Synergistic cooperation between progranulin and Jak2/Stat3 signaling determines definitive myeloid cell fate Authors: Abbigail McCune, Masuma Khatun Usha, Inga Baldus, CJ Willett, Noemi Ferrito, Radwa Barakat, Emilee Clemensen, Elizabeth Snella, Mark Morton, Jesus Lacal, Giulia Pavani, Clyde A Campbell, Raquel Espin-Palazon Published: 2026 ABSTRACT SUMMARYPerturbations in cell fate disrupt tissue homeostasis and drive diseases such as cancer. The JAK/STAT pathway is central to hematopoietic malignancies, yet its regulators in cell fate decisions remain poorly defined. Here, we identify progranulin (GRN) as a critical determinant of myeloid fate through the regulation of JAK2/STAT3 signaling. Using inducible stat3 and grna zebrafish models, knockout approaches, FACS, transcriptomics, and CUT&RUN, we show that Stat3 induces grna, which in turn amplifies stat3 expression, reinforcing myeloid over erythroid commitment. Lineage tracing and live imaging reveal that this Grna/Stat3 feedback operates during definitive, but not primitive, myelopoiesis. Functionally, definitive myeloid cells drive tissue repair, whereas primitive counterparts are less effective. Myeloid differentiation in a human leukemia line required GRN co-stimulation of JAK2/STAT3, highlighting its conserved role in fate determination. These findings uncover a regulatory axis controlling hematopoietic commitment and identify GRN as a potential therapeutic target in diseases with aberrant JAK2/STAT3 activation. INTRODUCTION Myeloid cells are heterogeneous in their ontogeny and arise from several distinct waves called the primitive (mesodermal-derived), transient definitive (erythromyeloid progenitor, EMP, -derived), and definitive waves (originating from hematopoietic stem and progenitor cells, HSPCs). EMPs and HSPCs are specified in the embryo from different subsets of hemogenic endothelial cells.1 While EMPs lack lymphoid potential and long-term repopulating engraftment properties, HSPCs are able to maintain the full repertoire of blood cells after birth and for the lifetime of an organism.2,3 However, EMPs maintain hematopoiesis during the fetal life.1 In addition, embryonically derived immune cells (from mesodermal precursors and EMPs) not only perdure during fetal life, but beyond birth.2 Particularly, lifelong tissue resident macrophages (TRMs), which maintain tissue homeostasis and tumor progression, seem to derive mostly from embryonic macrophages (mesodermal or EMP originated).2,4 However, genetic mapping of primitive and EMP-derived macrophages has been challenging due to the inability to separate temporally and spatially these two waves in the mouse. This, together with the inconsistent use of the terms “primitive” and “definitive” hematopoiesis, has generated confusion regarding the origin of TRM. In agreement with the most current recommendations,2 here we refer to primitive hematopoiesis as the production of blood cells from mesodermal progenitors, while the transient definitive (EMP-derived) and definitive hematopoietic programs (HSPC-derived) are of hemogenic origin. It has become clear in last years that embryonically originated macrophages orchestrate development, adult tissue homeostasis and disease progression, including tumorigenesis,3 placing these myeloid cells in the spotlight of a wide array of potential therapeutic interventions. Despite their critical role during life, little is known about the different embryonic macrophage subsets produced in the embryo. Uncovering the molecular mechanisms governing primitive and transient definitive myelopoiesis is needed to develop better tools that allow the characterization of the functional contributions of each myeloid embryonic population in tissue homeostasis and disease, a gap covered in this study. Zebrafish is advantageous for the study of the primitive and transient definitive waves of myelopoiesis because in the mouse model, both waves originate from the yolk sac and temporally overlap, making separation difficult.5 In contrast, the specification of zebrafish primitive and transient waves does not occur simultaneously and arises from two separate locations, the rostral and posterior blood islands, respectively.6,7 Progranulin (GRN) encodes a secreted precursor protein that is proteolytically cleaved into cysteine-rich granulin peptides. It is a multifunctional factor that promotes wound healing,8,9 exhibits pro- and anti-inflammatory properties,10-12 and contributes to tumorigenesis.13-15 In addition, human heterozygous GRN mutations lead to frontotemporal dementia. Despite several studies suggesting a promiscuous interaction of progranulin with several receptors such as Sortilin, Notch, and tumor necrosis factor (TNF) receptors,16-18 the mechanisms by which progranulin impacts these signaling pathways are poorly understood. We demonstrated that progranulin a (grna) was necessary for zebrafish myeloid development.19 However, its potential functional conservation in humans, the molecular mechanism, and the exact myeloid program remain unresolved and are addressed in this study. Signal transducer and activator of transcription 3 (STAT3) is a transcription factor that binds the DNA consensus sequence (TTCN2–4GAA) to modify gene transcription.20 STAT3 modulates a wide range of homeostatic functions and is often dysregulated in cancers. Most studies of STAT3 function have been conducted in cell lines, since the ablation of STAT3 results in embryonic and stem cell lethality,21,22 presenting a significant barrier to studying its role in vivo. STAT3 can drive contradictory cell responses, including B-cell proliferation, differentiation and growth arrest of monocytes, and anti-inflammatory properties of macrophages (φ), among others.23 This could be partially attributed to the specific gene target sets activated by Stat3 in different cell types.24 Conditional gene ablation approaches have revealed important functions for STAT3 in hematopoiesis, including the polarization of macrophages25 and emergency granulopoiesis.26 Nevertheless, most of STAT3 existing research has focused on either emergency myelopoiesis or abnormal myelopoiesis, limiting our understanding of STAT3 in homeostatic conditions. Whether Stat3 is required during normal granulocytic differentiation is still controversial based on mouse studies.26-28 However, it is clear that STAT3 signaling is disrupted in many cancers, including hematological malignancies. STAT3 hyperactivation is linked to numerous cancer types, including lymphoma, leukemia, pancreatic, colorectal, brain, renal, lung, ovarian, endometrial, cervical, breast, prostate cancer, glioma, head, melanoma, and squamous cell carcinoma.29 Prolonged activation of STATs in cancer cells is associated with drug resistance and poor prognosis,30 which promoted efforts to develop STAT3 inhibitors for cancer treatments.29 Despite the critical role of STAT3 in cancer progression, the development of a successful therapy that inhibits STAT3 has proven elusive, with currently no clinical drugs available that directly target STAT3.29 Uncovering mechanisms of STAT3 regulation could have the potential to improve the treatment of a plethora of hematological malignancies and solid cancers. In this study, we uncovered in vivo that in hematopoietic precursors, progranulin is a critical coregulator of the Interleukin-6/Janus Kinase 2/signal transducer and activator of transcription protein 3 (IL-6/Jak2/Stat3) pathway, resulting in the inhibition of the erythroid program and facilitating myeloid determination. Since the JAK/STAT pathway is linked to inflammation,31-33 hematological malignancies,34-36 and it is currently a therapeutic target for hematopoietic cancers,37-39 our discovery reveals a previously unrecognized target to correct JAK2/STAT3 aberrancies in hematopoietic contexts. Indeed, overexpression (OE) of human GRN, only when combined with IL-6 stimulation, was able to successfully differentiate a human erythroleukemia cell line into the myeloid lineage. In addition, the in vivo lineage tracing experiments and injury assays in this study uncovered important functional and molecular differences between these subsets of embryonically derived macrophages, resolving a long-standing gap in the field of developmental hematopoiesis and allowing for their functional characterization. While primitive macrophages failed to respond efficiently to tissue perturbances and were independent of Jak2/Stat3/Grna signaling, EMP-derived macrophages required Jak2/Stat3/Grna and were highly recruited to the wound, resulting in successful tissue regeneration. RESULTS grna is dispensable for the initial establishment of primitive myelopoiesis Developmental myelopoiesis is layered, producing ontologically different immune cells with potentially distinct functional outcomes. However, the exact molecular programs that drive the development of these ontogenically layered myeloid cells, and their presumable divergent functional outcomes, are poorly understood. We demonstrated that grna was required for embryonic and adult zebrafish myeloid differentiation,19 but it was unclear which embryonic myelopoietic program grna was driving and the molecular mechanism behind it. To address the first question, we took advantage of the timely and spatially compartmentalized zebrafish developmental myelopoietic programs (Figure 1A) and quantified the number of primitive myeloid cells at 24 h postfertilization (hpf) by whole-mount in situ hybridization (WISH) for microfibril-associated glycoprotein 4 (mfap4) within established grna mutants.19,40 The results showed normal numbers of mfap4+ macrophages at 24 hpf in grna−/− embryos (Figures 1B and 1B’), although a non-significant increased trend was noticed. In contrast, we showed previously a significant reduction of embryonic myeloid cells at 48 hpf,19 when the transient definitive EMP hematopoietic wave is already established6 (Figure 1A). These results demonstrate that grna is dispensable for the initial establishment of primitive myelopoiesis and suggest that grna could be required for EMP-derived myelopoiesis. grna is necessary for the development of EMP-derived macrophages (A) Myelopoiesis in vertebrates occurs in three waves: primitive (or mesodermal-originated myeloid cells), transient definitive (or EMP-derived myeloid cells), and definitive (HSPC-derived). These are spatially and temporally separated in the zebrafish, although there is an overlap in the persistence of the myeloid cells. The first wave initiates from the lateral plate mesoderm beginning at 10 hpf. The transient (EMP) wave of hematopoiesis begins at ~24 hpf, producing EMPs in the posterior blood island (PBI) from hemogenic endothelium cells (HE). By 48 hpf, the majority of EMPs are fully differentiated into erythroid or myeloid (macrophages and neutrophils) lineages. In the 3rd wave, definitive HSPCs are specified in the embryo from the aortic hemogenic endothelium from 1 to 2 days post-fertilization (dpf). Solid lines indicate the specification of the denoted precursors. Dotted lines indicate differentiated myeloid progeny. (B) Representative bright-field images of the head region of grna+/+ and grna−/− embryos subjected to WISH for mfap4. B’ manual quantification from B. Each dot represents the number of mfap4+ primitive macrophages in the anterior part of each embryo. Scale bars, 1 mm. ns, not significant. (C) Representative flow cytometric dot plots show Mpeg1:eGFP+; Gata1a:Dsred− primitive macrophage (green) and Mpeg1:eGFP+; Gata1a:dsRed+ EMP-derived (orange) populations in control (Grna mismatch Mo) and Grna morphant embryos. C’. Quantification from C. (D) Illustration depicts the experimental design in E. (E) Representative flow cytometric dot plots show DendraRed/Green primitive macrophage (orange) and DendraGreen EMP-derived macrophages (green) in Grna mismatch Mo control and Grna morphant embryos at 48 hpf. E’. Quantification from E. Each dot represents an experimental replicate of 10–15 pooled embryos. Horizontal lines indicate mean ± SD. EMP, Erythromyeloid progenitor. HSPC, hematopoietic stem and progenitor cells. Figure 1D created with BioRender.com. See also Figures S1 and S7. The EMP myelopoietic program is impaired in the absence of grna To address whether the decrease on embryonic myeloid cells observed during late embryogenesis19 (>36 hpf), but not early development (Figures 1B and 1B’), in the absence of grna was due to defective EMP-derived myelopoiesis, we performed lineage tracing approaches. Since EMPs co-express erythroid (gata1a) and myeloid markers,6,41 we reasoned that EMP-derived macrophages would be labeled by the dual expression of the macrophage marker Mpeg1:eGFP42 and the master regulator of erythroid development Gata1a:dsRed.43 Although gata1a downregulates after myeloid commitment, dsRed perdures for many hours after its transcription ceases,44 allowing for its detection in differentiated macrophages from its expression at the EMP stage. We hypothesized that macrophages derived from EMPs would be Mpeg+; Gata1+ in Tg(Mpeg:eGFP; Gata1:dsRed) embryos, whereas primitive macrophages would be Mpeg+; Gata1−, since gata1a does not colocalize with pu.1 in the rostral blood island.45,46 Flow cytometric analyses from homogenized pooled embryos at 36 hpf showed two mpeg1+ macrophage populations based upon the differential expression of Gata1a:dsRed (Figure 1C). We next ablated Grna with a validated Grna morpholino (Mo) and a Grna mismatch control morpholino19 in Tg(Mpeg:eGFP; Gata1:dsRed) embryos. Consistent with our hypothesis, Mpeg+; Gata1+ EMP-derived macrophages at 36 hpf were significantly reduced after Grna ablation (Figures 1C, 1C’, S1A, and S1A’). The increased percentages of Mpeg+; Gata1− primitive macrophages in the absence of Grna most likely reflected macrophage ratio shifts due to the loss of EMP-derived cells. To demonstrate the requirement for Grna during EMP myelopoiesis, we performed selective lineage tracing of primitive and EMP-derived macrophages by photoswitching Dendra2 in Tg(Mpeg1:Dendra2)47 embryos. DendraRed/Green primitive, and newly formed DendraGreen EMP-derived macrophages were then quantified at 48 hpf (Figure 1D). Consistent with our previous result, Grna depletion led to decreased Mpeg1:Dendra2Green EMP-derived macrophages, while Mpeg1:Dendra2Red/Green primitive macrophages were increased (Figures 1E and 1E’). These independent approaches demonstrate that Grna is required for EMP-derived myelopoiesis. To determine if EMPs are specified normally in the absence of grna, we quantified lmo2+; gata1a+ EMPs in Tg(lmo2:eGFP; ga-ta1a:dsRed)48 embryos as reported.6 EMP percentage was increased after Grna depletion compared to Std-Mo control embryos (Figures 2A and 2B). Additionally, Grna ablation led to a significant increase in gata1a expression within the EMP fraction (Figure 2C). Bertrand and colleagues described two lmo2+; gata1+ EMP morphological subtypes: immature myelomonocytic progenitors and immature erythroid progenitors.6 This made us hypothesize that Grna could be critical for EMP commitment toward a myeloid fate, and that grna depletion could skew these two subtypes. In agreement, cytospins from Grna-deficient fluorescence-activated cell sorting (FACS) purified lmo2+; gata1a+ EMPs at 36 hpf showed increased proportion of immature erythroid progenitors at the expense of immature myelomonocytic progenitors (Figures 2D and 2E). If Grna favored myeloid versus erythroid differentiation at the EMP level, then enforced grna expression under the vascular fli1b promoter should increase EMP myeloid commitment since EMPs specify from hemogenic endothelium.1,5,6,49 To address this, we generated a Tg(UAS:grna-P2A-RFP1)is506, hereafter referred to as Tg(UAS:grna), zebrafish line that, when crossed with Tg(fli1b:ga-l4,UAS:eGFP),50 overexpressed grna exclusively in endothelial cells, including the hemogenic endothelium (Figures S1B and S1C). As expected, WISH for the neutrophilic marker mpx and the myeloid marker mfap4 at 48 hpf showed significantly increased myeloid cells after grna OE (Figures 2F-2G’). Together, these results demonstrate that Grna is required to establish myeloid fate during the transient definitive wave of myelopoiesis. grna is critical for the establishment of the EMP-derived myeloid program (A) Representative flow cytometric dot plots show the percentage of Tg(lmo2:eGFP+; Gata1a:dsRed+) EMPs (orange) in pooled Std-Mo control and Grna morphant embryos at 36 hpf. (B) Quantification from A. Each dot represents an experimental replicate of 10–15 pooled embryos. B combines two independent experiments. (C) Histogram of gata1a expression within the EMP fraction from (A). n = 4 per experimental condition. (D) Representative cytospin images from Wright-Giemsa-stained Std-Mo control and Grna morphant FACS sorted Lmo2+; Gata1+ EMPs. Green and red asterisks represent myeloid and erythroid progenitors, respectively. Scale bars, 10 μm. (E) Quantification from D. (F and G) Representative bright-field images of control and grna overexpressing (OE) embryos subjected to WISH for mpx and mfap4, followed by F’–G’ manual quantification. Grna OE was achieved by a parental cross of Tg(14xUAS:grna-P2A-RFP1) and Tg(Flib:Gal4,UAS:eGFP). Embryos that lacked the 14xUAS:grna-P2A-RFP1 transgene were selected as controls. Each dot represents the number of mpx+ or mfap4+ cells in the caudal hematopoietic tissue (CHT) per embryo. Scale bars, 1 mm. Horizontal lines indicate mean ± SD. See also Figures S1 and S7. Transcriptomic analysis identifies defective Jak-Stat3 signaling in the absence of grna It is unclear the molecular mechanism by which progranulin drives definitive myelopoiesis. To gain mechanistic insight, we performed Panther analysis51 of differentially expressed genes in our published bulk RNAseq dataset from grna−/− and grna+/+ zebrafish kidney marrows.19 The results revealed several down-regulated pathways, including pro- and anti-inflammatory signaling (IL-4, IL-5, IL-10, IL-17, TNF), lipid metabolism, nuclear receptors, and STAT3 signaling pathway (Figure S2A). We found a multitude of STAT3-related genes dysregulated in the absence of Grna, including stat3, jak1, il6st, f2r, and socs3b (Figure S2B) (p-Adj 0.5; ≤1.5 and ≥1.5). While it is possible that these genes were differentially expressed in grna mutant marrows due to the hematopoietic lineage skew,19 a study in breast cancer and colorectal cancer cell lines showed that progranulin can enhance the transcriptional function of STAT3.14,15 This evidence prompted us to investigate if the IL-6/Jak/Stat3 pathway could be controlled by progranulin during myeloid lineage decisions. Stat3 ablation phenocopies grna deficiency To explore the role of Stat3 in developmental hematopoiesis, we first performed several Stat3 loss-of-function experiments in vivo. Stat3 knockdown using a morpholino52 reduced Stat3 protein levels in the embryo (Figure S3A) and decreased 2-fold neutrophil and macrophage numbers in the CHT at 48 hpf (Figures S3B-S3C’). Similarly, WISH analysis in previously validated stat3 mutants (stat3sa15744)53 confirmed the morphant phenotype (Figures 3A-3B’). In contrast, Stat3 depletion resulted in normal neutrophil and macrophage numbers at 24 hpf (Figures S3D-S3E’), phenocopying Grna depletion. This suggested that, similarly to Grna, Stat3 is dispensable for primitive myelopoiesis and essential during EMP myelopoiesis. We next FACS purified EMPs, neutrophils, and macrophages and quantified grna and stat3 expression levels by quantitative PCR (qPCR). As expected, both grna and stat3 transcripts were expressed by EMPs, and their expression increased in myeloid differentiated cells (Figures 3C and S3F). We next assessed EMP specification after Stat3 ablation. Similarly to Grna depletion, EMPs showed a significant increase in EMP numbers (Figures 3D and 3D’) and gata1a expression within the EMP fraction (Figure 3E) in the absence of Stat3. To address whether Stat3 drove the EMP myelopoietic program, we performed the lineage tracing experiments delineated in Figure 1D. As anticipated, Stat3 ablation resulted in decreased percentages of Mpeg1:Dendra2Green EMP-derived macrophages. However, Mpeg1:Dendra2Red/Green primitive macrophage percentages were unaffected (Figures 3F and 3G). To investigate whether Stat3 also regulates the adult HSPC-derived myelopoietic program, we isolated kidney marrow from stat3sa15744 mutant zebrafish and wild-type sibling controls at late juvenile stages and prior to the lethality associated with their systemic defects. Cytospins and Giemsa-Wright staining were performed, and cells were counted based on cytological features. The macrophage/monocyte population was significantly decreased within stat3sa15744/sa15744 mutant marrows, whereas erythroid precursors increased, though not significantly (Figure S3G). We next generated and validated a conditional constitutively active stat3 zebrafish line, Tg(UAS:ca-stat3-P2A-RFP1)is505 (Figures S4A and S4B), hereafter referred to as UAS:ca-stat3, to investigate whether Stat3 was sufficient to drive myeloid development. UAS:ca-stat3 was generated using a constitutively active Stat3 mouse transcript that triggers spontaneous Stat3 dimerization.54 ca-stat3 OE in EMPs was achieved by crossing Tg(fli1b:gal4,UAS:eGFP) with Tg(UAS:ca-stat3), resulting in a significant increase in mpx+ neutrophils (Figures 3H and 3H’). These results demonstrate that EMPs co-express grna and stat3, that both transcripts are up-regulated during EMP myeloid differentiation, and that, similarly to Grna, Stat3 is essential for EMP myeloid differentiation but dispensable for primitive myelopoiesis. EMP myelopoiesis, but not primitive myelopoiesis, is defective in the absence of stat3 (A and B) Representative bright-field WISH images of stat3+/+ and stat3sa15744/sa15744 sibling embryos. Scale bars, 1 mm. A’, B’ manual quantification from A and B, respectively. Each dot represents the number of mpx+ or mfap4+ cells in the CHT per embryo. (C) EMPs (lmo2:eGFP+; gata1:dsRed+), muscle (phldb:eGFP+), erythrocytes (hbbe1.1:eGFP+), neutrophils (mpx:eGFP+), and macrophages (mpeg1:eGFP) were purified by FACS at 48 hpf (or 36 hpf for EMPs), and RT-qPCR was performed for grna and stat3. Transcripts along the x axis are shown relative to ef1a and multiplied by 100,000. p-values: a = 0.0291, b = 0.6939 c = 0.0126, d = 0.0003, e = 0.0443, f = 0.8982, g = 0.0090, and h = 0.0402. p-values are calculated on comparison to grna or stat3 expression in EMPs. Bars represent the mean ± SEM of 3 independent samples. (D) Representative flow cytometric dot plots show lmo2+; gata1+ EMPs (orange) in Std-Mo control and Stat3 morphant embryos at 36 hpf. D’ quantification from D. Each dot represents an experimental replicate of 10–15 pooled embryos. (E) Histogram depicts gata1a expression in the EMP fraction from D. (F) Representative flow cytometric dot plots show the percentages of EMP derived macrophages (green) and primitive macrophages (orange) in Std-Mo control and Stat3 morphant Tg(mpeg1:Dendra2) embryos at 48 hpf. Photoconversion was performed at 28 hpf. (G) Quantification of EMP-derived macrophages (DendraGreen) and primitive macrophages (DendraGreen/Red) from F. Each dot represents an experimental replicate of 10–15 pooled embryos. (H) Representative bright-field images of control Tg(fli1b:Gal4+; UAS:ca-stat3−) and embryos overexpressing ca-stat3 in the endothelial compartment Tg(fli1b:Gal4+; UAS:ca-stat3+) at 48 hpf subjected to WISH for mfap4. Scale bars, 100 μm. H’ quantification from H. Each dot represents the number of mfap4+ cells per embryo. Arrowheads show marker-positive cells in the CHT. Horizontal lines indicate mean ± SD. See also Figures S3, S4, and S8; Table S1. The role of grna in myeloid differentiation is Stat3-dependent To investigate if progranulin expression could be driven directly by Stat3, we first queried whether STAT3 bound human GRN by using the regulatory feature of ensembl.org.55 We found one STAT3 binding site (BS) in the GRN 5′ regulatory region (Figure S5A). To interrogate whether Stat3 might contribute to grna expression in vivo, we assessed grna expression by WISH in Stat3 zebrafish morphants. Loss of Stat3 significantly reduced grna expression at 36 hpf (Figures 4A and 4A’). Similarly, RT-qPCR from dissected stat3 mutant kidney marrows showed a 2-fold decrease in grna expression (Figure 4B). We then utilized PROMO and TomTom56-58 to identify putative Stat3 BSs in the grna regulatory regions using the workflow we described.59 Six Stat3 BSs were identified based on similarity to known Stat3 binding motifs (Figure S5B). To verify whether Stat3 bound these sites, we utilized cleavage under targets and release using nuclease, followed by qPCR (CUT&RUN-qPCR)59 for Stat3 using a validated Stat3 antibody that recognizes the phosphorylated active Stat3 proteoform Stat3Tyr70852 All BSs, except for BS6, showed enrichment compared to a control immunoglobulin G (IgG) antibody (Figure 4C). To assess in vivo if Stat3 up-regulated grna expression, we crossed Tg(UAS:ca-stat3) with Tg(hsp70:Gal4)60 to have temporal control of ca-stat3 expression. Embryos were heat-shocked at 24 hpf to induce ca-stat3 expression and collected 7 h later. RT-qPCR for grna demonstrated that OE of ca-stat3 resulted in the up-regulation of grna in this short time frame after ca-stat3 induction (Figure 4D). In addition, fluorescent in situ hybridization (FISH) and confocal imaging were performed in these experimental settings to detect grna transcripts, together with DAPI nuclear staining to enable single-cell-resolution analysis (Figures 4E-4E”). While the number of grna-expressing cells was comparable between control and ca-stat3–overexpressing embryos (Figure 4E’), grna transcript intensity per cell was significantly increased following ca-stat3 induction (Figures 4E and 4E”). Together, these data demonstrate that Stat3 directly activates grna expression. When grna expression was enforced in Tg(hsp70:Gal4; UAS:grna) embryos, stat3 expression was significantly up-regulated after 7 h post-heat shock (Figure 4F). This made us hypothesize that progranulin could be amplifying the transcriptional activity of Stat3 in the context of myeloid cell differentiation, similarly to what was proposed in breast cancer lines.15 To assess if the effect of progranulin in myeloid differentiation was Stat3-dependent, we enforced grna expression within the endothelial and hemogenic fractions (including EMPs) by crossing Tg(fli1b:gal4,UAS:eGFP) with Tg(UAS:grna), in the absence or presence of Stat3. In Stat3-depleted embryos, grna OE failed to increase mpx+ cell numbers beyond those observed in Stat3-depleted embryos expressing endogenous WT grna levels. In contrast, grna OE in embryos with WT levels of Stat3 resulted in a significant increase in mpx+ cells when compared to control embryos (Figure 4G). We next overexpressed ca-stat3 in Grna-depleted embryos to assess if myeloid cell differentiation could be restored. ca-stat3 OE in Tg(fli1b:gal4,UAS:eGFP; UAS:ca-stat3-P2A-mRFP1) embryos partially but significantly restored the neutrophil numbers observed after Grna depletion alone (Figure 4H). These experiments suggest that the function of grna is Stat3-dependent during myeloid cell differentiation in vivo. Altogether, these results demonstrate that Stat3 directly drives grna expression, and that the effect of Grna during myeloid cell commitment is Stat3-dependent. Synergistic epistasis between Grna and Stat3 (A) Representative bright-field images of control and Stat3 Mo embryos subjected to WISH for grna, followed by A’ manual quantification. Each dot represents the number of grna+ cells per CHT. Scale bars, 100 μm. (B) grna expression measured by RT-qPCR from 43dpf stat3sa15744/sa15744 and stat3+/+ control sibling dissected kidneys. Three triplicates per condition. (C) Fold enrichment of Stat3-associated DNA fragments identified by CUT&RUN-qPCR using primers flanking the BSs denoted in Figure S4B. (D) RT-qPCR for grna transcripts in Tg(UAS:ca-stat3+; hsp70:Gal4+) and Tg(UAS:ca-stat3−; hsp70:Gal4+/−). (E) Confocal images showing grna transcripts by fluorescent in situ hybridization (FISH) and DAPI staining in the CHT of Tg(UAS:ca-stat3+; hsp70:Gal4+) and Tg(UAS:ca-stat3−; hsp70:Gal4−) sibling embryos. Cells outlined correspond to the maximum projection generated from four z stack confocal slices. Embryos were heat-shocked at 24 hpf and fixed 7 h later. Embryos were genotyped for Gal4 and mRFP1 after FISH. Scale bars, 25 μm. E’. Manual quantification of grna+ cells per CHT. E”. Mean fluorescence intensity of grna transcripts per cell. Each dot represents the average fluorescence intensity calculated from 6 to 8 grna+ cells per embryo. (F) stat3 expression in Tg(UAS:grna+; hsp70:Gal4+) and Tg(UAS:grna−; hsp70:Gal4+/−) measured via qPCR. (G) Embryos obtained from Tg(UAS:grna) × Tg(fli1b:Gal4,UAS:eGFP) crosses were injected with either Std-Mo (control) or Stat3-Mo and subjected to WISH at 48 hpf to assess the expression of mpx. (H) Embryos obtained from Tg(UAS:ca-stat3) × Tg(fli1b:Gal4,UAS:eGFP) crosses were injected with Std-Mo (control) or Grna-Mo and subjected to WISH at 48 hpf to assess mpx expression. Each dot represents the number of mpx+ cells in the CHT per individual embryo. Horizontal lines indicate mean ± SD. See also Figure S5 and Table S1. Jak2 activates Stat3 signaling during definitive myelopoiesis Janus Kinases are one of the main effectors that activate STATs.61 We decided to explore whether Jak2a loss of function phenocopied Stat3 and Grna deficiencies, since JAK2 is the canonical kinase responsible for STAT3 activation.62-64 As shown in Figures 5A and 5A’, Jak2a knockdown with a validated Mo65 led to decreased mpx and mfap4 numbers at 48 hpf, phenocopying Stat3 and Grna19 deficiencies. To interrogate whether myeloid numbers could be restored in the absence of Jak2a by the hyperactivated form of Stat3, we overexpressed ca-stat3 within the endothelial compartment using our Tg(UAS:ca-stat3) crossed with Tg(fli1b:Gal4, UAS:GFP). As expected, neutrophil numbers were restored to almost control levels (Figures 5B and 5B’). Overall, these data demonstrate that Jak2a acts upstream of the Stat3-Grna axis to drive definitive myelopoiesis. Jak2a is necessary for myeloid development (A) Representative bright-field images of Std-Mo control or Jak2a-Mo embryos at 48 hpf subjected to WISH for mpx and mfap4. Scale bars, 2.5 mm. A’. Quantification from A. (B) Representative bright-field images of Tg(UAS:ca-stat3; fli1b:Gal4) embryos at 48 hpf subjected to WISH for mpx. Embryos were also injected with Std control or Jak2a morpholinos. Scale bars, 1 mm. B’. Quantification from B. Each dot represents the number of mpx+ cells/CHT. Horizontal lines indicate mean ± SD. See also Figure S4 and Table S1. Definitive, but not primitive, macrophages respond to injury and promote tissue regeneration We demonstrated that grna−/− zebrafish had impaired ability to heal wounds due to reduced embryonic myelopoiesis.19 As loss of stat3 expression resulted in decreased EMP-derived myeloid cells, but not primitive myeloid cells, similarly to what we have demonstrated for grna,19 we hypothesized that stat3 mutant embryos would also have impaired regeneration abilities. We performed tail resection on increased stat3sa15744/+ mutants and imaged their healing progress (Figure S6A). stat3 mutants showed reduced regenerated area compared to their wildtype siblings at 48 h post-wounding (hpw) and 72 hpw (Figures 6A-6A”’). Since neutrophils and macrophages are critical during wound healing and regeneration66-69 and grna and stat3 mutants successfully specified primitive myeloid cells, but had reduced numbers of EMP-derived myeloid cells, we hypothesized that EMP-derived myeloid cells would be preferentially recruited to the wound, potentially being one of the main embryonic myeloid effectors of regeneration. To test this hypothesis, we photoconverted primitive macrophages by photoswitching Tg(Mpeg1:Dendra2) embryos, and quantified primitive (Dendra2Red/Green) and EMP-derived macrophages (Dendra2Green) recruited to the wound and those not responding to injury (residing within the CHT) (Figure 6B). As shown in Figures 6C-6C”, while ~51% (55% at 6hpw, and 47% at 24hpw) of EMP-derived macrophages were recruited to the wounded area, only ~27.5% (26% at 6hpw and 29% at 24hpw) of primitive macrophages were mobilized to the resected tail. We performed an additional tissue wounding assay in Tg(Mpeg1.1:eGFP) at 24 hpf, when macrophages are of primitive origin, to confirm their reduced capabilities to be recruited. In agreement to Figures 6C-6C”, only ~20% of primitive macrophages were recruited to the wound (Figures S6B-S6C’). These experiments demonstrate that primitive and EMP macrophages are functionally distinct, and that primitive macrophages recruit to the wound with low efficiency, while EMP-derived macrophages are the main myeloid effectors of tissue regeneration. Definitive macrophages efficiently respond to injury and promote tissue regeneration (A) Representative bright-field images of stat3+/+ and stat3sa15744/sa15744 mutant tails at the indicated hours post-wounding (hpw). Lines indicate the site of tail amputation. Scale bars, 50 μm. A’, A”, A”’ Quantification of regenerated area at the indicated times. Each dot represents the regenerated area per tail. Horizontal lines indicate mean ± SD. (B) Diagram of the workflow used in (C). Representative confocal images of photoconverted Tg(Mpeg1:Dendra2) embryonic tails, following fin resection, at the indicated hpw. EMP-derived macrophages are green. Primitive macrophages are yellow (orange arrows). All red cells were categorized as primitive. Scale bars, 100 μm. C’, C” Quantifications from C. See also Figure S6. GRN drives human myeloid commitment To assess the effects of progranulin (GRN) in a human system, we transduced the erythroleukemic K562 line with lentiviral vectors containing a GRN transgene and a hygromycin-resistance cassette (Figures 7A and S7B). K562 cells express low level of GRN,70-72 and can be further differentiated into early precursors of erythrocytes, granulocytes, and megakaryocytes, making this line a good model to study the impact of progranulin on hematopoietic cell lineage decisions.73 We hypothesized that GRN OE should affect lineage marker expression in K562. After viral transduction and hygromycin selection, we generated stable cell pools of GRN and mock-transduced K562. Compared to wildtype K562 and mock-treated cells, GRN-K562 showed a small but significant increase in the baseline expression of CD45, a known leukocyte marker (Figures 7B and 7C, untreated samples, UT). No significant changes were observed in baseline myeloid (CD11b, CD14, CD16), megakaryocytic (CD61), or erythroid (CD235ab and CD71) markers (Figure S7A). We hypothesized that, similarly to zebrafish, GRN-mediated effects on myeloid differentiation may require the activation of the JAK2/STAT3 axis. We propose that the co-activation of this pathway—through IL-6 stimulation—in combination with GRN OE would potentiate myeloid commitment. K562 cells are responsive to IL-6 and express both JAK2 and STAT3,74 making them a suitable model to test this interaction. We next stimulated transduced K562 with phorbol 12-mystate 13-acetate (PMA) as a control, or Interleukin-6 (IL6) to activate the JAK2/STAT3 axis. PMA has been used to induce monocytic/megakaryocytic differentiation in K562 through a STAT3-independent pathway (MEK/MAPK and NF-kB).64,75,76 PMA stimulation increased the expression of megakaryocytic and myeloid markers and reduced the expression of erythroid markers in WT K562 (Figure S7C), as shown.64,75,76 However, no significant differences were observed across the GRN and mock transduced pools. Only a mild increase in CD16 expression was observed in GRN-K562s after a 24-h induction with PMA (Figure S7D). IL-6 activates JAK/STAT3 by promoting the phosphorylation of STAT3-Y70577 and downregulates erythroid genes in K562.78 After 24 h of IL-6 stimulation, a significant increase in the macrophage marker CD11b and the percentage of CD11b-expressing K562 cells was observed in GRN-expressing K562 compared to mock and wild-type K562 controls, despite the failure of IL-6 treatment to increase CD11b in the absence of GRN OE (Figures 7D and 7E). No other significant differences were observed in erythroid, megakaryocytic, and myeloid markers (Figure S7). These results demonstrate a specific role for progranulin in promoting human monocyte commitment in response to IL-6 stimulation and validate the model established in vivo in zebrafish in this work (Figure 7F). Role of progranulin in human myeloid commitment (A) Illustration of the lentiviral vectors used to transduce K562. LTR, long terminal repeat; CMV, cytomegalovirus promoter; IRES, internal ribosome entry site; Hygro: Hygromycin resistance gene. (B) Representative flow cytometry histogram plots of CD45 expression. Gray (negative): Unstained cells. Green: GRN-K562. Orange (Mock): K562 transduced with Hygromycin control vector. Black (WT): untransduced K562. (C) CD45+ cell percentage normalized to WT K562 at baseline (UT, untreated), and after PMA or IL6 stimulation (*p < 0.05, ***p < 0.001; two-way ANOVA test, Sidak correction). ns, not significant. (D) Representative flow cytometric histogram plots of CD11b after IL-6 stimulation. (E) Bar plot of the percentage of CD11b+ cells normalized to WT K562 at baseline (UT, untreated), and after stimulation with PMA or IL-6 (*p < 0.05; two-way ANOVA test, Sidak correction) assessed by flow cytometry. ns, not significant. (F) Model of the synergistic role of progranulin and Jak2/Stat3 during myelopoiesis. Primitive myelopoiesis operates independently of the Stat3-Progranulin signaling pathway (left). Conversely, EMP-derived myelopoiesis depends on Jak2-Stat3-Progranulin signaling (right), where progranulin enhances Stat3 activity. The Jak2/Stat3/Progranulin axis is critical to drive definitive myeloid cell differentiation, and EMP-derived myeloid cells recruit efficiently to injured tissues, driving tissue repair. See also Figure S7 and Table S2. DISCUSSION Here, we asked whether macrophages originating from mesoderm (primitive) and from the transient embryonic definitive EMPs had distinct functional and molecular properties. The restricted expression of the zebrafish ohnologue grna to the hematopoietic cell compartment,19 and the spatial and temporal separation of the primitive and transient definitive zebrafish embryonic myelopoiesis79 allowed the resolution in vivo of the molecular and functional properties that distinguish these embryonic myelopoietic programs. We demonstrate that Grna enhances the transcriptional activity of Jak2-activated Stat3 to drive the definitive myelopoietic program, and that primitive myelopoiesis was independent of this mechanism. Our results reveal that each developmental myelopoietic program has unique molecular identities and divergent functional outcomes. In addition, we show that critical pathways for cell fate decision need to be modified by molecular regulators such as progranulin to effectively promote lineage fate determination, a mechanism highly conserved among vertebrates. It was assumed that macrophages were generated prior to hematopoietic stem cells (HSCs) to protect the embryo against potential perturbations before HSCs are ready to differentiate, which only occurs around the time of birth in mammals or ~14 dpf in zebrafish.80,81 Interestingly, our in vivo lineage-tracing resolved the unique functional characteristics of each embryonic macrophage population, both in the context of tissue disruption and normal development. We show that primitive macrophages fail to respond efficiently to tissue perturbances, while EMP-derived macrophages are efficiently recruited to the wound site. In agreement, grna19 and stat3 (this work and82,83) mutant animals were unable to properly regenerate the wounded tissue due to the lack of EMP-derived macrophages, while having normal numbers of primitive macrophages. While there is substantial evidence to support the presence of primitive and EMP-derived macrophages in zebrafish embryos, the existence of primitive neutrophils is more controversial,84,85,86,87 which may explain why the effect of Grna on neutrophils is more apparent. It is important to notice, however, that these mutants also lacked EMP-derived neutrophils, which could also contribute to the resolution of inflammation and accelerate tissue repair.88 A recent study concluded that primitive hematopoiesis sustained most embryonic erythromyeloid lineages, and that macrophages derived from the definitive wave did not contribute substantially to tail fin regeneration.89 While seemingly contradictory to our data, the authors did not distinguish between primitive and the transient definitive EMP-derived macrophages and defined both populations as “primitive macrophages.” In contrast, the lineage tracing experiments presented here precisely distinguished these two embryonic myelopoietic programs, resolving a long-standing question in the field. Adult, or HSPC-derived, hematopoiesis occurs in the mammalian bone marrow and the teleost kidney marrow. Therefore, blood lineage assessment in these vertebrate organs addresses HSPC differentiation potential. Global Stat3 knockout in mice is embryonic lethal,21 and Stat3 ablation in HSPCs only extended their survival to a few weeks after birth,90 limiting the in vivo study of Stat3 during hematopoiesis. Although most zebrafish stat3 mutants prematurely die before reaching adult-hood,53,91 we were able to analyze the kidney marrows of stat3−/− late juvenile survivals, allowing a comprehensive assessment of HSPC differentiation in vivo in the adult stat3−/− marrow. Wright-Giemsa stains of cytospun stat3−/− marrows and pathologist examination revealed decreased myeloid differentiation and increased erythroid differentiation, mimicking the phenotype observed at embryonic stages, when EMP differentiation was interrogated. The fact that, in contrast to the primitive hematopoietic program, both the definitive and the transient definitive hematopoietic programs (HSPC and EMP, respectively) were impaired in the absence of stat3 (this work) and grna,19 implies that EMPs and HSPCs use the Stat3-Grna program as a shared molecular mechanism of myeloid differentiation, suggesting a global characteristic of hemogenic endothelium-originated myeloid programs. This could have a positive impact on future in vivo drug screenings, since screenings performed at embryonic stages could reveal candidates that could be targeted to correct adult myeloid aberrancies. Because many of these mutations (Jak2,92 Stat3,21 and gp13093) are lethal, analyzing hematopoietic differentiation at the EMP stage can provide key mechanistic insights into adult hematopoiesis, as shown here. Around 200 × 109 erythrocytes are produced daily in humans,94 being the most abundant cell type in the human body.95 The balance between erythroid and myeloid factors decides the differentiation of hematopoietic progenitors into each blood lineage, with Gata1 acting as the master regulator of erythroid differentiation.96 The increase on gata1a expression observed in the context of stat3 deficiency mirrored our results in grna depleted animals.19 It could be speculated that the default progression of hematopoietic progenitors could be toward the erythroid lineage to sustain the high demand of erythrocytes required, and that only active repression of gata1 by the Stat3/Grna program in hematopoietic progenitors would deviate their differentiation toward the myeloid lineages. In agreement, studies in mice and zebrafish have demonstrated that Gata1 and Stat3 antagonize each other.97,98 Our data support a model in which progranulin is a critical factor to enhance Stat3 activity to outbalance erythroid fate and promote myelopoiesis (Figure 7F). We have therefore resolved the mechanism by which gata1a is inhibited by Grna, a question that remained open in our previous study.19 Through the generation of the transgenic zebrafish lines Tg(UAS:grna) and Tg(UAS:ca-stat3), we were able to precisely control temporally and spatially the OE of grna and ca-stat3, respectively, and assess myeloid development and gene expression. These experiments, together with CUT&RUN-qPCR for Stat3, facilitated the identification in vivo of the synergistic cooperation among Grna and Stat3. Since it has been demonstrated that Stat3 activates its own expression,99 and we showed here that Stat3 directly targets grna transcription, this could result in a positive feedback amplification model that rapidly outcompetes erythroid factors and drives the commitment of hematopoietic progenitors toward myeloid fate (Figure 7F). Our in vivo data agree with previous studies in breast cancer cell lines showing that progranulin increased STAT3 phosphorylation and nuclear translocation.15 Together, this suggests that the progranulin/Stat3 feedback loop could be a universal mechanism of Stat3 control. Although the subcellular role of progranulin is not fully understood, lipidomic studies have revealed that progranulin might be a critical effector of cell lipid composition and metabolism.12,100,101 On the other hand, STAT3 is hyperactivated when subjected to cycles of palmitoylation-depalmitoylation.102 It would be insightful to test whether progranulin indirectly modify Stat3 function through the alteration of the cell lipid land-scape. Regardless, the molecular mechanism revealed here uncovered progranulin as a previously un-reported target to modulate Jak/Stat3 activity during hematopoiesis. Moreover, we show that the Jak2/Stat3/Grna axis is also crucial for neutrophil development, impacting granulopoiesis, an effect that was unclear when using Stat3 knockout mice models in previous studies. This has fundamental implications in hematological disease, since Stat3 is involved in lymphocytic, myeloid leukemias, multiple myelomas, and non-Hodkin lymphoma.29 JAKs are the main activators of STAT3, often driven by IL-6.78 Our transcriptomic data from grna−/− marrows showed decreased il6st (also known as glycoprotein 130, or gp130) expression. It is well known that the IL-6/IL-6Rα/gp-130 complex activates JAK/STAT3. Our human in vitro data showing that myeloid cell fate is achieved only when IL-6 stimulation is combined with GRN OE supports the idea that IL-6 is at the apex of the JAK/STAT3/GRN activation axis in the context of definitive myeloid development. This has significant implications for cancer, since IL-6 is the most important regulator of the JAK/STAT3 pathway in tumors. Because this pathway is hyperactivated in hematologic malignancies alongside reduced GRN expression, boosting progranulin may restore myeloid differentiation and help eliminate leukemic clones. Limitations of the study While our study establishes progranulin as a regulator of JAK2/STAT3 signaling that drives definitive myeloid fate across two vertebrate systems, several limitations remain. First, the incomplete rescue of myeloid cells by hyperactivated Stat3 in grna morphants suggests that progranulin may do more than amplify STAT3; it may refine target gene selection. Determining whether progranulin reshapes the STAT3 transcriptional program will require further study. Second, although primitive macrophage specification appears normal without grna, its expression in these cells suggests potential roles in their maintenance or function. Finally, the precise mechanism by which progranulin modulates STAT3 activity remains unresolved. STAR★METHODS EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS Ethics declarations and approval for animal experiments The zebrafish research in this study was performed according to the Guidelines for Ethical Conduct in the Care and Use of Animals.108 All zebrafish experiments were performed according to Iowa State University Animal Care and Use Committee IACUC-20-025 and IACUC-20-024 approved protocols, and in compliance with ARRIVE guidelines, the American Veterinary Medical Association (2020), and NIH guidelines for the humane use of animals in research guidelines.109,110 The age/developmental stage of animals is noted in the figure or legend when applicable. Sex was not included as a biological variable because sex determination in zebrafish is not complete at the embryonic and juvenile stages examined in this study. Zebrafish husbandry Wild-type AB* and transgenic zebrafish (Danio rerio) embryos and adults were mated, staged, raised, and processed following the procedures from the Zebrafish Book.111 Adult fish were maintained in a circulating aquarium system at 27°C on a 14 hr light/10 h dark cycle. Most experiments were carried out with embryo or larvae stages. The following zebrafish lines were utilized: grna54a 40, stat3sa15744,112 Tg(gata1:dsRed),43 Tg(mpeg1:eGFP),42 Tg(lmo2:eGFP),48 Tg(hsp70:gal4),60 fli1rstpl50GT 5550 (denoted as Tg(fli1b:ga-l4,UAS:eGFP) throughout the manuscript), Tg(14xUAS:grna-P2A-mRFP1)is505 (this work), Tg(14xUAS:stat3C-P2A-mRFP1)is506 (this work), Tg(phldb1:gal4-mCherry)104, Tg(mpx:eGFP)i1141,105 and Tg(mpeg1:Dendra2).47 Genotyping of stat3sa15774 was carried out using TaqMan Real-Time PCR Assay (ThermoFisher). Cell culture K562 cells (ATCC CCL-243), isolated lymphoblasts from a 53-year old female with myelogenous leukemia, were maintained in RPMI 1640 medium containing 2 mM glutamine and supplemented with 10% fetal bovine serum (FBS, Lonza, Basel, Switzerland) 10 mM HEPES, 1 mM sodium pyruvate and penicillin and streptomycin (100U/ml each; Gibco, Waltham, MA, USA). HEK293T cells (ATCC, CRL-3216), epithelial-like cell isolated from the kidney of a fetus, were maintained in DMEM/F-12 GlutaMAX (Gibco, #A4192001), supplemented with 10% fetal bovine serum (FBS; NeoBiotech RNase A powder, REF #NB-03-0163), at 37°C in a humidified incubator with 5% CO2. Large stocks of the cell lines were generated after receipt. Authentication was performed by ATCC prior to distribution. After thawing and every four passages thereafter, K562 cell identity was monitored by assessing marker expression, including CD235a and CD71, and by evaluating cellular morphology. Cells were routinely tested for mycoplasma contamination by PCR every four passages and consistently tested negative. METHOD DETAILS Generation of Tg(Xla.Cry:eGFP,14xUAS:grna-P2A-mRFP)is506 zebrafish The grna zebrafish cDNA was amplified by RT-qPCR from RNA isolated from adult AB* WT zebrafish kidney marrow using the primers specified in Table S1. The forward and reverse primers incorporated EcoRI and XhoI restriction enzyme sites, respectively. Five-hundred nanograms RNA template was used for reverse transcription with Superscript III (Invitrogen, 18080051). Amplified grna was directionally cloned into pCS2+113. To build the Tol2 (Xla.Cry:eGFP,14xUAS:grna-P2A-mRFP) construct, grna was amplified from a previous grna-p2a plasmid generated in lab by PCR using Q5 High-Fidelity DNA Polymerase (NEB #M0491S) and the primers denoted in Table S1. The P2A sequence was as follows (GCTACTAACTTCAGCCTGCTGAAGCAGGCTGGAGACG TGGAGGAGAACCCTGGACCT). The grna-P2A sequence was also amplified using the primers denoted in Table S1 as P2A (for UAS:grna-P2A-mRFP generation). Amplified grna-P2A sequence was cloned into the transposon vector Tol2(14xUAS:mRPF, gcry1:eGFP) (pDB790).114 The plasmid backbone was linearized with Kpn1 (NEB R3142S). The NEBuilder HiFi DNA Assembly Cloning Kit (NEB #E5520) was utilized to subclone the PCR-amplified grna-P2A (4:1 ratio insert:vector) into the UAS backbone. Tol2 transposase mRNA was in vitro transcribed from 1 μg linear pT3TS-Tol2 plasmid115 to generate capped, polyadenylated mRNA using a mMessage SP6 Transcription Kit (Thermo Fisher Scientific, AM 1340). Tol2 mRNA was purified using a GeneJET RNA Cleanup and Concentration Micro Kit (Thermo Fisher Scientific, K0841) and eluted in RNase-free water. A total volume of 520 pL containing 80 ng/μL transposase mRNA and 100 ng/μL of the correspondent plasmid DNA was injected into one-cell stage AB* embryos using a PLI-90A pico-injector Warner instrument. Tg(Xla.Cry:eGFP,14xUAS:grna-P2A-mRFP) F0 adults were screened at 48 hpf to identify two independent founders transmitting Tol2 integration through the germline based on GFP gamma crystalline lens marker expression. Individual F1 adults were used to generate F2 fish and assess grna expression by qPCR. A single F2 adult showing Mendelian segregation of transmitted alleles and the highest grna expression was outcrossed to AB* to establish the independent line Tg(Xla.Cry:eGFP,14xUAS:grna-P2A-mRFP)is506, referred to in text as Tg(UAS:grna). Generation of Tg(Xla.Cry:eGFP,14xUAS:stat3C-P2A-mRFP) is505 zebrafish The McGrail lab gifted the plasmid backbone, pTol2<14XUAS, gcry1:GFP, but it was initially published with the vector designation pDB790.114 The pTol2 backbone was purified using Monarch PCR & DNA Cleanup Kit (T1030L) following the manufacturer recommendations. The constitutively active stat3 (ca-stat3) sequence was cloned from Stat3-C Flag pRc/CMV (Addgene plasmid #8722; https://www.addgene.org/8722/; RRID:Addgene_8722)54 using the primers listed as “ca-stat3” Table S1 (65°C annealing, 18 cycles). The stat3 sequence is rendered constitutively active by substitution of two cysteine residues (substituting residues A661 and N663) within the C-terminal loop of the SH2 domain. A P2A sequence (GCTACTAAC TTCAGCCTGCTGAAGCAGGCTGGAGACGTGGAGGAGAACCCTGGACCT) was cloned into pTol2<14XUAS, gcry:GFP using the primers denoted as “P2A (for ca-stat3)” (Table S1) (72°C annealing, 18 cycles). All cloned sequences were amplified by PCR using Q5 High-Fidelity DNA Polymerase (NEB #M0491S). The plasmid backbone was linearized with Kpn1 (NEB R3142S). The NEBuilder HiFi DNA Assembly Cloning Kit (NEB #E5520) was utilized to subclone the PCR-amplified ca-stat3 with the P2A sequence (10:3:1 ratio insert P2A:insert-ca stat3:vector) into the UAS backbone. A total volume of 520 pL containing 80 ng/μL transposase mRNA and 100 ng/μL of the correspondent plasmid DNA was injected into the one-cell stage AB* embryos using a PLI-90A pico-injector Warner instrument. At 48 hpf, embryos were dechorionated by hand, sorted based on the GFP γ-crystalline (referred to as γ-crys throughout paper) lens marker expression, and raised to adulthood. Potential founders were outcrossed with AB* to identify transmitters. We identified founders that exhibited increased stat3 expression based on qPCR data. Subsequent F1-F2 progeny were used for the experiment outlined in this work. Tg(Xla.Cry:eGFP,14xUAS:stat3C-P2A-mRFP) is505 is referred to in text as Tg(UAS:ca-stat3). Morpholino injection Antisense targeting morpholinos (GeneTools, LLC) were resuspended in water to obtain a 2 mM stock solution. Morpholinos were diluted with nuclease-free water (Fisher Scientific, BP2484100) and phenol red solution (Sigma, P0290) to 0.4mM (Standard control), 0.4mM (jak1), 0.6mM (jak2a), 0.4mM (stat3), 0.4mM (grna mismatch control), and 0.4mM (grna). Morpholinos were injected into the yolk ball of 1-cell stage embryos with a 500pL drop size using a PLI-90A pico-injector Warner instrument. Morpholino sequence information is located in Table S1. Gene overexpression by mRNA injection ca-stat3 mRNA was prepared using the Stat3-C Flag pRc/CMV plasmid purchased from snape gene (#8722)54 and mMESSAGE mMACHINE T7 ULTRA Transcription Kit (Invitrogen: AM1345). The Stat3C transcript contains two-point mutations replacing an asparagine and alanine residue with cysteine residues, resulting in Stat3 spontaneous homodimerization without undergoing phosphorylation.54 mRNA was diluted to 200ng/μL with nuclease-free water (Fisher Scientific, BP2484100) and phenol red solution (Sigma, P0290), and 500pL injected at one-cell stage. Whole mount in situ hybridization (WISH) Whole mount in situ hybridization (WISH) was carried out using procedures from 116. Probes for the grna, mfap4, and mpx transcripts were generated using the DIG RNA Labeling Kit (Roche Applied Science, 11277073910) as per manufacturer instruction from linearized plasmids. All probes were used at a concentration of 0.5ng/μL. Anti-Digoxigenin-AP Fab fragments (11093274910) and NBT/BCIP solution (11681451001) were used for probe labeling and detection. Animals were analyzed using a Leica M165FC stereomicroscope equipped with a DFC295 color digital camera (Leica) and FireCam software (Leica), and the number of labeled cells manually counted. stat3sa15744 genotyping by TaqMan SNP assay stat3sa15744/+ heterozygous zebrafish were incrossed to generate WT, mutant, and heterozygous individuals, and the progeny was subjected to WISH and subsequent genotyping. After WISH, embryos were imaged in Phosphate Buffered Saline with 0.1% Tween 20 (PBT) and individually genotyped using a Custom TaqMan SNP Genotyping Assay (Thermo Fisher, 4332075, Assay ID: ANGZY9Y) in a CFX Connect Real-Time PCR System (BioRad CFX Maestro 2.0, v5.0.021.0616). A qPCR program was used with an initial denaturing temperature of 94°C for 9 min. This was followed by 38 cycles of the following 3 steps: denaturing at 96°C for 20 s, annealing at 60°C for 1 min, and extension at 72°C for 30 s. Individual stat3+/+, stat3sa15744 /+, stat3sa15744 /sa15744 embryos were clustered using allelic discrimination. Fluorescence in situ hybridization (FISH) and fluorescence intensity quantification Fluorescence in situ hybridization (FISH) was carried out as previously described with the following two adjustments.117 1) Probe washes were performed at 65°C as follows: 15 min in 75% HB/2xSSC; 15 min in 50% HB/2xSSC; 15 min in 25% HB/2xSSC. Hybridization buffer (HB) contains 50% Formamide, 5x SSC, 0.1% Tween 20, 9 mM citric acid, 5mg/ml Torula RNA, 50 μg/ml heparin. 2) Embryos were not incubated overnight with glycerol, but rather imaged directly. Antisense mRNA probe for grna transcripts was generated from linearized plasmids using the DIG RNA Labeling Kit (Roche Applied Science, 11277073910) as per manufacturer instruction and used at a concentration of 0.5 ng/μL. grna antisense probe was labeled and detected with anti-DIG POD (Sigma, 11207733910) antibody and TSA Plus Cy3 Solution (Fisher Scientific, 509049411). Following the detection of grna transcripts by FISH, we performed DAPI staining to mark cell nuclei for 15 min in 1:1000 dilution in PBT from stock solution of 5mg/mL. After adding Plus Cy3 TSA, all samples were protected from light exposure until the commencement of imaging. Images were taken at 40X magnification using a Stellaris 5 confocal microscope with LASX software (Leica, v4.8.1.29271). Fluorescence intensity of grna transcripts was quantified using LAS X software (Leica, v4.8.1.29271; Quantify application). For each embryo, 6–8 grna+ cells in the CHT were manually selected as region of interests (ROIs) from maximum projections of four confocal z-stacks based on grna FISH and DAPI staining. Mean and integrated fluorescence intensities were recorded, and each plotted data point represents the average intensity of 6–8 cells per embryo. Flow cytometry and FACS analysis Embryo dissociation and homogenization was performed as previously reported.59,118 Briefly, embryos were dechorionated with pronase (MilliporeSigma; 11459643001) and dissociated using 500μL of 50 μg/mL liberase (MilliporeSigma 5401127001) in Dulbecco’s Phosphate Buffered Saline solution with Ca2+ and Mg2+ (MilliporeSigma, D8662). The resulting suspension was then filtered through a 30 μm nylon mesh (Fisher Scientific; NC9084441) using a 21-gauge syringe (Fisher Scientific; BD 309624), and the liberase solution inactivated post-filtration with cold PBS. Cells were collected by centrifugation at 300rcf for 7min at 4°C. Supernatant was removed via gentle pipetting. For live flow cytometric acquisitions, SYTOX Red (Life Technologies; S34859) was added to exclude dead cells. For FACS analysis, cells were sorted into 350μL ice-cold RTL buffer (Qiagen, 79216). Populations were sorted based on fluorescent markers at 48 hpf unless specified otherwise: muscle Tg(phldb1:gal4-mCherry), neutrophils Tg(mpx:eGFP), macrophages Tg(mpe-g1:eGFP), EMPs Tg(lmo2:eGFP, gata1:dsRed) (36 hpf), erythrocytes Tg(hbbe1.1:eGFP). Flow cytometric acquisitions and FACS isolations were performed on a Melody (BD) with FACS Chorus 2.0, version 1.1.20.0 software, and analyses using FlowJo software (v10.3 or v10.8.1, Tree Star). Representative gating strategies for the flow cytometric analyses presented in this manuscript are provided in Figure S8. Quantitative RT-PCR (qPCR) analysis RNA from sorted cells was collected with QIAshredder (Qiagen, 79656) followed by RNA isolation with miRNeasy (Qiagen, 217684). When whole embryos were used, they were first stored at −80°C overnight in 500 μL of trizol (Fisher Scientific,15596026) before being homogenized with an automatic homogenizer (Fisher Scientific,12-141-361) and pestel (Fisher Scientific, 12141363). RNA was then extracted via trizol-chloroform (Fisher Scientific, C298-500) extraction followed by clean up with the GeneJet RNA Purification Kit (Thermo Fisher, K0732). cDNA was generated with QuantiTect Reverse Transcription Kit (Qiagen, 205311). Primers to detect zebrafish transcripts are described in Table S1 qPCR was performed in CFX real-time PCR detection system (BioRad), and relative expression levels of genes were calculated by the following formula: Relative expression = 2(gene CQ – housekeeping CQ)*100,000. Log Fold increase was calculated with the following formula: Log Fold change= (Population of interest) 2[(geneCQ-housekeepingCQ)–averageCQ from EMP]/(EMP population)2[(geneCQ-housekeepingCQ)–averageCQ from EMP]. Cytological staining Cytospin preparations were made from 1,000 to 2,000 Lmo2+; Gata1+ fluorescence-activated cell sorting (FACS) -sorted cells cytocentrifuged at 1000 rpm for 5 min with medium acceleration onto charged glass slides (Superfrost Plus Microscope Slides, 22-037-246, Fisherbrand) in a Thermo Scientific Shandon CytoSpin 4. For kidney marrow cytology; whole embryos were dissected from juvenile fish following peripheral blood drawl. Dissected kidneys were immediately placed in 500 μL of ice-cold 0.9 x PBS with 5mg/ml heparin (VWR BT139975-1G). Kidneys were then manually dissociated with a P1000 pipette. 100 μL of solution was loaded into the cytospin following the same settings as previously stated. Cytospin preparations were stained with Wright-Giemsa for morphological analyses and differential hematopoietic cell counts. Briefly: Wright stain solution (Fisher Scientific, 5029782) was placed upon the air-dried smear for 2 min and washed with distilled water. Subsequently, the slides were placed for 10 min in a Coplin jar containing filtered Giemsa stain (Sigma-Aldritch, GS1L0-1L) and rinsed with distilled water. Samples were air-dried and mounted with Cytoseal XYL (Fisher Scientific, 22050262) and a coverslip (VWR, 48366-089). Images were taken with a 100× objective lens using an EVOS M5000 (Thermo Fisher Scientific) with EVOS software. 100–200 cells per sample were imaged and counted for FACS isolated cells. At least 200 cells were manually counted from kidney marrow samples. For quantification, EMPs were considered erythroid-biased based on round nuclei with condensed coarse chromatin. Myeloid-biased EMPs were classified as those with irregular nuclei shape and fine chromatin. Morphological analyses and differential cell count of kidney marrow hematopoietic cells Differential cell counts were assessed in 43 dpf stat3sa15744/sa15744 or stat3+/+ siblings zebrafish after cardiocentesis, kidney dissection, cytospin and staining with Wright-Giemsa. The morphological features to identify each cell type are the following. Early myeloid precursors, round to ovoid in shape with a round to ovoid-shaped pink nucleus. Chromatin pattern finely granular and semi frequently with distinct round nucleoli. High nuclei to cytoplasmic (N:C) ratio with a moderate amount of deep blue cytoplasm. Late myeloid precursors (immature neutrophils): round to ovoid to bilobed nuclei with light purple to darker pink coloration. Slightly clumped chromatin pattern and no distinct nucleoli. Moderate N:C ratio with a blue to light blue cytoplasm. Neutrophils (mature): smaller in size than late myeloid precursors. Round to ovoid to bilobed to segmented dark purple nucleus smaller than late myeloid precursors. Moderate N:C ratio with a light blue cytoplasm and often bubbly vacuolated appearance. Macrophages: round to ovoid in shape. Nuclear shapes included round, ovoid, bilobed, and lobulated forms that have a pale pink to purple coloration. The chromatin pattern is finely stippled and occasionally has distinct nucleoli. Moderate N:C ratio that is moderately high with a blue appearance frequently vacuolated and sometimes contained melanin pigment. Lymphocytes: small round to ovoid-shaped cells. Round to ovoid-shaped nucleus, coarsely clumped chromatin pattern, and no distinct nucleoli. High N:C ratio with a scant amount of pale blue cytoplasm. Eosinophil: The nucleus is ovoid to bilobed in shape with a coarsely clumped chromatin pattern. The cytoplasm contains many small pink granules. Erythroblast: Round to ovoid in shape. They have a nearly perfectly round nucleus with a coarsely stippled chromatin pattern and often one to a few, small to moderate-sized, distinct, round nucleoli. The N:C ratio is moderately high with a moderate amount of deep blue cytoplasm. Later erythroid precursor: Smaller than a rubriblast. They have a round nucleus with an irregularly clumped chromatin pattern. The cytoplasm is blue to light purple in coloration. Blast cell: Round to oval in shape and smaller than a myeloblast and erythroblast. They have a moderately high N:C ratio with blue cytoplasm. They have an ovoid-shaped nucleus with a lacy granular chromatin pattern with indistinct to a few distinct round nucleoli. Thrombocyte: Small ovoid to round-shaped cells with a very high N:C ratio and light blue cytoplasm that is occasionally wispy or contains a few small vacuoles. The ovoid to ellipsoid nucleus has a dense, dark purple, coarsely stippled chromatin pattern. Erythroid progenitors were distiguished from myeloid progenitors by round and centered nuclei with coarse chromatin, while the latter was identified by bean-shaped and uncentered nucleus with fine chromatin. Dendra photoconversion We utilized Tg(mpeg1:dendra2) zebrafish47 to lineage trace primitive macrophages and distinguish them from EMP-derived macrophages. Prior to photoconversion embryos were dechorionated with pronase (11459643001, Sigma) and anesthetized in 1% tricaine (Pentair; TRS5). Embryos were photoconverted Dendra2Green to Dendra2Red via exposure to a DAPI filter under the 10× objective of an Axio Imager A2 (Zeiss) for 2 min. Photoconversion was carried out at 28 hpf, based on when the Dendra signal was visible in the yolk ball but not the PBI via microscopy. Photoconverted embryos were then covered from the light and allowed to develop. At 48 hpf, embryos were either analyzed via flow cytometry using a BD FACS Melody with BD chorus software (Biosciences, v3.0) or confocal imaging on a Stellaris 5 confocal microscope with LASX software (Leica, v4.8.1.29271). Cleavage under targets and release using nuclease (CUT&RUN) CUT&RUN was performed following previously published procedures59 using the CUT&RUN Assay Kit (Cell Signaling; 86652). Briefly, 22 hpf AB* embryos were disassociated and filtered as described in the “Flow cytometry and FACS analysis” section. 1,000,000 cells were used per sample. Cells were labeled with a monoclonal mouse antibody against p-Stat3, Tyr708 (MBL Life Science, D128-3, dilution 1:25), or a rabbit (Cell Signaling, #66362, dilution 1:20) IgG Isotype control antibody as recommended by the manufacturer. DNA purification was carried by phenol/chloroform as indicated by the CUT&RUN Assay Kit instructions (Cell signaling, 86652S), and qPCR analysis of the predicted Stat3 binding sites (BSs) using the following primers: BS1 (−6,059), BS2 (−4,341), BS3 (−3,000), BS4 (+34,498), BS5 (+35.061), and BS6 (+41, 827) (Table S1). The manufacturer instructions were followed with the following modifications: The digitonin dilution used was 5:1000; incubation time for the primary antibody was overnight at 4°C; incubation for pAGMase was 1 h at 4°C; the digestion time was 30 min. To identify potential binding sites for Stat3 in the grna regulatory regions, the entire grna gene sequence and 10kb 5′ upstream and 3′ downstream were analyzed using two binding site prediction software: Tomtom119 and PROMO57 following the workflow previously described.59 qPCR targeting the predicted Stat3 binding sites was performed using the primers indicated in Table S1 and as described.59 To calculate the fold enrichment, qPCR Cq values for each BS were normalized against spike-in DNA as described and control primers that amplify tert, a locus Stat3 lacking predicted Stat3 BSs. Whole-mount immunohistochemistry (WIHC) Zebrafish embryos were fixed using 2% paraformaldehyde (Sigma-Aldrich, 158127-500G; Lot # STBH9750)and incubated at 4°C overnight. Fixed embryos were then blocked using a solution of 0.3% Triton X-100 (VWR, 97063-864) in PBS with 4% bovine serum albumin (BSA) (Sigma, A3311-10G) and incubated at 4°C for 1 h. Primary antibodies were diluted in 0.3% Triton, 4% BSA blocking solution, 200 μL of antibody solution was applied to embryos for three days and incubated in a Bio Rad Mini Rocker (1660710EDU) at 4°C. Primary antibodies used were: anti-p-STAT3 (MBL Life Science, D128-3, 1:250), anti-Stat3 (F-2) (Santa Cruz, sc-8019, 1:50), anti-pStat3 (Tyr-708) (MBL International Corporation, D128-3, 1:250). Primary antibodies were washed with a solution of 0.3% Triton X-100 in PBS (wash solution) for 2 min 20 to 35 times. Secondary antibodies were diluted with 0.3% Triton, 4%BSA block and incubated overnight at 4°C on a rotator. Secondary antibodies used included Donkey anti-Rabbit Alexa Flour 594 (Life Tech A21207, 1:1000), 4′,6-diamidino-2-phenylindole (DAPI) (1:1000, Life Science Technology 62248), and Donkey anti-Mouse Alexa Flour 488 (Life Tech A21202_3704259636, 1:1000). After incubation with secondary antibodies, all samples were protected from light exposure until the commencement of imaging. Embryos were incubated in 70% glycerol (Sigma, G5516-10ML) at 4°C for 10 min and transferred into 100% glycerol for confocal imaging. Images were taken using a Stellaris 5 confocal microscope with LASX software (Leica, v4.8.1.29271). Fin amputation and wound healing assay The following assay methodology is based on that described.19 Tg(mpeg1:dendra2) or Tg(mpeg1:eGFP) embryos were anesthetized with tricaine and cut across the dorsoventral length of the caudal fin at 48 or 24 hpf respectively, just after the end of the notochord, using a size 15 surgical blade (19-200-218, Fisher Scientific). Individual embryos were isolated in 24-well plates to allow tracking of each embryo and imaged at the indicated times using a Stellaris 5 confocal microscope with LASX software (Leica, v4.8.1.29271). 48 hpf stat3sa15744 embryos were also subjected to fin amputation, but a Leica M205 FCA stereomicroscope equipped with a THUNDER imager, DFC9000 sCMOS camera (Leica), and LASX imaging software (Leica, v4.7.028176) used for imaging. stat3sa15744 embryos were allowed to recover for three days following fin resection, with rejenerated fin area being recorded every 24h. Comparison of images from the initial time point to each later time point was performed using CorelDraw to allow the original image to be made semi-transparent before being overlayed and aligned. Images across time points were aligned based on the position of the notochord. Quantification of the regenerated area was determined using Fiji (ImageJ, Java 1.8.0_322) software. Heat shock treatment Tg(hsp70:Gal4) was crossed with Tg(14xUAS:grna-P2A-mRFP1) or Tg(14xUAS:ca-stat3-P2A-mRFP1). At the indicated developmental stages, 15 embryos were transferred to a falcon tube (Fisher Scientific, 14-432-22) containing 45mL of embryo media (pre-warned to 38°C). The tube and embryos were then submerged into a Thermo Neslab Ex7 Digital One Circulating Bath (Thermo Fisher) set at 38°C for 1hr. Following the heat shock, embryos were transferred to room temperature embryo media and placed in a 28.5°C incubator until analysis. RNA-seq data analysis Previously published bulk RNA sequencing (Gene Expression Omnibus: GSE155258)19 were analyzed and represented using Consesus Path BD106,120 and ROSALIND (https://rosalind.bio/), with a HyperScale architecture developed by ROSALIND, Inc. (San Diego, CA). Reads were trimmed using cutadapt.121 Quality scores were assessed using FastQC.122 Reads were aligned to the Danio rerio genome build danRer10 using STAR.123 Individual sample reads were quantified using HTseq124 and normalized via Relative Log Expression (RLE) using DESeq2 R library.125 Read Distribution percentages, violin plots, identity heatmaps, and sample MDS plots were generated as part of the QC step using RSeQC.126 DEseq2 was also used to calculate fold changes and p-values and perform optional covariate correction. Clustering of genes for the final heatmap of differentially expressed genes was done using the PAM (Partitioning Around Medoids) method using the fpc R library.127 Hypergeometric distribution was used to analyze the enrichment of pathways, gene ontology, domain structure, and other ontologies. The topGO R library,128 was used to determine local similarities and dependencies between GO terms in order to perform Elim pruning correction. Several database sources were referenced for enrichment analysis, including Interpro,129 NCBI,130 MSigDB,131,132 REACTOME,133 WikiPathways.134 When using Consensus Path BD, the p-value cut-off for the genes input into the software was 0.05. The p-value cut off for the identified pathway-based sets was 0.01 with a minimum overlap of 2 within the input list. We utilized Wiki pathways, Kegg, Smpdb, and Reactome as the defined pathway databases. Enrichment was calculated relative to a set of background genes relevant for the experiment. Lentivirus production and stable cell line generation Human GRN was amplified from pLX304-GRN (clone HsCD00438248, The ORFeome Collaboration, DNA SU Plasmid Repository) and cloned into pCS2+113 using the primers indicated in Table S1. GRN was subcloned from pCS2+-GRN into pLVX_IRES_Hyg (Clon-tech, 632185) using the primers denoted in Table S1 pLVX _IRES_Hyg with no insert was utilized as mock control. DNA constructs were validated by sequencing. Lentiviral particles were produced using a second-generation lentiviral packaging system through co-transfection of HEK293T cells with packaging plasmids pVSV-G (Addgene, #138479), psPAX2 (Addgene, #12260), and the lentiviral expression vector pLVX_IRES_Hyg containing the human GRN gene. Lentiviruses were collected, filtered, and concentrated with Lenti-X Concentrator (Takara Bio, Cat. Nos. 631231 & 631232) following the manufacturer instructions, after which they were resuspended in PBS and stored at −80°C until further use. Lentiviral transduction To generate a stable pool of GRN-expressing K562, 3.0 × 105 cells were transduced with GRN lentiviral vectors containing a hygromycin-resistance cassette as previously described.135 Viral transduction was performed overnight in the presence of 8 μg/mL polybrene. After 8 days, hygromycin selection was started (200 μg/mL, Sigma-Aldrich, St. Louis, MI, USA). Cell pellets were collected after selection and vector copy number was measured in cell pools as previously described.136 Stable cell pools of GRN and mock-transduced K562 had an average vector copy number of 2.9 and 1.4, respectively. Flow cytometry analysis was performed on GRN-expressing K562 selected pools induced with 100 ng/mL interlukin-6 (IL6) or Phorbol 12-myristate 13-acetate (PMA) for 24 h (CytoFLEX Flow Cytometer, Model: V3-B3-R3 Beckman Coulter, Indianapolis, IN, USA). Staining was performed as previously described19 and cells were analyzed using FlowJo. RNA was extracted with RNAesy micro kit according to manufacturer’s instruction (Qiagen, Hilden, Germany) and cDNA was generated as previously described.137 Real time qPCR was performed using TaqMan Fast Advanced Master Mix (Thermofisher). The expression of GRN was normalized using human GAPDH135 as a reference gene (NM_002046.6) and represented as 2ΔCt for each sample or as fold changes (2ΔΔCt) relative to the control. QUANTIFICATION AND STATISTICAL ANALYSIS Each experiment was performed two to three independent times. Data was analyzed by either unpaired t test when comparing two conditions, or one-way ANOVA with Tukey’s Honest Significant Difference post-hoc test when comparing three or more conditions, using GraphPad software (Prism 8). In all figures, significance is represented by the specific p-value and was defined as p-value <0.05. The middle bars denote the mean, and the error bars represent SD. Illustrations were generated using CorelDRAW (Corel Corporation), unless otherwise indicated. Supplementary Material Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2026.117477. KEY RESOURCES TABLEREAGENT or RESOURCESOURCEIDENTIFIERReagentsFetal bovine serumLife technologies26140-079HEPESGlibco15630080Sodium pyruvateGlibco11360070Penecilin- StreptomyocinGlibco15140122Hygromycin B (50mg/mL)Roche10843555001RPMI 164Glibco11875085Q5® High-Fidelity DNA PolymeraseNew England BiosciencesM0491SKpn1New England BiosciencesR3142SNuclease-free waterFisher ScientificBP2484100Phenol red solutionSigmaP0290NBT/BCIP solutionRoche11681451001PronaseMilliporeSigma11459643001LiberaseMilliporeSigma5401127001Dulbecco’s Phosphate BufferedSaline solution with Ca2+ and Mg2MilliporeSigmaD866230 μm nylon meshFisher ScientificNC908444121-gauge syringeFisher ScientificBD 309624SYTOX™ RedLife TechnologiesS34859ChloroformFisher ScientificC298-500TrizolFisher Scientific15596026HeparinVWRBT139975-1GWright stain solutionFisher Scientific5029782Giemsa stainSigma-AldritchGS1L0-1LCytoseal XYLFisher Scientific22050262PronaseSigma11459643001TricainePentairT RS5ParaformaldehydeSigma-Aldrich158127-500GTriton X-100VWR97063-864Bovine serum albuminSigmaA3311-10GLenti-X ConcentratorTakara Bio631231 & 631232RNase A powderNeoBiotechNB-03-0163DMEM/F-12 GlutaMAXGibcoA4192001DAPILife Science Technology62248TaqMan Fast Advanced Master MixThermo Fisher4444556TaqMa Fast Advanced Master MixThermo Fisher4444556Antibodiesp-Stat3, Tyr708MBL Life ScienceD128-3Rabbit IgG(control negative)Cell Signaling66362; RRID:AB_2924329Anti-Digoxigenin-AP Fab fragmentsRoche11093274910; RRID:AB_514497Donkey anti-Rabbit Alexa Flour 594Life TechnologiesA21207; RRID:AB_141637Donkey anti-Mouse Alexa Flour 488Life TechnologiesA21202; RRID:AB_141607Goat Anti-Chicken Alexa Flour 488Life TechnologiesA11039; RRID:AB_2534096APC Mouse Anti-Human CD11bBD Pharmingen550019; RRID:AB_398456PE Mouse Anti-Human CD14BD Pharmingen555398; RRID:AB_395799PE/Cyanine7 anti-human CD16 AntibodyBiolegend302016; RRID:AB_314216FITC anti-human CD45 AntibodyBiolegend304006; RRID:AB_314394Pacific Blue™ anti-human CD235ab AntibodyBiolegend306612; RRID:AB_2116241Pacific Blue™ anti-human CD45 AntibodyBiolegend304022; RRID:AB_493655CD34 Monoclonal Antibody (4H11), PE-Cyanine7Invitrogen25-0349-42; RRID:AB_1963576FITC Mouse Anti-Human CD61BD Pharmingen555753; RRID:AB_396094PE Mouse Anti-Human CD71BD Pharmingen555537; RRID:AB_395921Experimental ModelsZebrafish: AB*ZIRCZDB-GENO-960809-7Zebrafish: grna54Sholenberger et al.40N/AZebrafish: stat3sa15744Kettelborough et al.103ZDB-ALT-130411-4288Zebrafish: Tg(gata1:dsRed)Traver et al.43ZDB-TGCONSTRCT-070117-38Zebrafish: Tg(mpeg1:eGFP)Ellet et al.42ZDB-ALT-120117-1Zebrafish: Tg(lmo2:eGFP)Zhu et al.48ZDB-TGCONSTRCT-080122-3Zebrafish: Tg(hsp70:gal4)Scheer et al.60ZDB-TGCONSTRCT-070117-42Zebrafish: fli1rstpl50GTCraig et al.50ZDB-ALT-151120-2Zebrafish: Tg(14xUAS:grna-P2A-mRFP1)is505This paperN/AZebrafish: Tg(14xUAS:stat3C-P2A-mRFP1)is506This paperN/AZebrafish: Tg(phldb1:gal4-mCherr)Distel et al.104N/AZebrafish: Tg(mpx:eGFP)i114Renshaw et al.105ZDB-TGCONSTRCT-070118-1Zebrafish: Tg(mpeg1:Dendra2)Harvie et al.47ZDB-TGCONSTRCT-150430-2Cell line: K562 CellsATCCCCL-243Cell line: HEK293T cellsATCCCRL3216OligonucleotideMorpholino: StandardGeneToolsN/AMorpholino: stat3Yamashita et al.52N/AMorpholino: grna mismatchCampbell et al.19N/AMorpholino: jak2aMa et al.65N/AMorpholino: grnaCampbell et al.19N/ACloning primer: GRN Table S1 N/ACloning primer: grna-P2A Table S1 N/ACloning primer: ca-stat3 Table S1 N/ACloning primer: P2A Table S1 N/ACloning primer: grna Table S1 N/AqPCFR: GRN Table S1 N/AqPCR primer: efa1 Table S1 N/AqPCR primer: grna Table S1 N/AqPCR primer: stat3 Table S1 N/AqPCR primer: stat3-BS1 Table S1 N/AqPCR primer: stat3-BS2 Table S1 N/AqPCR primer: stat3-BS3 Table S1 N/AqPCR primer: stat3-BS4 Table S1 N/AqPCR primer: stat3-BS5 Table S1 N/AqPCR primer: stat3-BS6 Table S1 N/AqPCR primer: tert Table S1 N/ACustom TaqMan SNP Genotyping AssayThermo Fisher4332075 Assay ID: ANGZY9YPlasmidsPlasmid: Stat3-C Flag pRc/CMVAddgene8722pVSV-GAddgene138479psPAX2Addgene12260pLX304-GRNThe ORFeome Collaboration, DNA SU Plasmid Repositorclone HsCD00438248pLVX_IRES_HygClontech632185Software and algorithmsConsesus Path BDHerwig et al.106N/AROSALIND https://rosalind.bio/ N/AFIJISchneider et al.107LASXLeicaN/AFlowJoTree StarN/AFireCamLeicaN/AFACS ChorusBDN/AEVOSThermo Fisher ScientificN/ATomTomGupta et al.56N/APROMOMesseguer et al.57N/AGraphPadPrism 8N/ACorelDrawCorelDrawGraphic Suite 2020HardwareCytoFLEX Flow CytometerBeckman Coulter Life SciencesV3-B3-R3PLI-90A pico-injectorWarner instrument64-1738M165FC stereomicroscopeLeicaN/ADFC295 color digital cameraLeicaN/ACFX Connect Real-Time PCR SystemBioRadCFX Maestro 2.0, v5.0.021.0616BD FACSMelody™ Cell SorterBD BioscienceN/AAutomatic homogenizerFisher Scientific12-141-361PestelFisher Scientific12141363Superfrost™ Plus Microscope SlidesFisherbrand22-037-246Shandon CytoSpin 4Thermo ScientificN/ACoverslipVWR48366-089Axio Imager A2ZeissN/AEVOS M5000Thermo Fisher ScientificN/AStellaris 5 confocal microscopeLeicaN/ASize 15 surgical bladeFisher Scientific19-200-218THUNDER imagerLeicaN/ADFC9000 sCMOS cameraLeicaN/AFalcon tubeFisher Scientific14-432-22Thermo Neslab Ex7 Digital One Circulating BathThermo FisherN/AMini RockerBioRad1660710EDUCommercial AssaysSuperscript IIIInvitrogen18080051NEBuilder® HiFi DNA Assembly Cloning KitNew England BiosciencesE5520mMESSAGE mMACHINE SP6 Transcription KitThermo Fisher ScientificAM1340Monarch® PCR & DNA Cleanup KitNew England BiosciencesT1030LmMESSAGE mMACHINE™ T7 ULTRA Transcription KitInvitrogenAM1345DIG RNA Labeling KitRoche Applied Science11277073910miRNeasyQiagen217684QIAshredderQiagen79656QuantiTect Reverse Transcription KitQiagen205311GeneJet RNA Purification KitThermo FisherK0732CUT&RUN Assay KitCell Signaling86652 KEY RESOURCES TABLE Highlights Granulin a (grna) drives definitive but not primitive myeloid fate in zebrafish Granulin a (grna) drives definitive but not primitive myeloid fate in zebrafish Grna amplifies Jak2/Stat3 signaling to enforce the myeloid program Grna amplifies Jak2/Stat3 signaling to enforce the myeloid program Primitive- and EMP-derived macrophages respond differentially to injury stimuli Primitive- and EMP-derived macrophages respond differentially to injury stimuli Granulin directs human myeloid lineage commitment Granulin directs human myeloid lineage commitment ACKNOWLEDGMENTS This work was supported by the NIH-NIDDK R01 (R01DK131162), NIH-NIDDK R03 (R03DK125661), NIH-NIDDK K01 award (7K01DK115661), and the Roy J. Carver Charitable Trust (#21–5532) to R.E.-P. This work was supported in part by the Children’s Hospital of Philadelphia Breakthrough Funds to C.W; the Regional Ministry of Education of Castilla y Leon (SA013G24) to J.L.; the United States Department of Agriculture National Institute of Food and Agriculture Hatch project IOW05706 and Iowa State University start-up funds to C.C. and R.E.-P. The findings and conclusions in this publication are those of the author(s) and should not be construed to represent any official USDA or US Government determination or policy. We thank Sam Birkeland, Kya Thiel, Andrea Behanish, Amelia Case, and Jessica Tupy for technical assistance and fish husbandry. Footnotes RESOURCE AVAILABILITY Lead contact Inquiries regarding further information and resources should be directed to the lead contact, Raquel Espin Palazon (espin@iastate.edu). Materials availability All unique/stable reagents and zebrafish lines generated in this study are available from the lead contact with a completed materials transfer agreement. Data and code availability This paper analyzes existing, publicly available data, accessible at https://doi.org/10.1182/bloodadvances.2020003096. This paper does not report original code. Microscopy and flow cytometric data reported in this paper will be shared by the lead contact upon request. DECLARATION OF INTERESTS The authors declare no competing interests. REFERENCES Associated Data Supplementary Materials