population throughout embryonic development, as well as in the adult bone marrow37. This stem cell enhancer is also active in zebrafish HSCs and erythroid progenitors25,38, whereas its expression does not completely recapitulate endogenous runx1 expression25,37. Because runx1:mCherry expression is restricted in hematopoietic cells, this line can be used for imaging of HSPCs not only in embryos but also in juvenile animals25,39,40. As a parallel view, myb:GFP and cd41:GFP are also widely utilized to visualize developing HSPCs in zebrafish embryos27,31, while these lines are usually combined with an endothelial mCherry line to capture nascent HSCs derived from hemogenic endothelium. While myb:GFP is not shown, cd41:GFP can also label HSCs in the adult kidney24. Cd41 is, however, expressed broadly in erythroid, myeloid, and megakaryocyte/thrombocyte lineages in both mammals and zebrafish32,41,42. Indeed, our transcriptome data also showed that cd41 (itga2b) is highly expressed in gata2a− runx1+ erythroid and/or myeloid-primed progenitors. Because of these similar expression patterns, we did not utilize a combination of cd41:GFP and runx1:mCherry to isolate HSCs. In contrast, we found that gata2a:GFP expression in the hematopoietic cell fraction was restricted mainly in the lymphoid lineage, a part of the myeloid lineage, and HSCs in the kidney. Thus, this minimum lineage overlapping between gata2a:GFP and runx1:mCherry enables HSCs to be isolated to the highest degree of purity to date.

Our competitive repopulation assays suggest that the frequency of HSCs is approximately 540 times higher in gata2a+ runx1+ cells than in KMCs. The frequency of HSCs in the zebrafish kidney has been estimated by two groups via limiting dilution transplantation assays; they showed that HSCs are present at approximately 1 in 65,500 cells43 or 1 in 38,140 cells44 within WKMCs (including erythrocytes). Since WKMCs contain approximately 55% of erythrocytes, these observations suggest that the frequency of HSCs in the gata2a+ runx1+ fraction is estimated at 1 in 32 to 55 cells. On the other hand, Tamplin et al. also estimated by limiting dilution transplantation assays that the frequency of HSCs in the runx1:mCherry+ fraction of the kidney is at approximately 1 in 35 cells25. Our data showed that the percentage of gata2a+ runx1+ cells within the runx1:mCherry+ fraction was approximately 4.6%, and that gata2a− runx1+ cells never showed long-term hematopoietic reconstitution. Based on these observations, the frequency of HSCs in the gata2a+ runx1+ fraction could also be estimated at approximately 1 in 1.6 cells. Due to the lack of inbred strains, however, contributions of donor-derived cells following transplantation largely vary in zebrafish, reflecting the difficulty in precise estimation of the HSC frequency by limiting dilution transplantation assay. As one point of view, the percentage of gata2a+ runx1+ cells in the zebrafish kidney (0.16% in KMCs) is close to that of c-kit+ Sca-1+ Lineage-marker− (KSL) cells in the murine bone marrow, which are observed at 0.1–0.2% in total nucleated cells5,45. These observations suggest that gata2a+ runx1+ cells in the zebrafish kidney may be equivalent to KSL cells in the murine bone marrow.

It has been shown in zebrafish that long-term repopulating HSCs are present in the FSClow SSClow “lymphoid” fraction in the adult kidney17. Based on our analysis, however, most gata2a+ runx1+ cells were detected in the FSCmid fraction rather than FSClow fraction, whereas both gata2a− runx1+ cells and gata2a+ runx1− cells abundantly contain FSClow cells. Ma et al. previously reported the putative HSC fraction, cd41:GFPlow SP cells, which are detected at approximately 0.02% in WKMCs24. Interestingly, these cd41:GFPlow SP cells also show slightly higher intensity of FSC than cd41:GFP− SP cells. In addition, murine HSCs in the bone marrow are larger in cell diameter compared with mature lymphocytes46. These observations suggest that HSCs in zebrafish are slightly larger in size compared with FSClow “lymphoid” cells in the kidney.

High-throughput genome-wide transcriptome analysis is now commonly used in all fields of life science research and is promising to characterize cell types along differentiation processes. Tang et al. reported single-cell transcriptomic profiling of wide variety of blood cell types isolated from the zebrafish kidney. This analysis revealed that transcriptional programs in predicted blood cell types, including HSPCs, erythroid cells, neutrophils, thrombocytes, T cells, B cells, and NK cells, are highly consistent with those in each blood cell subset reported in mouse and human, suggesting that hematopoietic differentiation programs are highly conserved amongst vertebrates38. Our data also showed that gata2a+ runx1+ cells highly expressed HSC-related genes, such as gata2b, runx1t1, kita, mpl, gfi1aa, meis1b, egr1, fgd5b, pbx1b, fosb, myca, and apoeb, of which orthologues are shown to be predominantly expressed in mammalian HSCs47,48,49,50. These data strongly suggest that regulatory mechanisms underlying self-renewal and differentiation of HSCs are also highly conserved amongst vertebrates. Furthermore, the recent advent of direct gene knockout methods in zebrafish based on the CRISPR/Cas9 system enables to rapidly screen the function of candidate genes in the embryonic to adult stage14,15,16. In combination with a gata2a:GFP and runx1:mCherry line, it is now possible to perform rapid genome-wide interrogation of gene function in HSCs using the zebrafish model. Thus, our purification strategy of HSCs in the zebrafish kidney will open new avenues to elucidate molecular cues that needed to regulate HSCs.

Methods

Zebrafish husbandry

Zebrafish strains, AB*, Tg(gata2a:GFP)la3 (ref.28), Tg(Mmu.Runx1:NLS-mCherry)cz2010 (here denoted as runx1:mCherry) (ref.25), Tg(bactin2:loxP-BFP-loxP-DsRed)sd27 (here denoted as bactin2:BFP) (ref.51), Tg(kdrl:Cre)s898 (ref.31), and Tg(bactin2:loxP-STOP-loxP-DsRed)sd5 (ref.31), were raised in a circulating aquarium system (AQUA) at 28.5 °C in a 14/10 h light/dark cycle and maintained according to standard protocols52. All experiments were performed in accordance with a protocol approved by the Committee on Animal Experimentation of Kanazawa University.

Cell preparation and flow cytometry

Kidney marrow cells (KMCs) were prepared as previously described51 with some modifications. Cells were obtained by pipetting of a dissected kidney in 1 mL of ice-cold 2% fetal bovine serum (FBS) in phosphate buffered saline (PBS) (2% FBS/PBS). After centrifugation, the pellet was gently mixed with 1 mL of distilled water by pipetting to lyse erythrocytes by osmotic shock. Subsequently, 1 mL of 2X PBS was added. Cells were then filtered through a 40-stainless mesh and washed with 2% FBS/PBS by centrifugation. Just before flow cytometric analysis, the Sytox Red (Thermo Fisher Scientific) was added at a concentration of 5 nM to exclude dead cells. Flow cytometric acquisition and cell sorting were performed on a FACS Aria III (BD Biosciences). Data analysis was performed using the Kaluza software (ver. 1.3, Beckman Coulter). The absolute number of cells was calculated by flow cytometry based on the acquisition events, maximum acquisition times, and the percentage of each cell fraction.

RNA-seq and qPCR

For sorted cells, whole-transcript amplification and double-strand cDNA synthesis was performed according to the method of Quartz-Seq53. Cells were directly sorted in a lysis buffer containing 1 μg/mL of polyinosinic-polycytidylic acid, and total RNA was extracted using RNeasy Mini Kit (Qiagen). Reverse transcription (RT) was performed using Super Script III (Thermo Fisher Scientific) and an RT primer, which contains oligo-dT, T7 promoter, and PCR target region sequences. After digestion of remaining RT primers by exonuclease I (Takara), a poly-A tail was added to the 3′ ends of the first-strand cDNAs using terminal transferase (Sigma). The second-strand DNA was then synthesized using MightyAmp DNA polymerase (Thermo Fisher Scientific) and a tagging primer, which contains oligo-dT and PCR target region sequences. PCR amplification was performed using a suppression primer, which allow to amplify small-size DNA that contains complementary sequences at both ends of the template DNA. The amplified double-strand cDNA was purified using QIAquick PCR Purification Kit (Qiagen). Library preparation was performed using Nextera XT DNA Library Preparation Kit (illumina). Next generation sequencing of cDNA libraries was performed by GENEWIZ using the Illumina NextSeq. 500 (illumina), and base-calling was performed using the Illumina RTA software (ver. 2.4.11). Sequence reads were mapped to the zebrafish reference genome (GRCz11) using HiSAT2 (version 2.1.0). Reads per million (RPM) were calculated using the Subread (ver. 1.6.4). Hierarchical clustering of each subset was performed in R (ver. 3.5.0) with the Bioconductor Heatplus package. Genes that were over two-fold enriched in gata2a+ runx1+ cells, gata2a− runx1+ cells, or gata2a+ runx1− cells were selected and used for gene ontology enrichment analysis using the Gene Ontology Resource website (http://geneontology.org). The data have been deposited in Gene Expression Omnibus (GEO) (National Center for Biotechnology Information) and are accessible through the GEO database (series accession number, GSE132927). For the kidney tissue, total RNA was extracted using RNeasy Mini Kit, and cDNA was synthesized using ReverTra Ace qPCR RT Master Mix (Toyobo). Quantitative real-time PCR (qPCR) assays were performed using TB Green Premix Ex Taq II (TaKaRa) on a ViiA 7 Real-Time PCR System according to manufacturer’s instructions (Thermo Fisher Scientific). Primers used for qPCR were listed in Supplementary Table S2. The expression of ef1a was used for normalization.

X-ray irradiation and transplantation

In zebrafish, sublethal dose of γ-irradiation (23–25 Gy) has been shown to be sufficient to ablate hematopoietic cells in the adult kidney18,43. Three to six zebrafish were placed in a 90 mm petri dish in system water, and animals were sublethally irradiated with X-ray on a Faxitron RX-650 (Faxitron, 130 kVp, 1.15 Gy/min) for 20 min (approximately 23 Gy). At 2 days post-irradiation, animals were transplanted with cells using a retro-orbital injection method43.

Colony assay

Recombinant zebrafish Epoa and Gcsfb (also known as colony stimulating factor 3b (Csf3b)) proteins were generated as previously described54. The coding region of epoa and gcsfb was amplified by PCR using kidney cDNA and primers listed in Supplementary Table S2, and ligated into the pET-16b vector. Recombinant Epoa and Gcsfb were purified from Escherichia coli using QIAexpressionist kit (Qiagen) according to the manufacturer’s procedure. For colony assays, cells were sorted into a round-bottom 96-well plate or a glass-bottom of 35 mm dish filled with the 1X ERDF condition medium containing 20% FBS, 2.5% carp serum, and 600 ng/mL of Epoa and/or Gcsfb. Cells were cultured at 30 °C, 5% CO2 for 7 days. Colonies were enumerated using an EVOS fl microscope (Thermo Fisher Scientific).

Statistical analysis

Statistical differences between groups were determined by one-way ANOVA with Dunnett’s test or Fisher’s Exact test. A value of p < 0.05 was considered to be statistically significant.