G-CSF induced hematopoietic stem cell mobilization from the embryonic hematopoietic niche does not require neutrophils and macrophages

Abstract

Hematopoietic stem cell transplantation requires a collection of hematopoietic cells from patients or stem cell donors. Granulocyte colony-stimulating factor (G-CSF) is widely used in the clinic to mobilize hematopoietic stem and progenitor cells (HSPCs) from the adult bone marrow niche into circulation, allowing a collection of HSPCs from the blood. The mechanism by which G-CSF acts to mobilize HSPCs is unclear, with some studies showing a direct stimulation of stem cells and others suggesting that myeloid cells are required. In this study, we developed a heat-inducible G-CSF transgenic zebrafish line to study HSPC mobilization in vivo. Live imaging of HSPCs after G-CSF induction revealed an increase in circulating HSPCs, demonstrating a successful HSPC mobilization. These mobilized HSPCs went on to prematurely colonize the kidney marrow, the adult zebrafish hematopoietic niche. We eliminated neutrophils or macrophages using a nitroreductase-based cell ablation system and found that G-CSF still mobilizes HSPCs from the niche. Our findings indicate that neutrophils and macrophages are not required for G-CSF-induced HSPC mobilization from the embryonic hematopoietic niche.

Granulocyte colony-stimulating factor (G-CSF) induces the production of mature neutrophils, promotes the proliferation and differentiation of hematopoietic stem and progenitor cells (HSPCs), and facilitates the release of HSPCs from the bone marrow^1,2. Its capability to release HSPCs has made G-CSF one of the pioneering cytokines used for HSPC mobilization, a process involving the release of HSPCs from the bone marrow to collect them from the peripheral blood during stem cell transplantation. To date, G-CSF remains the most widely employed mobilizing agent and can be combined with other drugs or cytokines, such as AMD3100 or GROβ, to enhance mobilization efficiency^3–5.

Neutrophils play a pivotal role in G-CSF-induced HSPC mobilization. Depletion of neutrophils using antibodies against Gr1 significantly reduces HSPC mobilization by G-CSF, demonstrating that neutrophils are required for mobilization⁶. Mature neutrophils secrete matrix metalloproteinases (MMPs) into the bone marrow, thereby creating a proteolytic environment. This proteolytic environment disrupts HSPC retention signaling molecules, such as CXCL12/CXCR4 or VCAM1, in the bone marrow niche and releases HSPCs^6–10. However, experiments conducted on mice genetically deficient in neutrophil proteases have failed to demonstrate that proteases are necessary for mobilization, as HSPCs can still be mobilized in the absence of MMPs¹¹. Although neutrophils are essential for G-CSF-induced mobilization, the precise role of neutrophil-derived MMPs in this process remains unclear.

The significance of neutrophils was soon challenged by a study demonstrating a successful HSPC mobilization from the bone marrow in transgenic mice with G-CSF receptors exclusively expressed in the monocytic lineage but not in neutrophils¹². This finding indicated that G-CSF signaling to macrophages/monocytes alone is sufficient for HSPC mobilization. Notably, endosteal macrophages play a critical role in regulating osteoblast function in the bone marrow, thus maintaining the bone marrow microenvironment¹³. Furthermore, CD169+ macrophages promote the expression of HSPC retention factors, such as Cxcl12, Vcam1, and Scf, in Nestin+ mesenchymal cells within the bone marrow¹⁴. G-CSF has been shown to deplete these endosteal macrophages, leading to a decrease in Cxcl12, Vcam1, and Scf expression¹⁵. While this suggests that macrophage depletion facilitates HSPC mobilization, macrophage activation and expansion have also been shown to promote LECT2-dependent HSC mobilization, implying more complex mechanisms behind macrophage regulation of HSPC mobilization¹⁶. Overall, the roles of neutrophils and macrophages in HSPC mobilization remain ambiguous.

The caudal hematopoietic tissue (CHT) is a specialized microenvironment that supports the developmental maturation, expansion, and differentiation of HSPCs in zebrafish^17,18. Like the adult hematopoietic niche, the CHT comprises multiple cell types, including endothelial cells, mesenchymal stromal cells, macrophages, and neural crest cells, which communicate with HSPCs through various extrinsic signals. For example, endothelial cells in the CHT secrete cytokines such as Kit ligand and oncostatin M to recruit and sustain HSPCs in the hematopoietic niche¹⁹. Additionally, CHT macrophages have been found to selectively expand or eliminate HSPCs through Calreticulin-dependent interactions²⁰. Therefore, extrinsic signals from different cell types in the CHT are crucial in regulating HSPC functions and have implications for HSPC mobilization.

In this study, we used a cell ablation technique to investigate the necessity of neutrophils and macrophages in G-CSF induced HSPC mobilization from the embryonic hematopoietic niche. To achieve this, we developed a stable transgenic line with a heat-inducible csf3b, a zebrafish ortholog of human CSF3, which translates into G-CSF. We showed that G-CSF effectively mobilizes HSPCs from the CHT. Using a nitroreductase (NTR)-based cell ablation system, we further demonstrated that G-CSF successfully mobilizes HSPCs in the absence of neutrophils or macrophages. Collectively, our findings show that neutrophils and macrophages are not necessary for G-CSF induced HSPC mobilization from the embryonic hematopoietic niche.

Results

Dynamics of HSPC egress from the CHT and colonization of the kidney

To characterize the location of HSPCs during development, we performed whole-mount in situ hybridization (WISH) of cmyb, an HSPC marker gene, at multiple time points in zebrafish embryos. Consistent with previous reports on HSPC development¹⁷, HSPCs were found to reside in the CHT from 2 – 4 days post fertilization (dpf) and subsequently appeared in the kidney at 4.5 dpf (Fig 1A). To gain a more precise and quantitative understanding of the dynamics of HSPC egress in the CHT, we live-imaged HSPCs at multiple time points using the Tg(runx1+23:mCherry) line. The number of HSPCs in the CHT was then counted with Imaris, a microscopy image analysis software. At 2 dpf, when HSPCs begin colonizing the CHT, an average of 10 runx1+23:mCherry⁺ HSPCs were observed in the CHT. This number rapidly and consistently increased to approximately 100 HSPCs during 2 – 4 dpf, a period of HSPC expansion in the embryonic niche. Notably, the number of CHT-resident HSPCs plateaued at 4 – 5 dpf and decreased afterward, suggesting that HSPC egress commences just after 4 dpf and eventually surpasses the rate of expansion (Fig 1B).

Figure 1. Dynamics of HSPC egress during development.

A) WISH for cmyb of wild-type zebrafish embryos during development. Arrowheads point to cmyb signal at the kidney. B) Number of runx1+23:mCherry⁺ HSPCs in the CHT from 2 dpf through 8 dpf. Mean +/− s.d., One way ANOVA; **p<0.01, ***p<0.001, ****p<0.0001. C) Image of circulating runx1+23:mCherry⁺ HSPCs in embryo tail at different time points. Arrowheads point at circulating HSPCs. D) The number of circulating runx1+23:mCherry⁺ HSPCs in the tail during development from 72 hpf through 120 hpf. Mean +/− s.d., One way ANOVA; ***p<0.001, ****p<0.0001.

To further characterize HSPC egress in greater detail, we recorded time-lapse images of runx1+23:mCherry⁺ HSPCs in the tail and quantified the number of HSPCs in circulation. Until 92 hours post fertilization (hpf), fewer than four HSPCs were observed in circulation at baseline at any time. Starting at 102 hpf, this number tripled to 11 circulating HSPCs, indicating that HSPC egress initiates between 92 – 102 hpf (Fig 1C and D, Movies S1 - S3). Additionally, the rate of HSPC release, as depicted by the slope of the graph in Fig 1C, also increased over time. This suggests that HSPCs steadily egress from the CHT beginning at 92 hpf, rather than all cells abruptly leaving the CHT at a specific time point. Overall, these data indicate that HSPCs gradually begin to egress from the CHT at 92 hpf and colonize the kidney marrow at 4.5 dpf during development.

G-CSF mobilizes HSPCs from the embryonic hematopoietic niche

We sought to gain insight into HSPC mobilization by expressing G-CSF in developing embryos. csf3b is a zebrafish ortholog of the human gene CSF3, which encodes for G-CSF²¹. To express G-CSF in an inducible manner, we developed a stable transgenic line Tg(hsp70l:csf3b) – hereafter referred to as the “iG-CSF fish” or “iG-CSF embryo” (Fig 2A). Expression of csf3b was validated with WISH 2 hours post heat shock, confirming that the iG-CSF embryo is a successful inducible system (Fig 2B). To validate that a functional G-CSF protein is being produced, we assessed neutrophil expansion in iG-CSF embryos 32 hours post heat shock in vivo. We observed an increase in the number of neutrophils in response to G-CSF induction (Fig 2C). These results show that iG-CSF produces functional G-CSF in response to heat shock.

Figure 2. G-CSF mobilizes HSPCs from the CHT.

A) A schematic of hsp70l:csf3b Tol2kit plasmid. B) WISH for csf3b 2 hours post heat shock. C) Neutrophils in heat shocked wild-type and iGCSF embryo at 92hpf. D) Snapshot images of runx1+23:mCherry⁺ HSPCs in the CHT of 92 hpf embryos. E) The number of circulating runx1+23:mCherry⁺ HSPCs in wild-type and iG-CSF embryos. Mean +/− s.d., Kruskal Wallis test; *p<0.05, **p<0.01. F) The number of CHT-resident runx1+23:mCherry⁺ HSPCs. Mean +/− s.d., Kruskal Wallis test; *p < 0.05

To assess whether G-CSF can induce HSPC mobilization from the CHT, we heat shocked iG-CSF at 60 hpf, a time point when HSPCs are actively colonizing the CHT. The activity of the hsp70l promoter is known to decline over time after heat-shock²². To maintain high expression of csf3b until 92 hpf, a time point right before a developmental egress begins, the iG-CSF embryos were heat-shocked again at 84 hpf. This led to a 1.7-fold increase in circulating HSPCs at 92 hpf in heat-shocked iG-CSF embryos compared to wild-type embryos (Fig. 2D and E, Movies S4 - S7). At the same time, the number of CHT-resident runx1+23:mCherry⁺ HSPCs decreased in response to G-CSF (Fig 2F). WISH also revealed a slight decrease in cmyb-expressing CHT-resident HSPCs in response to csf3b (Supplementary Figure 1). These data show that G-CSF mobilizes HSPCs from the CHT.

HSPCs mobilized from the embryonic niche prematurely colonize the adult niche

Kidney marrow blood vessels, a structure that HSPCs eventually colonize, develop as early as 2 dpf. A functional kidney equipped with glomerular filtration apparatus forms at 4 dpf.²³ This suggests that although HSPCs typically start colonizing the kidney at 4.5 dpf, the kidney may be ready to receive HSPCs as early as 4 dpf. Hence, we reasoned that HSPCs prematurely mobilized from the CHT might start colonizing the kidney marrow at 4 dpf. To test this hypothesis, we imaged the kidney of iG-CSF embryos at 92 hpf and observed HSPCs that had prematurely colonized the kidney (Fig 3A). cmyb WISH confirmed that HSPCs prematurely colonize the kidney marrow in heat shocked iG-CSF embryos (76%, n = 31/41), whereas HSPCs in embryos without G-CSF have not yet reached the kidney marrow (non-heat shocked WT: 11%, n = 4/36; heat shocked WT: 12.8%, n = 5/39; non-heat shocked iG-CSF: 7.8%, n = 3/38) (Fig 3B). These results show that mobilized HSPCs can prematurely colonize the kidney marrow.

Figure 3. G-CSF promotes early colonization of the kidney marrow without increasing the rate of HSPC proliferation.

A) Kidney marrow image of 92 hpf iG-CSF embryo with runx1+23:mCherry⁺ HSPCs. Yellow arrowhead denotes HSPCs in the kidney. Pronephric tubule with autofluorescence is indicated with a dotted line. B) WISH of cmyb for 92 hpf iG-CSF embryos (representative images). Arrowhead points to cmyb⁺ HSPCs at the kidney marrow. C) EdU staining of runx1+23:mCherry⁺ HSPCs in heat shocked wild-type and iG-CSF embryos at 92 hpf. D) The percentage of EdU-positive HSPCs in the CHT. Mean +/− s.d., Unpaired t test. E) The percentage of runx1+23:mCherry⁺ HSPCs over all live cells quantified with flow cytometry of wild-type and iG-CSF embryos. Mean +/− s.d., n = 2 replicates.

Increased circulating HSPCs and premature kidney colonization may be attributable to increased total number of HSPCs rather than mobilization. In fact, the Cyclophosphamide/G-CSF regimen promotes HSPC proliferation before mobilization, and G-CSF administration alone was shown to switch dormant HSCs into self-renewing HSCs^24–26. However, some studies suggest that G-CSF alone does not promote dormant HSCs to proliferate in a steady-state bone marrow^27,28. To test if G-CSF promoted HSPC proliferation, we performed EdU staining of iG-CSF embryos at 92 hpf. The percentage of proliferating HSPCs in the CHT was similar with or without G-CSF (Fig 3C and D), indicating that G-CSF does not promote HSPC proliferation in the embryonic niche. In addition, we compared the number of total HSPC numbers at 92 hpf through flow cytometric analysis of iG-CSF embryos crossed with Tg(runx1+23:mCherry) embryos. G-CSF did not increase the total number of runx1+23:mCherry⁺ HSPCs (Fig 3E). Thus, the induction of G-CSF at 60 – 92 hpf did not induce HSPC proliferation or increase the total number of HSPCs. Taken together, these findings indicate that the increase in circulating HSPCs and premature kidney colonization are attributable to HSPC mobilization from the CHT rather than HSPC proliferation.

Neutrophils and macrophages are not required for G-CSF-dependent HSPC mobilization from the embryonic niche

HSPCs lacking G-CSF receptors can mobilize from the bone marrow after G-CSF treatment²⁹. This suggests that G-CSF-dependent mobilization is not HSPC autonomous and involves other cell types. Neutrophils are one such candidate population. G-CSF expands and activates neutrophils, which secrete proteolytic enzymes such as MMPs into the surrounding environment. The cleavage of key retention factors such as CXCL12 and VCAM1 then leads to the mobilization of HSPCs from the bone marrow. Indeed, the induction of G-CSF in iG-CSF zebrafish embryos increased the number of neutrophils by ~3.4-fold (Fig 2C, Supplementary Figure 2A). This suggests that neutrophils play a role in G-CSF induced HSPC mobilization from the embryonic niche, similar to the adult niche.

To assess whether neutrophils are required for the G-CSF induced HSPC mobilization from the CHT, we optimized a nitroreductase (NTR) based cell ablation system. In this system, the bacterial protein NTR is expressed in target cells via a tissue-specific promoter. NTR converts Nifurpiriniol (NiF), an otherwise benign compound, into a DNA crosslinking cytotoxin (Fig 4A). We crossed the iG-CSF line with Tg(mpx:GAL4; UAS:NTR-mCherry), in which neutrophils express NTR. NiF treatment from 56 hpf to 92 hpf ablated up to ~73% of neutrophils (Supplementary Figure 2A). Even the surviving mCherry⁺ neutrophils had fewer pseudopods and were more spherical in shape, suggesting they are mostly not functional. To assess if neutrophils are necessary for G-CSF induced HSPC mobilization from the CHT, we performed cmyb WISH in neutrophil-depleted iG-CSF embryos. While 80% (n = 40/50) of iG-CSF embryos with intact neutrophils displayed premature kidney colonization at 92 hpf, we also observed premature kidney colonization in 77% (n = 36/47) of neutrophil-depleted iG-CSF embryos (Fig 4B). In summary, these data show that neutrophils are not required for G-CSF induced HSPC mobilization from the CHT and that neutrophil-independent pathways are involved.

Figure 4. G-CSF mobilizes HSPCs in the absence of neutrophils or macrophages.

A) Scheme of Nifurpirinol treatment for ablation of nitroreductase expressing neutrophils or macrophages. B) cmyb WISH of 92 hpf iG-CSF embryos with or without neutrophils. Arrowhead points to cmyb signal at the kidney. C) cmyb WISH of 92 hpf iG-CSF embryos with or without macrophages. Arrowhead points to cmyb signal at the kidney.

Macrophages are another cell type that regulates G-CSF induced mobilization of HSPCs from the bone marrow. Macrophages are known to promote the retention of HSPCs by maintaining CXCL12 levels in the bone marrow. G-CSF treatment leads to a loss of mouse bone marrow-resident monocytes/macrophages, disrupting osteoblast maintenance and ultimately resulting in HSPC mobilization¹². To assess if macrophages are involved in the G-CSF induced mobilization from the CHT, we quantified the number of macrophages after G-CSF expression. Unlike in the mouse bone marrow, the number of macrophages remained the same, if not slightly increased, after G-CSF induction. NiF treatment from 56 hpf to 92 hpf completely ablated macrophages (Supplementary Figure 2B). To test if macrophages are required for G-CSF induced HSPC mobilization from the CHT, we performed cmyb WISH in macrophage-depleted iG-CSF embryos. Ablation of macrophages in iG-CSF embryos did not prevent HSPCs from prematurely colonizing the kidney (76% of WT embryos; 78% of iG-CSF embryos), showing that HSPCs successfully mobilized in the absence of macrophages (Fig 4C). This demonstrates that macrophages are dispensable for G-CSF induced HSPC mobilization from the CHT. Overall, our results show that neutrophils and macrophages are not required for G-CSF-dependent HSPC mobilization from the embryonic niche.

Discussion

HSPCs egress from the embryonic hematopoietic niche during the fetal-to-adult hematopoietic transition. The shift in the hematopoietic niche is conserved across different organisms, from zebrafish and frogs to chickens and humans²¹. By live imaging HSPCs in developing zebrafish embryos, we characterized the dynamics of HSPC egress from the CHT during development. A rapid and constant increase in the number of CHT-resident HSPCs from 2 – 4 dpf supports the accepted concept that CHT supports HSPC expansion. We showed that HSPC egress is a gradual process, as evidenced by a sequential increase in the number of circulating HSPCs and an increase in the rate of HSPC egress.

The molecular mechanisms involved in the HSPC mobilization from the fetal niche remain elusive. Using a transgenic zebrafish with heat-inducible G-CSF expression, we showed that G-CSF mobilizes HSPCs from the embryonic niche, similar to the adult bone marrow, and increases circulating HSPCs by 1.7 fold. The effect is smaller than that observed in adult mammals, where G-CSF increases circulating HSPCs by 10~100 fold depending on the regimen scheme. While G-CSF is usually administered for at least 4 days in adult mammals before peripheral blood collection, we are limited to 1.5 days of induction of G-CSF in this study because HSPCs reside in the embryonic niche for just 2 days. We believe that the G-CSF dose accounts for the difference between adult mammals and zebrafish embryos and that increased G-CSF dosage will enhance HSPC mobilization in zebrafish embryos.

Intriguingly, mobilized HSPCs prematurely colonized the kidney marrow. These data demonstrate that G-CSF successfully mobilized HSPCs from the CHT and, more importantly, indicate that the kidney marrow is ready to support HSPCs as early as 92 hpf. The embryonic to adult niche transition is likely a gradual process where the embryonic niche fades out, and the adult niche fades in. The embryonic niche, which forms faster than the adult niche, may be a more favorable environment early in development. As more organs fully develop, the adult niche eventually becomes a more favorable environment where all HSPCs migrate. For example, the expression of adhesion factors may slowly decrease in the embryonic niche while increasing in the adult niche, making the adult niche more supportive of HSPCs. Identifying such factors that govern the embryonic-to-adult niche transition remains an interesting question.

Some studies suggest that G-CSF promotes HSPC division in the mammalian bone marrow^24–26. We did not observe an enhanced HSPC proliferation in iG-CSF embryos. This may be attributable to the difference in cell cycle states between adult and embryonic HSPCs. HSPCs in the adult niche, especially long-term HSCs, are mostly quiescent and can be switched to enter the cell cycle^26,30,31. Embryonic HSPCs, in contrast, are highly proliferative in the fetal niche³². Hence, G-CSF may not promote proliferation in already actively cycling embryonic HSPCs.

Neutrophils are one of the major regulators of G-CSF induced mobilization, yet it is unclear whether they are required for mobilization. We showed that neutrophils are not required for the release of HSPCs from the embryonic hematopoietic niche in response to G-CSF, as evidenced by premature colonization of the kidney marrow even after the ablation of neutrophils in the iG-CSF embryos. This indicates that neutrophil-independent mechanisms regulate G-CSF-induced HSPC mobilization from the CHT, which needs further investigation. One caveat of our system is that we could only partially eliminate neutrophils with the nitroreductase system. About 27% of neutrophils remain after the NTR-based ablation. This seemingly small number of neutrophils may be enough to induce the release of HSPCs from the embryonic niche. However, given the small number of neutrophils that remain and the spherical morphology of neutrophils indicative of dormancy, our conclusion that neutrophils are not required for G-CSF induced mobilization from the embryonic hematopoietic niche is justified.

Macrophages also play a role in G-CSF induced HSPC mobilization, but their necessity for mobilization remains elusive. Unlike in mice, we noticed that macrophages are not lost in response to G-CSF. This difference may be attributable to the origin of embryonic macrophages that arise from primitive hematopoiesis, while macrophages in the adult niche are mostly generated through definitive hematopoiesis. By ablating macrophages, we demonstrated that they are unnecessary for G-CSF induced HSPC mobilization from the CHT, hinting that macrophage-independent pathways modulate HSPC mobilization from the embryonic hematopoietic niche.

In conclusion, our study characterizes the dynamics of HSPC egress during development and suggests that neutrophil and macrophage-independent pathways are involved in G-CSF induced HSPC mobilization from the embryonic hematopoietic niche. G-CSF directly binds to its receptor, CSF3R, expressed in neutrophils and macrophages. scRNA-seq of zebrafish embryo³³ shows that the G-CSF receptor is also expressed on HSPCs in addition to neutrophils and macrophages, implying that HSPCs can directly respond to G-CSF³⁴ (Supplementary Figure 3A). scRNA-seq of human fetal liver³⁵ demonstrates that human fetal HSCs and MPPs, along with neutrophils, macrophages, and hepatocytes, express G-CSF receptor, showing a conservation in mammals (Supplementary Figure 3B). Notably, human fetal hepatocytes also express G-CSF receptors and may respond to G-CSF. Furthermore, in both adult zebrafish marrow³⁶ and mice marrow³⁷, most HSPCs and some long-term HSCs also express G-CSF receptor, indicating that the myeloid-independent mobilization mechanism may be conserved in the adult hematopoietic niche (Supplementary Figure 3C and D). A study that used chimeric mice containing both wild-type and G-CSFR−/− HSPCs showed that both types of HSPCs mobilize equally after G-CSF treatment in the chimeric mice, suggesting that a non-cell autonomous signal is needed for HSPC mobilization²⁹. This concept, taken together with our studies and the expression of G-CSFR on HSPCs, supports a model in which HSPCs first directly respond to G-CSF through G-CSFR and secrete cytokines to elicit changes in the hematopoietic niche, which then mobilizes stem cell through a non-cell autonomous signal. For example, cytokines secreted by HSPCs may decrease the expression of Cxcl12 or Vcam1 in the niche endothelial cells. Future research could focus on the factors that HSPCs secrete in response to G-CSF and how the niche is affected by these factors. This could lead to new strategies to increase HSPC mobilization for HSPC transplants.

Methods and Materials

Zebrafish Husbandry and Strains

Wild-type zebrafish TU, casper-EKK, and transgenic lines runx1+23:mCherry [runx1+23:NLS-mCherry], mpx:GAL4; UAS:NTR-mCherry, mpeg1:Gal4; UAS:NTR-mCherry were used in this study. Full transgene names are listed in brackets. All animals were housed at Boston Children’s Hospital and handled according to approved Institutional Animal Care and Use Committee (IACUC) of Boston Children’s Hospital protocols.

Zebrafish Embryo Heat Shock

E3 media was preheated in a 1.5 ml Eppendorf tube. 20 embryos were transferred to the pre-heated E3 media, and the Eppendorf tube was placed into a 38 C heat block for 30 minutes, tilted sideways to avoid embryos piling on top of each other.

Generation of Transgenic Animals

The hsp70l:csf3b Tol2kit plasmid was generated by first PCR amplifying csf3b cDNA from pCS2-zfCSF3b plasmid provided by Petr Bartůněk. (Fwd: CACCAAAAGCCTCCAGATCCGCAC / Rev: CCAGTTAGCATTGAGACACACTG) The PCR amplicon was then integrated into a pENTR^™/D-TOPO^™ vector using pENTR^™/D-TOPO^™ Cloning Kit (Thermo Fisher) to create pME-csf3b. Multi-site Gateway recombination (Invitrogen) was used to assemble p5E-hsp70l, pME-csf3b, p3E-polyA, and pDest-cmlc2GFP into the hsp70l:csf3b Tol2kit plasmid. 25 pg of the plasmid and 200pg of Tol2 mRNA were co-injected into one-cell stage zebrafish zygotes. Viable embryos were selected at 24 hpf, and germline transgenesis was validated through mating of the injected embryos 3 months post fertilization.

Whole-Mount RNA In Situ Hybridization (WISH)

In situ hybridization was performed using a standard protocol³⁸. cmyb probe was generated by linearizing a plasmid containing cmyb cDNA sequence. csf3b probe was generated by PCR amplification using a cDNA library pooled from 2 ~ 6 dpf embryos using specific primers (Fwd: CATTAACCCTCACTAAAGGGAAGATTTAACACTGGAGGAGCGTGT/Rev: TAATACGACTCACTATAGGGTGCGTGCAGTTAGCATTGAGA). cDNAs were then in vitro transcribed with T7 RNA polymerase (Roche) and digoxigenin-linked nucleotides.

Fluorescent Visualization of HSPCs and Image Analysis

Time-lapse microscopy was performed using a Yokogawa CSU-X1 spinning disk mounted on an inverted Nikon Eclipse Ti microscope equipped with dual Andor iXon EMCCD cameras and a climate controlled (maintained at 28.5°C) motorized x-y stage to facilitate tiling and imaging of multiple specimens simultaneously. Screening of injected constructs and imaging of WISH embryos was performed using a Nikon SMZ18 stereomicroscope equipped with a Nikon DS-Ri2 camera. All images were acquired using NIS-Elements (Nikon), blinded, and processed using Imaris (Bitplane). Embryos were mounted in 0.8% LMP agarose with tricaine (0.16 mg/ml) in glass bottom 6-well plates and covered with E3 media containing tricaine (0.16 mg/ml).

Quantification of Circulating HSPCs

For each Tg(runx1+23:mCherry) embryos, a 30 seconds long movie of their tail was taken with the spinning disk Nikon Eclipse Ti microscope described above. For each movie, the number of mCherry⁺ cells in circulation were counted at 0s, 15s, and 30s snapshot. The average of the three values were used as a representative number of circulating HSPCs for each zebrafish embryo.

Zebrafish EdU Labeling

1 nanoliter of 500μM EdU was injected into embryonic circulation via the Duct of Cuvier at 92 hpf. Embryos were kept at 4°C for 1 hour, fixed in 4% paraformaldehyde for 1 hour, permeabilized with 0.1% Triton for 20 minutes at room temperature and labeled with Alexa Fluor 647 using the Click-iT reaction (Thermo Fisher) for 30 minutes according to manufacturer instructions. Embryos were washed with PBS+0.5% Triton and blocked for 1 hour in 10% Normal Goat Serum, 0.5% Bovine Serum Albumin, 0.5% Triton. Samples were incubated in Rat anti-mCherry Alexa Fluor 594 (Invitrogen M11240, 1:200) for 1 hour at room temperature and washed 5 times with PBS+0.5% Triton.

Flow cytometry

Embryos were chopped with a razor blade in cold PBS and then incubated in Liberase (Roche) for 20 minutes at 37°C before filtering the dissociated cells through a 40 μm filter and transferring to 2% FBS. We used 3nM DRAQ-7 for live/dead stain (Abcam). Flow cytometric analysis was performed on a BD FACSAria II. Data were analyzed with FlowJo software version 10.

Nifurpirinol treatment

Nifurpirinol was added to 4 ml E3 media in 6 well plates at a final concentration of 5 uM. 20–30 embryos were raised per well and replaced with a fresh drug every 24 hours. The 6 well plate was covered with aluminum foil during drug treatment.

Statistical analysis

Graphs and statistical analyses were done with Prism (Graphpad). For all graphs, error bars indicate mean +/− standard deviation. p-values were obtained with Mann-Whitney test, One-way ANOVA followed by multiple comparison, or Kruskal-Wallis test for all analyses.

Supplementary Material

1

Supplementary Movie 1

Circulating mCherry⁺ HSPCs in the posterior half of 72 hpf embryo

2

Supplementary Movie 2

Circulating mCherry⁺ HSPCs in the posterior half of 92 hpf embryo

3

Supplementary Movie 3

Circulating mCherry⁺ HSPCs in the posterior half of 120 hpf embryo

4

Supplementary Movie 4

Circulating mCherry⁺ HSPCs in a wild-type, non-heat shocked 92hpf embryo

5

Supplementary Movie 5

Circulating mCherry⁺ HSPCs in an iG-CSF, non-heat shocked 92hpf embryo

6

Supplementary Movie 6

Circulating mCherry⁺ HSPCs in a wild-type, heat shocked 92hpf embryo

7

Supplementary Movie 7

Circulating mCherry⁺ HSPCs in an iG-CSF, heat shocked 92hpf embryo

8

Supplementary Figure 1. WISH of cmyb for 92 hpf wild-type and iG-CSF embryos at the CHT. Graph shows the phenotypic distribution of cmyb expression scored in wild-type and iG-CSF embryos.

9

Supplementary Figure 2. Ablation of neutrophils and macrophages using the NTR system. A) Representative images of neutrophils in 92 hpf wild-type and iG-CSF embryos with/without NiF treatment. Bar graph shows the number of mCherry⁺ neutrophils in each condition. B) Representative images of macrophages 92 hpf wild-type and iG-CSF embryos with/without NiF treatment. Bar graph shows the number of mCherry⁺ macrophages in each condition.

10

Supplementary Figure 3. G-CSF receptor is expressed in fetal and adult HSPCs across species. A) UMAP visualization of cells obtained from zebrafish embryo at 72 hpf and csf3r gene expression. Arrowhead points at csf3r expressing HSPCs. B) UMAP visualization of fetal liver cells obtained between 7 and 17 post conception weeks, and CSF3R gene expression. Arrowhead points at CSF3R expressing HSP_MPP cluster. C) UMAP visualization of adult zebrafish kidney marrow cells, and csf3r gene expression. D) Diffusion map of HSPCs collected from adult mouse bone marrow, and Csf3r gene expression.

Highlights.

• G-CSF mobilizes hematopoietic stem and progenitor cells (HSPCs) from the embryonic hematopoietic niche.

• Mobilized HSPCs prematurely colonize the adult hematopoietic niche.

• Neutrophils or macrophages are not required for G-CSF-induced mobilization from the embryonic hematopoietic niche.