
==== Front
Nat Commun
Nat Commun
Nature Communications
2041-1723
Nature Publishing Group UK London

51920
10.1038/s41467-024-51920-7
Article
The heart is a resident tissue for hematopoietic stem and progenitor cells in zebrafish
http://orcid.org/0000-0003-4084-0250
Bornhorst Dorothee 12
http://orcid.org/0009-0001-2510-5242
Hejjaji Amulya V. 12
Steuter Lena 12
Woodhead Nicole M. 3
Maier Paul 4
http://orcid.org/0000-0002-5423-7295
Gentile Alessandra 58
http://orcid.org/0009-0001-8006-2732
Alhajkadour Alice 3
http://orcid.org/0000-0002-6907-4564
Santis Larrain Octavia 3
http://orcid.org/0000-0001-8007-9975
Weber Michael 4
http://orcid.org/0000-0001-5672-3791
Kikhi Khrievono 6
http://orcid.org/0000-0002-5594-4549
Guenther Stefan 7
Huisken Jan 4
http://orcid.org/0000-0001-9146-4860
Tamplin Owen J. 3
http://orcid.org/0000-0002-0382-0026
Stainier Didier Y. R. 5
http://orcid.org/0000-0002-5592-9680
Gunawan Felix felix.gunawan@ukmuenster.de

12
1 https://ror.org/00pd74e08 grid.5949.1 0000 0001 2172 9288 Institute of Cell Biology, Faculty of Medicine, University of Münster, Münster, 48149 Germany
2 https://ror.org/00pd74e08 grid.5949.1 0000 0001 2172 9288 ‘Cells-in-Motion’ Interfaculty Center, University of Münster, Münster, 48149 Germany
3 https://ror.org/01y2jtd41 grid.14003.36 0000 0001 2167 3675 Department of Cell and Regenerative Biology, School of Medicine and Public Health, University of Wisconsin-Madison, Madison, WI 53705 USA
4 https://ror.org/01y9bpm73 grid.7450.6 0000 0001 2364 4210 Multiscale Biology, Faculty of Biology and Psychology, Georg-August-University Göttingen, Göttingen, 37077 Germany
5 https://ror.org/0165r2y73 grid.418032.c 0000 0004 0491 220X Department of Developmental Genetics, Max Planck Institute for Heart and Lung Research (MPI-HLR), Bad Nauheim, 61231 Germany
6 Flow Cytometry and Cell Sorting Core Facility, MPI-HLR, Bad Nauheim, 61231 Germany
7 Deep Sequencing Platform, MPI-HLR, Bad Nauheim, 61231 Germany
8 https://ror.org/0220mzb33 grid.13097.3c 0000 0001 2322 6764 Present Address: Centre for Developmental Neurobiology, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, London, UK
31 8 2024
31 8 2024
2024
15 75891 12 2023
20 8 2024
© The Author(s) 2024
2024
https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
The contribution of endocardial cells (EdCs) to the hematopoietic lineages has been strongly debated. Here, we provide evidence that in zebrafish, the endocardium gives rise to and maintains a stable population of hematopoietic cells. Using single-cell sequencing, we identify an endocardial subpopulation expressing enriched levels of hematopoietic-promoting genes. High-resolution microscopy and photoconversion tracing experiments uncover hematopoietic cells, mainly hematopoietic stem and progenitor cells (HSPCs)/megakaryocyte-erythroid precursors (MEPs), derived from EdCs as well as the dorsal aorta stably attached to the endocardium. Emergence of HSPCs/MEPs in hearts cultured ex vivo without external hematopoietic sources, as well as longitudinal imaging of the beating heart using light sheet microscopy, support endocardial contribution to hematopoiesis. Maintenance of these hematopoietic cells depends on the adhesion factors Integrin α4 and Vcam1 but is at least partly independent of cardiac trabeculation or shear stress. Finally, blocking primitive erythropoiesis increases cardiac-residing hematopoietic cells, suggesting that the endocardium is a hematopoietic reservoir. Altogether, these studies uncover the endocardium as a resident tissue for HSPCs/MEPs and a de novo source of hematopoietic cells.

The endocardium lines the interior of the heart chambers and has been debated as a source of hematopoietic lineages. Here they show that the endocardium may act as a source of, and resident tissue for, hematopoietic stem and progenitor cells in zebrafish, providing evidence for diversity in origins and residences of hematopoietic cells.

Subject terms

Heart stem cells
Haematopoietic stem cells
Heart development
https://doi.org/10.13039/501100004870 University of Münster | Medizinische Fakultät, Westfälische Wilhelms-Universität Münster (Medical Faculty, WWU Münster) I-GU122208 Gunawan Felix https://doi.org/10.13039/100008662 Joachim Herz Stiftung (Joachim Herz Foundation) https://doi.org/10.13039/100010663 EC | EU Framework Programme for Research and Innovation H2020 | H2020 Priority Excellent Science | H2020 European Research Council (H2020 Excellent Science - European Research Council) Max Planck Institute for Heart and Lung Researchissue-copyright-statement© Springer Nature Limited 2024
==== Body
pmcIntroduction

A specialized subset of endothelial cells undergoes endothelial-to-hematopoietic transition (EHT) to give rise to hematopoietic stem and progenitor cells (HSPCs)1–5. EHT occurs in 1-3% of endothelial cells in the aorta-gonad-mesonephros (AGM) in mammals, the first vascular source of hematopoiesis1–5. These hemogenic endothelial cells upregulate factors critical to promote hematopoiesis, including the transcription factors Runx1, Gata2, and cMyb, the receptor Integrin αIIb/CD41, and members of the Notch, BMP, and Wnt signaling pathways6–8. The HSPCs then migrate into intermediary niches where the hematopoietic population expands before residing in their final niche, the bone marrow in mammals. HSPCs constitute only ∼1 in 10,000 bone marrow cells9, but continuously sustain blood production10–12. HSPCs can differentiate into common myeloid precursors and megakaryocyte-erythroid precursors (MEPs), the latter of which give rise to platelets and erythrocytes13. Extensive research has mostly been directed towards understanding how the endothelial cells at the first hematopoietic sites give rise to HSPCs4,5,14. However, increasing evidence has pointed to the capacity of endothelial cells in tissues outside of the AGM, such as the lung15,16 and the head17, to undergo EHT.

In hematopoietic niches, endothelial cells also play crucial roles in regulating HSPC maintenance and pluripotency18,19. Endothelial cells in the intermediary niches (the fetal liver in mammals and the caudal hematopoietic tissue (CHT) in zebrafish) and the bone marrow secrete angiocrine factors that sustain HSPC quiescence and self-renewal20. Endothelial-specific deletion of factors required for survival of endothelial cells or maintenance of HSPCs leads to loss of HSPCs18,20,21, highlighting the importance of endothelial-HSPC communications in the niche. Whether endothelial cells can support the maintenance of HSPCs in organs other than the well-studied intermediary niches14,22 and the bone marrow remains poorly understood. This knowledge gap limits our understanding of how differences in the locations where hematopoietic cells reside contribute to heterogeneity in HSPC characteristics23,24.

One subpopulation of endothelial cells proposed to undergo EHT is the endocardial cells (EdCs), which form the endothelial lining of the heart. The endocardium and the contractile myocardium, which consists of cardiomyocytes (CMs), compose the first layers of the developing heart25,26. EdCs possess a distinct molecular profile compared with other endothelial cells27–29 and a degree of plasticity to differentiate into other cells30, including valve mesenchymal cells and coronary vessels31–34. There are also surprising findings regarding the contribution of EdCs to CM35,36 or mural cell37 populations.

Research concerning endocardial contribution to the hematopoietic lineage is contentious. In mouse, several studies have reported a subset of early-stage EdCs differentiating into hematopoietic cells of erythroid, myeloid, and macrophage identities38–42. The function of EdCs in giving rise to macrophages has been proposed to be particularly relevant, with cardiac valves being hyperplastic in the absence of endocardial-derived macrophages39. These studies also reported higher phagocytic activities in endocardial-derived macrophages compared with other macrophages30, suggesting that cardiac-derived macrophages are functionally distinct from others39. However, other recent studies that developed endocardial-specific Cre lines performed careful lineage tracing of EdCs during development and reported an almost complete absence of endocardial-derived macrophages or circulating hematopoietic cells in mouse embryos43,44. Thus, the prevalence of endocardial-to-hematopoietic transition in mouse remains elusive. Reliance on using fixed and sectioned cardiac tissues for imaging, coupled with potential challenges with the specificity of Cre expression in the existing transgenic lines, have also raised difficult obstacles in addressing this intriguing question in mouse.

In this work, we use zebrafish as a model to longitudinally image the heart at high resolution and manipulate cardiac and hematopoietic development. The cardiovascular and hematopoietic systems share many similarities between zebrafish and mammals45,46, with endothelial cells in the dorsal aorta (DA; first site of hematopoiesis) undergoing EHT47, migrating into the CHT (intermediary niche)48, and finally into the kidney (presumptive adult niche)48. A recent study uncovered a potential for EdCs to undergo EHT prior to 24 hpf, with most endocardial-derived cells fated to become neutrophils49. In our study, combining high-resolution imaging in live animals and single-cell transcriptomic analysis, we provide several lines of evidence that the endocardium can not only give rise to hematopoietic cells but also maintain a stable population of hematopoietic cells, which consists mainly of HSPCs/MEPs.

Results

A subset of endocardial cells expresses enriched levels of hematopoietic genes

The endocardium has emerged as a dynamic tissue showing remarkable plasticity30. Most studies on endocardial morphogenesis focus on specification and patterning of the valve50–53, but the involvement of non-valve EdCs in driving endocardial tissue morphogenesis remains largely unexplored. To uncover other endocardial subpopulations, we performed single-cell RNA sequencing from EdCs isolated by fluorescence-activated cell sorting (FACS) of Tg(kdrl:nls-mCherry)-positive cells from dissected hearts (Fig. S1A). We focused on two developmental time points, 48 and 72 hpf, to identify morphogenetic processes that pattern the endocardium as the heart continues to loop and grow, valves take shape, and trabecular networks start to form54.

From this dataset, we classified 11 clusters of EdCs with distinct enriched gene signatures (Fig. 1A, S1B, Supplementary Table 1). We identified two clusters that exhibit enrichment of genes involved in valve development53,55,56, with one cluster composed of 1830 cells being consistent with valve EdCs (alcama, has2, hey2) and the other composed of 359 cells consistent with fibroblast-like valve interstitial cells (col1 genes, postnb, twist1b) (Fig. 1A, S1B, Supplementary Table 1). The largest cluster is composed mostly of 2274 non-valve quiescent EdCs in the G1 phase (Fig. 1A, S1D). Two other clusters are primarily composed of non-valve EdCs that were undergoing DNA synthesis (620 cells in the S phase) or dividing (657 cells in the G2/M phase) (Fig. 1A, S1D). A small cluster of 299 cells expresses genes consistent with lymphatic endothelial cells (lyve1a/b, flt4, mafbb; Fig. 1A, S1B, Supplementary Table 1). The remaining clusters were not as clearly defined in their transcriptomic signatures, suggesting that they are previously unidentified endocardial subpopulations with currently unknown functions.Fig. 1 Hematopoietic cells are stably maintained and grow in numbers on the endocardium.

A Single-cell RNA sequencing analysis of EdCs sorted by FACS for Tg(kdrl:nls-mCherry) expression at 48 and 72 hpf. The UMAPs reveal a subcluster of 390 cells that expresses genes affiliated with hematopoietic differentiation mostly at 72 hpf. B UMAP of the hematopoietic endocardial cell cluster indicates 5 subclusters. C Violin plots showing that genes enriched in hematopoietic cells at 72 hpf, such as myb (HSPCs/MEPs), mpx (neutrophils), gata1 (erythrocytes), and mpeg1.1 (macrophages), are enriched in this endocardial hematopoietic cluster, with HSPCs/MEPs and erythrocytes being the predominant cell types. D–F Representative confocal images of Tg(cd41:GFP) expression labeling HSPCs (low GFP; red asterisks) and platelets (high GFP; white asterisks) attached to the endocardium marked by Tg(kdrl:nls-mCherry) expression. D At 50 hpf, cd41:GFP+ cells are almost absent in the heart. E, F From 74 hpf onwards, cd41:GFP+ cells are attached to the endocardium. G HSPC (cd41:GFPlow+) numbers progressively increase from 56 hpf and remain until at least 10 dpf. n = 10 (50 hpf), 13 (56 hpf), 19 (74 hpf), 27 (98 hpf), 17 (120 hpf), 33 (7 dpf), and 19 (10 dpf) hearts. H Schematic heat map representing the positional mapping of cd41:GFP+ cells, the majority of which (66.5%) attaches to the outer curvature of the ventricle. n = 20 hearts. (I, J) Quantification of co-surface localization of cd41:GFP+ cells and endocardium labeled by Tg(kdrl:BFP-CAAX) expression in the ventricle. HSPCs share 11%, 27.4%, and 29.3% of surface areas with the ventricular endocardium at 74, 96, and 120 hpf, respectively. n = 20 (74 hpf), 28 (96 hpf), and 23 (120 hpf) hearts, respectively. K, K’ Representative confocal images of EdU pulse labeling assays (50–74 hpf) in Tg(cd41:GFP), Tg(kdrl:nls-mCherry) larvae. Most cd41:GFP+ cells attached to the heart are EdU-positive at 74 hpf (yellow asterisks). L Quantification of the proportions of cd41:GFP+ cells positive for EdU after a pulse assay between 50–74 and 74–98 hpf. n = 14 (50–74 hpf) and 17 (74–96 hpf) hearts. One-way ANOVA with Tukey’s multiple comparison test (G, I, J) or Unpaired, two-tailed t test (L) was used. Scale bars, 30 µm (D–F, K). Source data are provided as a Source Data file.

Interestingly, one cluster most distinct from the other clusters and found almost exclusively at 72 hpf (14 cells at 48 hpf and 376 cells at 72 hpf) exhibits an enrichment of genes that promote hematopoiesis or are expressed in differentiated hematopoietic cells (Fig. 1A, B; “endocardial-hematopoietic cluster” or EHC). At 72 hpf, the EHC accounts for 9.1% of total EdCs (376 out of 4,125 cells). Within the EHC, the majority of cells express high levels of genes known to be involved in promoting HSPC fate in the CHT57, including myb, drl, drll.2, hdr, ikzf1, and gata2b (Fig. 1B, C, S1C, S2A–E). Many EHC cells also express genes involved in erythrocyte differentiation, including gata1a and hbae3 (Fig. 1C, S1C, S2G). Two additional subclusters express high levels of platelet-promoting genes58, including mpl and nfe2 (Fig. 1B, S2I, 165 cells). Together, the cells expressing genes enriched in HSPCs, erythrocytes, and platelets constitute 360 out of the 390 cells (Fig. 1C, S2A–E, G, I), suggesting that most cells within the EHC are HSPC/MEPs. Although it has recently been reported that hematopoietic cells derived from the endocardium at 24 hpf are primarily neutrophils49, we found only a small subset of cells (15 cells) among 72 hpf EdCs expressing high levels of neutrophil markers (mpx and lyz) within subcluster 2 (Fig. 1C, S2A, F). Another small subcluster of 24 cells expresses macrophage-enriched markers (mpeg1.1, mpeg1.2, and irf8) (Fig. 1C, S2A, H). However, we could not detect any cell differentiation trajectories that uncover populations of EdCs in the process of EHT, likely due to low numbers of EdCs actively undergoing EHT at the timepoint of cell isolation. In conclusion, our single-cell transcriptomic dataset uncovers a previously unknown subgroup of EdCs with a hematopoietic signature.

Hematopoietic cells are stably attached to the endocardium

We sought to visualize the presence of the hematopoietic cells within the endocardium by performing confocal imaging of transgenic lines that label HSPCs (Tg(cd41:GFP) and Tg(Runx1:mCherry) expression), as well as endothelial cells (Tg(kdrl:GFP) and Tg(kdrl:nls-mCherry) expression) at 50, 74, 98, and 120 hpf. At 50 hpf, we observed almost no hematopoietic cells attached to the endocardium (Fig. 1D), consistent with our RNA-sequencing dataset that shows almost no hematopoietic cells at 48 hpf (Fig. 1A). Strikingly, at 74 hpf, we observed hematopoietic cells, the majority of which were positive for the HSPC/platelet marker (Tg(cd41:GFP)+ cells; Fig. 1E, F; Supplementary Movies 1, 3, 4), attached to the endocardium. Many hematopoietic cells were also positive for the erythrocyte marker (Tg(gata1:DsRed)+; Fig. S3A), but only very few were positive for the neutrophil marker (Tg(mpx:GFP)+; Fig. S3B), consistent with our single-cell sequencing results (Fig. 1A). Almost all the hematopoietic cells attached to the endocardium exhibited a rounded shape and nucleus, suggestive of undifferentiated HSPCs, instead of mature erythrocyte or myeloid morphologies that are larger and elongated (Fig. 1E, F). Importantly, these hematopoietic cells were not part of the circulating blood as observed through the imaging of beating hearts, which revealed their stable attachment to the endocardium without any movement in the cardiac lumen (Supplementary Movies 1, 2). These hematopoietic cells were never attached to the cardiac valve (Fig. 1E, F), suggesting that valve cells lack the adhesive properties for hematopoietic cells.

As most endocardial-associated hematopoietic cells appeared positive for HSPC/MEP identity (Fig. 1A–C, S1C, S2A–I; Supplementary Movies 1, 3, 4), we focused on the emergence and maintenance of these cells in the heart. We labeled them in the heart over several developmental time points using Tg(cd41:GFP) expression (hereafter referred to as cd41:GFP+ cells), a commonly used and reliable marker where low GFP signal is detected in HSPCs and high GFP signal in platelets. cd41:GFP+ cells were almost absent at 50 hpf, but started appearing in very low numbers at 56 hpf in 52% (11 out of 21) embryos (Fig. 1D, G; Supplementary Table 2). The numbers of cd41:GFPlow+ and total cd41:GFP+ cells significantly increased at 74 hpf (Fig. 1E, G; average of 10.1 cd41:GFPlow+ cells and 23.1 cd41:GFP+ cells per heart) and continued to rise at 98 hpf (Fig. 1F, G; average of 25 cd41:GFPlow+ cells and 33 cd41:GFP+ cells per heart), and 120 hpf (Fig. 1G; average of 40.2 cd41:GFPlow+ cells per heart). The cd41:GFPlow+ cells persisted at stable numbers until at least 10 days post fertilization (dpf) (Fig. 1G; 32.7 and 27.8 cd41:GFPlow+ cells at 7 and 10 dpf, respectively). By 74 hpf, 96% of the hearts had cd41:GFP+ cells on their endocardium; from 100 hpf onwards, all hearts contained cd41:GFP+ cells (Supplementary Table 2). Positional mapping of cd41:GFPlow+ cells revealed the highest frequency of HSPC attachment on the luminal surface of the outer curvature of the ventricle, close to the outflow canal (Fig. 1H; Supplementary Movies 1–4; 66.5% of endocardial-attached HSPCs), uncovering a regional preference for hematopoietic attachment.

To measure the adherence of the HSPCs to the endocardial surface, we quantified the degree of overlap between the HSPCs, as labeled by Tg(cd41:GFP) expression, and the endocardium, as labeled by Tg(kdrl:BFP-CAAX) expression. We found that over time, there was a progressive increase of the surface of HSPCs that overlapped with the endocardium, indicating that HSPCs gain significantly higher contact area with the endocardium, particularly between 74 and 96 hpf (Fig. 1I, J, S4A–C). The increasing contact areas between these two cell types suggest that the HSPCs become more adherent to the endocardium over time.

To strengthen our findings that the endocardial hematopoietic cells are HSPCs, we imaged and quantified Tg(Runx1:mCherry)-positive cells. Runx1 expression has been consistently used not only in zebrafish22, but also in mouse to mark HSPCs and hemogenic endothelial cells59. Similar to cd41:GFP+ cells, we uncovered a progressive increase in the number of Runx1:mCherry+ hematopoietic cells attached to the endocardium starting from 56 hpf (Fig. S5A, B). Imaging of Tg(cd41:GFP) and Tg(Runx1:mCherry) expression in the same animal revealed that most cardiac-residing hematopoietic cells were positive for both low GFP and mCherry expression (Fig. S5C, D), consistent with previous co-expression studies of these HSPC-marking transgenes22. We also performed fluorescence in situ hybridization for myb60, another HSPC marker that was also enriched in the EHC of our single-cell sequencing dataset (Fig. 1C, S2D), and observed its expression in the EdCs in the outer curvature of the ventricle (Fig. S5E). Based on their gene expression profiles and cell morphology, we conclude that HSPCs/MEPs are stably maintained on the endocardium.

We then assessed whether the heart supports HSPC proliferation, which is an essential function of a hematopoietic niche, by performing an EdU pulse labeling assay. Between 50 and 74 hpf, when we observed the highest growth of endocardial-residing hematopoietic cells, the majority (57%) of cd41:GFP+ cells associated with the heart were EdU-positive (Fig. 1K, L). Between 74 and 98 hpf, the proportion decreased to 31% of cd41:GFP+ cells (Fig. 1L). These results suggest that hematopoietic cells can proliferate on the luminal surface of the endocardium.

Some endocardial cells can undergo EHT

The appearance of endocardial-residing hematopoietic cells raised the question as to whether these cells arise de novo from EdCs or migrate from the DA. A significant limitation we encountered was a lack of EdC- or DA-specific transgenic lines. Thus, we used a photoconversion-based approach for cell lineage tracing. We utilized the Kaede protein that turns from green to red upon excitation with UV wavelength53 and when expressed specifically in endothelial cells ((Tg(fli1a:Gal4); Tg(UAS:Kaede), hereafter referred to as Tg(fli:Kaede)), enables us to track a spatially confined group of endothelial cells.

We first photoconverted the entire endocardium (Fig. 2A, A’) at 22 hpf, when the heart is not yet fully contractile and the atrium has not yet connected to the cardinal vein. We grew the photoconverted animals for two days and imaged the hearts at 74 hpf, the stage when we observed many endocardial-residing hematopoietic cells. We indeed found photoconverted Kaede-red cells that appear to be budding off from the endocardium at 74 (Fig. 2B, B’), 96 (Fig. S6A), and 120 (Fig. S6B) hpf, suggesting that EdCs give rise to hematopoietic cells that persist in the heart. Interestingly, when we examined other hematopoietic niches, we found photoconverted Kaede-red cells in the CHT (Fig. 2C), the intermediary niche that allows for HSPC population expansion, as well as the thymus (Fig. 2D), where adaptive immune cells reside and proliferate. Based on the Kaede-red/Kaede-green color ratio, we quantified photoconverted cells that reside in non-photoconverted tissues. We observed approximately 16.7 and 0.92 endocardial-derived cells residing in the CHT (Fig. 2E) and the thymus (Fig. 2F), respectively. Endocardial-derived Kaede-red cells were found in all the CHTs we imaged, but only in 57% (8 out of 14) of the thymuses we imaged. The low numbers of endocardial-derived cells in the thymus compared with the CHT suggest that most endocardial-derived hematopoietic cells are not fated to become immune cells. Our data suggest that the endocardium gives rise to cells that not only remain attached to the endocardium, but also circulate and eventually reside in extracardiac hematopoietic niches.Fig. 2 Endocardial cells contribute to cardiac-residing hematopoietic cells.

A–B’ Confocal Z-stack projection (A) and single confocal plane (A’) of photoconverted endocardium at 22 hpf, and follow-up imaging of the same animals (B, confocal Z-stack projection; B’, single confocal plane) 52 h post-conversion. Extruding Kaede-red+ cells are observed on the endocardium at 74 hpf (white asterisk). C, D Confocal image of 74 hpf fli:Kaede+ caudal hematopoietic tissue (CHT; C) and thymus (D) from larvae with their endocardium photoconverted at 22 hpf. Kaede-red+ cells (white asterisk) surrounded by endothelial cells are visible in the CHT (C); few Kaede-red+ cells visible in the thymus (white asterisks) (D). E, F Quantification of Kaede-red+ cells from the photoconverted endocardium (22 hpf) reveals 16.7 and 0.92 endocardial-derived cells in the CHT (74 hpf, E) and thymus (74 hpf, F), respectively, 52 h post-conversion. n = 14 CHT and 14 thymi. G–H’ Single confocal plane of a heart with photoconverted endocardium in the ventricular outer curvature (G, G’) or the atrium (H, H’) at 48 hpf and at 48 h post-conversion (G’–H”). G”, H” Closeup single plane of the 96 hpf fli:Kaede+ heart, where EdCs in the ventricular outer curvature (G”) or the atrium (H”) of the same animals had been photoconverted at 48 hpf. G” Kaede-red+ cells observed on the ventricular endocardium (white asterisks) near Kaede-green+ cells (red asterisks). H” Kaede-red+ cells derived from the atrium observed close to the atrial and ventricular endocardium (white asterisks). Kaede-green+ cells (red asterisks) are also present in the ventricle. I, J Quantification of Kaede-red+ cells derived from the ventricular outer curvature (I) or the atrium (J) found in other regions of the heart and the CHT, 48 h post-conversion. Several Kaede-red+ cells from the ventricular outer curvature found in the CHT, but not in other heart regions. Atrial-derived Kaede-red+ cells are present in the CHT and ventricular outer curvature. n = 14 hearts with photoconverted ventricular outer curvature and 14 hearts with photoconverted atrium, respectively. One-way ANOVA with Tukey’s multiple comparison test (I, J) was used. Scale bars, 30 µm (A–D, G–H”). Source data are provided as a Source Data file.

As our analysis showed that most hematopoietic cells were found in the outer curvature of the cardiac ventricle (Fig. 1H), we investigated whether EdCs possess heterogeneous, region-specific capability to give rise to hematopoietic cells. We photoconverted different regions of 48 hpf hearts, which exhibit spatially distinct regions: the atrium, the ventricular outer curvature (VOC), and ventricular inner curvature (VIC). We then grew the animals for 24 or 48 h and quantified Kaede-red cells in the other regions of the heart as well as in the CHT at 72 and 96 hpf. Interestingly, we observed Kaede-red cells in the VOC not only when cells from the same region were photoconverted (Fig. 2G–G”, S7A, white asterisks), but also when atrial EdCs were photoconverted (Fig. 2H–H”, S7E, white asterisks; average of 1.9 cells). Whereas VOC-derived cells were not frequently found in other cardiac regions (Fig. 2I; 6 out of 14 (43%) hearts at 96 hpf), atrial-derived Kaede-red cells were attached to the VOC (Fig. 2H’, H”, J; 9 out 14 (64%) hearts at 96 hpf). Quantification of endocardial-derived cells revealed that the VOC and the atrium contribute to approximately the same numbers of cells in the CHT (Fig. 2I, J, S7B–D, S7F–H; 3.4 and 2.2 cells at 72 hpf, 9.1 and 7.8 cells at 96 hpf from the VOC and atrium, respectively). The VIC contributed almost no cells to the other regions of the heart (Fig. S8A-C, E), and fewer cells than the other cardiac regions to the CHT particularly at 72 hpf (Fig. S8D, E; average of 0.5 cells at 72 hpf and 6.6 cells at 96 hpf). Thus, our data suggest that the endocardium in the VOC and atrium have the highest propensity to differentiate into hematopoietic cells that remain in the heart or integrate into other niches.

To further test the potential of EdCs to autonomously generate hematopoietic cells in isolation, we dissected and cultured the hearts ex vivo for 24 h. We adopted an established protocol for culturing embryonic hearts61, which were extracted from 48 hpf embryos positive for Tg(cd41:GFP) and/or Tg(Runx1:mCherry) expression (Fig. 3A). At this stage, the heart consists of only the endocardium and myocardium45. Imaging the hearts immediately post dissection showed that only a minority of them contained hematopoietic cells at 48 hpf (Fig. 3B, C; 2 out of 13 hearts were cd41:GFP+ (15%); 3 out of 12 hearts were Runx1:mCherry+ (25%)). We grew the 48 hpf hearts in Dulbecco’s media supplemented with fetal bovine serum for 24 h, and selected the ones that were still contracting in the media for further analysis. Strikingly, the majority of the hearts grown ex vivo for 24 h contained cd41:GFP+ (16 out of 17 hearts (94%)) or Runx1:mCherry+ (15 out of 19 hearts (79%)) cells (Fig. 3B, C). These results further indicate that between 48 and 72 hpf, EdCs can give rise to HSPCs without contribution from the DA or other extra-cardiac sources.Fig. 3 Ex vivo transplants and longitudinal light sheet imaging of the beating heart show endocardial cells activating HSPC markers.

A Schematics of ex vivo growing of hearts extracted at 48 hpf and cultured for 24 h. B Representative confocal images of a 48 hpf extracted Tg(Runx1:mCherry), Tg(cd41:GFP), Tg(kdrl:BFP-CAAX) heart, showing no Runx1:mCherry+ or cd41:GFP+ cells. After culturing for 24 h, Runx1:mCherry+ and cd41:GFP+ cells are present in the heart. C Significantly higher numbers of Runx1:mCherry+ and cd41:GFP+ cells are observed in hearts at the end of the 24 h culture compared with 48 hpf extracted hearts. n = 13, 17, 12, and 19 hearts, from left to right. D Schematics of a 72 hpf zebrafish larva at the illumination axes of the light sheet microscope. Movie stack acquisitions (one movie per plane) were used to reconstruct a synchronization of the beating heart every 30 min for 7 h. E–J Single plane of an endocardial cell in the ventricular outer curvature (E–G, magenta, white arrow) or in the ventricular inner curvature (H–J, magenta, white arrow). At 0 min (E, H), no Tg(cd41:GFP) expression is observed. At 30 min (F, I), Tg(cd41:GFP) expression appears noticeable (F); at 60 min (G, J), strong Tg(cd41:GFP) expression is visible in endocardial cells that remain integrated in the heart. The hearts are outlined in dashed lines. C Unpaired, two-tailed t test was used. Scale bars, 30 µm (B, E–J). Source data are provided as a Source Data file. Figure A, D created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license.

Furthermore, using a custom-built light sheet microscope, we recorded high-speed movies of the beating heart (Fig. 3D, Supplementary Movie 5). We followed individual EdCs over a 7-hour period starting from 72 hpf in the beating hearts of Tg(cd41:GFP)+, Tg(kdrl:nls-mCherry)+ larvae (Fig. 3E–J, Supplementary Movie 5). From the reconstructed images, we identified EdCs that did not initially express Tg(cd41:GFP) at the start of imaging but then gradually turned on Tg(cd41:GFP) expression (Fig. 3E–J, yellow asterisk in Supplementary Movie 5). These EdCs were found in several regions of the cardiac ventricle and remained integrated within the endocardium by the end of the imaging period. Together, our results suggest that some EdCs can undergo EHT, activate HSPC/MEP markers, and contribute to hematopoietic populations in the heart and other niches.

Dorsal aorta endothelial cells are a major contributor to endocardial-residing hematopoietic population

We next investigated the contribution of endothelial cells from the DA to cardiac-residing cells by photoconverting them and imaging the hearts 48 h post-conversion. We narrowed down specific DA regions that might contribute cells to the heart and converted the anterior-most DA region (Fig. 4A), the middle DA region above the yolk extension (Fig. 4D), and the posterior-most DA region (Fig. 4G). Interestingly, we found that many cells derived from the DA end up residing in the endocardium. While the yolk extension DA region gave rise to some cells (Fig. 4E, F, M; average of 2.25 cells at 74 hpf and 4.5 cells at 98 hpf), the posterior DA region yielded the most cardiac-residing cells (Fig. 4H, I, M; average of 6 cells at 74 hpf and 8 cells at 98 hpf). In contrast, the anterior DA region contributed almost no cells to the heart (Fig. 4B, C, M; average of <1 cell at 74 and 98 hpf). Photoconverting the middle and posterior DA regions at 22 hpf (Fig. 4J) yielded an average of 9.75 DA-derived cells attached to the endocardium at 74 hpf (Fig. 4K, M), which amounted to 42% of the total cd41:GFP+ cells in 74 hpf hearts (average of 23.1 cells). At 98 hpf, the number of DA-derived Kaede-red cells attached to the endocardium significantly increased to 21 cells (Fig. 4L, M), which represented 63% of the total cd41:GFP+ cells (average of 33 cells). These results indicate that DA-derived cells attach and significantly expand in numbers once attached to the endocardium, which could be driven by proliferation or capture of more circulating cells on the endocardium.Fig. 4 Dorsal aorta endothelial cells are major contributors to cardiac-residing hematopoietic cells.

A, D, G, J Representative confocal images of Kaede green-to-red photoconversion of Tg(fli1:Kaede)+ endothelial cells in the anterior-most region (A), middle region (D), posterior-most region (G), or combined middle and posterior regions (J) of the dorsal aorta (DA) at 22 hpf. B–C’ Z-stack projections and single confocal planes of the heart at 74 (B, B’) and 98 (C, C’) hpf show minimal to no contribution of anterior DA-derived Kaede-red+ cells to cardiac-residing cells. E–F’ Z-Stack projections and single confocal planes of the heart at 74 (E, E’) and 98 (F, F’) hpf show moderate contribution of the middle DA region to the heart (white asterisks). H–I’ Z-Stack projections and single confocal planes of the heart at 74 and 98 hpf demonstrate a significant contribution of posterior DA cells to the heart. K–L’ Photoconversion of endothelial cells in the middle and the posterior DA (combined) regions at 22 hpf lead to the highest numbers of DA-derived cells attached to the endocardium (white asterisks). M Quantification of DA-derived Kaede-red+ cells attached to the endocardium at 74 and 98 hpf. The posterior DA subregion contributes the highest numbers of cardiac-residing cells. Photoconverting endothelial cells in both the middle and posterior DA (combined) regions yields the most cardiac-residing cells. n = 14 (anterior), 24 (middle), 23 (posterior), and 12 (combined) hearts at 74 hpf, and 14 (anterior), 13 (middle), 16 (posterior), and 15 (combined) hearts at 96 hpf. A one-way ANOVA with Tukey’s multiple comparison test was used. Scale bars, 30 µm (A–L). Source data are provided as a Source Data file.

We also found that photoconverting the DA at 38 hpf still led to the appearance of photoconverted cells in the heart (Fig. 5G, H), indicating that DA endothelial cells retain their capacity to undergo EHT at these later stages. In 74 hpf hearts, we found similar numbers of Kaede-red cells derived from the DA endothelial cells photoconverted at 38 hpf compared with ones photoconverted at 22 hpf (Fig. 5M; 7.8 cells from 38-hpf DA, 9.75 cells from 22 hpf DA), suggesting that most or all DA endothelial cells retain their EHT capacity and attachment capability to the endocardium at 38 hpf. Altogether, these results show that many cardiac-residing hematopoietic cells derive from the DA, particularly the posterior-most region of the DA.Fig. 5 HSPC attachment to the endocardium is dependent on Itga4-Vcam1b ligand-receptor interaction.

A Violin plots from the endocardial single-cell RNAseq dataset show that itga4 expression is enriched in the hematopoietic endocardial population (Cluster 5), whereas vcam1b expression is elevated in other endocardial cell populations. B RNAscope in situ hybridization for vcam1b mRNA in the 74 hpf zebrafish heart. vcam1b appears to be strongly expressed in the EdCs of the ventricular outer curvature (B’, labeled with Tg(kdrl:GFP) expression, white arrows), but appears undetectable in the valves (B”). C Representative confocal images of control, itga4−/−, and vcam1b−/− Tg(Runx1:mCherry), Tg(kdrl:GFP) hearts at 74 hpf. White asterisks indicate Runx1:mCherry+ cells attached to the endocardium. D Significantly fewer Runx1:mCherry+ cells were found in itga4−/− and vcam1b−/− hearts compared with control. n = 12 control, 20 itga4−/−, and 9 vcam1b−/− hearts. E Representative confocal images of Tg(cd41:GFP), Tg(kdrl:nls-mCherry) hearts in control, itga4, and vcam1b morphants at 74 hpf. F Quantification of cd41:GFP+ cells on the endocardium reveals a significant reduction in itga4 or vcam1b morphants compared with control. n = 38 control, 14 itga4, and 18 vcam1b morphant hearts. G, I, K Representative confocal images of photoconverted endothelial cells in the middle and the posterior regions of the DA at 38 hpf in control (G), itga4 morphants (I), and vcam1b morphants (K). H Photoconverted Kaede-red+ cells (white asterisks) are attached to the endocardium in 74 hpf control hearts. J, L Noticeably lower numbers of Kaede-red+ cells from the DA are present in 74 hpf hearts of itga4 (J) or vcam1b (L) morphants. M DA-derived Kaede-red+ cell numbers in the heart are significantly reduced in itga4 and vcam1b morphant larvae compared with control. n = 13 control, 14 itga4, and 12 vcam1b morphant hearts. D, F, M One-way ANOVA with Tukey’s multiple comparison test was used. Scale bars, 30 µm (B, C, E, G–L). Source data are provided as a Source Data file.

Hematopoietic cells attach to the endocardium via an Integrin α4-Vcam1-dependent adhesion mechanism

We hypothesized that an active mechanism in which adhesion molecules mediate endocardial-hematopoietic attachment was necessary to withstand the high-flow intracardiac environment. One essential ligand-receptor pair that mediates HSPC homing into their niche is the ligand Vcam1 (Vascular cell adhesion molecule 1), which is mainly present on the endothelial cell surface of HSPC niches, and its receptor Itgα4 (Integrin α4), which is present on the surface of HSPCs62. In zebrafish, loss of either itga4 or vcam1b does not impact EHT in the DA but does result in the failure of HSPCs to seed the CHT63.

Our single-cell transcriptome dataset showed that itga4 expression was present at high levels in the EHC (Fig. 5A; Cluster 5). vcam1b expression is high in other EdCs that might be hematopoietic niche cells (Fig. 5A; Cluster 1-4, 8) but not in the EHC itself (Fig. 5A; Cluster 5) or in the valve cell population (Fig. 5A; Cluster 6). Using an RNAscope assay to detect vcam1b mRNA in the heart, we found that within the endocardium, vcam1b appeared enriched in the EdCs in the VOC (Fig. 5B, B’). In contrast, vcam1b did not appear detectable in the valve EdCs (Fig. 5B, B”). The high vcam1b expression in ventricular EdCs suggests that the endocardium exhibits similarities in molecular signatures with endothelial cells in HSPC niches and might explain the preferential adhesion of hematopoietic cells to EdCs in the VOC but not in the valve.

We tested the function of Vcam1 and Itgα4 in HSPC-endocardial attachment by quantifying Runx1:mCherry+ cells in itga4 and vcam1b mutants or cd41:GFP+ cells in animals with morpholino-mediated knockdown of itga4 or vcam1b. itga4 and vcam1b mutants and morphants do not appear severely defective in development, with wild-type like body lengths and cardiac morphology (Fig. S9A, B), as well as unaffected cardiac valve function (Supplementary Movies 6-8). Quantification of Runx1:mCherry+ cells in itga4 and vcam1b mutants compared with heterozygous siblings revealed an almost complete loss of Runx1:mCherry+ cells (Fig. 5C, D; asterisks indicate Runx1:mCherry+ cells in Fig. 5C). Similarly, significantly lower numbers of cd41:GFP+ cells were attached to the endocardium of itga4 and vcam1b morphants at 74 hpf (Fig. 5E, F). These results indicate that Itgα4-Vcam1b-based adhesion is necessary to maintain endocardial hematopoietic cells, and that the endocardium may share some niche characteristics with the CHT.

We then determined whether the endocardium fails to mediate attachment of circulating hematopoietic cells from the DA upon loss of itga4 or vcam1b function, similar to what was reported for the CHT vasculature. We photoconverted the DA endothelial cells in control, itga4, and vcam1b morphants at 38 hpf (Fig. 5G, I, K), and quantified endocardial-residing hematopoietic cells at 74 hpf (Fig. 5H, J, L). We found a substantial reduction of DA-derived hematopoietic cells on the endocardium when itga4 or vcam1b was knocked down (Fig. 5G–M; averages of 2.5 and 3.5 cells in itga4 and vcam1b morphants, respectively, compared with 7.8 cells in control). These results indicate that Itga4 and Vcam1b promote the attachment of circulating hematopoietic cells to the endocardium.

Hematopoietic cell attachment to the endocardium is at least partly independent of cardiomyocyte trabeculation and reduced shear stress

Our results narrowed the time window when endocardial-residing hematopoietic cells appear in the heart to be between 56 and 72 hpf. During this period, trabecular cardiomyocytes delaminate and form ridges in the cardiac lumen64. These ridges are computationally predicted to produce pockets of low shear stress65, which led us to hypothesize that they allow the hematopoietic cells to attach to the endocardium. To test this hypothesis, we used two hypotrabeculation models: nrg2a mutants, which fail to activate ErbB266, and the ErBB2 pharmacological inhibitor PD168393 (PD)67. In both models, the endocardium and myocardium remained single-layered without trabecular ridges (Fig. 6A–C). However, at 74 hpf, cd41:GFP+ cells remained attached to the endocardium in both hypotrabeculation models (Fig. 6A–C). Quantification of cd41:GFP+ cells shows no significant reduction in the number of endocardial-attached HSPCs (Fig. 6D, E), and no changes in HSPC position within the heart of nrg2a mutants (Fig. S10A, B), indicating that cardiac trabeculation is not necessary for hematopoietic cells to attach to the endocardium.Fig. 6 HSPC attachment to the endocardium is at least partly independent of cardiac trabeculation and contraction-induced biomechanical forces.

A–C Representative confocal image of 74 hpf Tg(cd41:GFP), Tg(kdrl:nls-mCherry) hearts. In nrg2a−/− mutants or larvae treated with the ErBB2 inhibitor PD168393 (PD), both leading to hypotrabeculation, cd41:GFP+ cells remain attached to the endocardium. D, E Quantitative analysis reveals no significant difference in endocardial-attached cd41:GFP+ cells in both models of hypotrabeculated hearts, PD drug treatment (D) or nrg2a mutants (E), compared with control. D n = 12 control, 18 PD-treated hearts; (E) 7 control, 10 nrg2a−/− hearts. F–I Suboptimal concentrations of MS-222 (G), tnnt2a morpholino (H), or Blebbistatin (I) were used to reduce cardiac contractions compared with control (F). No noticeable changes in cd41:GFP+ cell numbers were observed. J Quantifications show no significant changes in cd41:GFPlow+/HSPC/MEP numbers upon reduction of cardiac contraction. n = 23 control, 13 tnnt2a morphant, 6 Blebbistatin-treated, and 14 MS222-treated hearts. Two-way ANOVA with Sidak’s multiple comparison test (D, E) or One-way ANOVA with Tukey’s multiple comparison test (J) was used. Scale bars, 30 µm (A–C, F–I). Source data are provided as a Source Data file.

Although the endocardium is similar to the CHT in its dependence on Vcam1-Itgα4 for HSPC lodgment, it exhibits apparent differences in tissue organization. Unlike the fenestrated CHT vasculature22, the endocardium is constantly exposed to blood flow. Combined with our observations that cardiac hypotrabeculation did not affect HSPC/MEP numbers in the heart, we hypothesized that hematopoietic cell attachment to the endocardium was at least partly independent of mechanical forces. Stopping the heartbeat in tnnt2a loss-of-function morphants, or with high concentrations of the anesthetic Blebbistatin, led to failure of the heart to balloon and grow, which unsurprisingly led to the almost-complete absence of HSPCs in the heart (Fig. S11A, B). To mitigate secondary defects from severe cardiac problems, we used a low concentration of anesthetics from 48 to 74 hpf (MS-222/Tricaine or Blebbistatin) or a low dose of tnnt2a morpholino to reduce the heartbeat by around 30% (Fig. S11C). Consistent with our hypothesis, the numbers of endocardial-residing cd41:GFPlow+ cells were not significantly altered upon reduced cardiac contraction and reduced shear stress (Fig. 6F–J). These data suggest that the attachment of hematopoietic cells to the endocardium is at least partly independent of biomechanical forces from the blood flow or heartbeat.

Abrogation of primitive erythropoiesis leads to increased cardiac-residing HSPCs/MEPs

The emergence of hematopoietic cells on the endocardium after 56 hpf suggests that this process is driven by the definitive and not primitive hematopoiesis. We investigated whether abrogating primitive erythropoiesis affects HSPC/MEP appearance in the heart by knocking down gata1, a gene encoding a transcriptional promoter of primitive erythropoiesis68. When the gata1 morpholino is injected at sub-optimal concentrations, gata1 morphants lack red blood cells for only the first 96 h of development65 (own observations). Surprisingly, gata1 knockdown led to a significantly increased number of endocardial-attached cd41:GFPlow+ cells (Fig. 7A, B). Analysis of gata1 morphant hearts at 48 and 54 hpf shows that cd41:GFPlow+ cells did not emerge or attach earlier than in control embryos (Fig. S12A-E), but instead significantly increased at 74 hpf when cd41:GFPlow+ cells were consistently found in almost all wild-type hearts (Fig. 7D; Fig. S12E).Fig. 7 Perturbing primitive erythropoiesis enhances the presence of cardiac-residing hematopoietic cells independently of blood flow.

A–C Representative confocal images of 74 hpf control (A) and gata1 morphant (B) Tg(cd41:GFP), Tg(kdrl:nls-mCherry) hearts, as well as gata1 morphant Tg(cd41:GFP), Tg(kdrl:nls-mCherry) hearts treated with MS-222 (C). Noticeable increase of cd41:GFP+ cells present in gata1 morphant hearts treated with DMSO (B) or MS-222 (C). D Quantification of cd41:GFP+ cells in untreated or MS222-treated control and gata1 morphants. Untreated gata1 morphant hearts exhibited significantly increased cd41:GFPlow+ cell numbers compared with control, whereas MS-222 treated gata1 morphant hearts exhibited a trend towards increased numbers of cd41:GFPlow+ cell numbers. n = 17 control, 15 gata1 morphant, and 13 MS222-treated gata1 morphant hearts. E, F Representative confocal images of control and gata1 morphant Tg(cd41:GFP), Tg(kdrl:nls-mCherry) CHTs at 74 hpf. G In contrast with the heart, total numbers of cd41:GFP+ cells did not change in the CHT upon loss of primitive erythrocytes in gata1 morphants. n = 11 control and 17 gata1 morphant CHTs. H, I Quantification of cd41:GFPlow+ and gata1:DsRed+ cell numbers show a significant increase in the total numbers and proportions of double-positive cells upon gata1 knockdown. (H) n = 12 control, 9 gata1 morphant hearts; (I) n = 11 control, 9 gata1 morphant hearts. J, K Representative confocal images of control and gata1 morphant Tg(cd41:GFP), Tg(gata1:DsRed), Tg(kdrl:BFP-CAAX) hearts at 74 hpf. Asterisks indicate cells positive for both Tg(cd41:GFP), Tg(gata1:DsRed) expression. Noticeably higher numbers of double-positive cells were observed in gata1 morphant hearts. One-way ANOVA with Tukey’s multiple comparison test (D), two-way ANOVA with Sidak’s multiple comparison test (G, H) or unpaired, two-tailed t test (I) was used. Scale bars = 30 µm (A–C, J, K), 100 µm (E, F). Source data are provided as a Source Data file.

Erythrocytes constitute a significant source of shear stress in the vascular lumen, which promotes HSPC formation in the zebrafish DA69 and the mouse aorta-gonad-mesonephros70.Although our results suggest that HSPC/MEP emergence in the heart is partly independent of shear stress, we tested whether the increase of cd41:GFP+ cells in gata1 morphants was due to altered biomechanical forces using a low concentration of MS-222. We still observed a trend of increasing numbers of endocardial-residing cd41:GFP+ cells in gata1 morphants with reduced heartbeat (Fig. 7C, D), suggesting that the effects of abrogating primitive erythrocytes on HSPC/MEP presence in the heart may be independent of erythrocyte-induced biomechanical forces.

We further tested whether abrogating primitive erythrocytes induced a global response from other vascular tissues to produce or maintain an increased number of HSPCs. We quantified cd41:GFP+ cells in the CHT at 74 hpf in control (Fig. 7E) and gata1 morphants (Fig. 7F). Interestingly, we did not observe any significant changes in the numbers of cd41:GFP+ cells (low or high GFP signal) in the CHT (Fig. 7G). These results show that abrogating primitive erythrocytes does not lead to a non-specific global increase in HSPCs but does influence the numbers of cardiac-residing HSPCs/MEPs.

We then investigated whether the absence of primitive erythrocytes could influence the differentiation state of the increased HSPCs/MEPs in the heart. We combined Tg(cd41:GFP) expression to mark HSPCs and Tg(gata1:DsRed) expression to mark erythrocytes in the same larvae, imaged them in control and gata1 morphants, and quantified the proportion of double-positive cells in both conditions (Fig. 7H–K). We only considered gata1:DsRed+ cells that exhibited a round morphology and not the elongated mature erythrocyte morphology. In control hearts, a significant percentage (~50%; Fig. 7I) of cd41:GFP+ cells were also positive for Tg(gata1:DsRed) (Fig. 7H–J), suggesting that these HSPCs/MEPs are naturally primed to become erythrocytes. This proportion of cd41:GFP+; gata1:DsRed+ cells significantly increased in gata1 morphant hearts (average of 70%; Fig. 7H, I, K), suggesting that the lack of circulating primitive erythrocytes not only increased endocardial-associated HSPCs but also promoted their differentiation state into erythrocytes. Altogether, these data suggest that when circulating primitive erythrocytes are lost in the first few days of development, attachment of more HSPCs/MEPs primed to differentiate into erythrocytes to the endocardium was induced.

Discussion

Our results provide evidence that the zebrafish endocardium can sustain a population of HSPCs/MEPs for at least the first 10 days of development (Fig. 8). Our data are consistent with the previous report by Nakano et al.38 that mouse EdCs in the outflow tract give rise to cells positive for the HSPC marker CD4, suggesting that a spatially confined subset of EdCs undergo EHT. However, important distinctions exist between our findings in zebrafish and the reports on mouse endocardial hematopoietic cells. First, whereas HSPCs/MEPs are only temporarily found in mouse hearts between E8.5 and E10.5, these cells persist and multiply in the zebrafish heart (Fig. 1D–J). Second, in mouse, HSPCs appear in the endocardium even earlier (E8.5) than the AGM (E10.5); we found that in zebrafish, hematopoietic cells appear later in the heart (~72 hpf) compared with the DA (~24 hpf) (Fig. 1D, E, J). Loss of Integrin α4 and Vcam1, which does not affect EHT in the DA but affects HSPC incorporation into the CHT63, significantly reduces hematopoietic cell attachment to the endocardium, suggesting that the heart acts more as a secondary niche than a primary hematopoietic production tissue. Finally, other reports note that one of the major cells derived from EdCs is macrophages, which are particularly important for valve remodeling. We did not observe hematopoietic cell attachment to the valves, and only a minor population of macrophages in our transcriptomic dataset. Whether zebrafish EdCs generate macrophages during later development or in adults, and whether the endocardial-derived hematopoietic cells differentiate into macrophages upon valve injury remain interesting questions. In mouse, limitations with Cre lineage tracing have also called the endocardial contribution into hematopoietic lineages into question43,44. In contrast, our study utilizes imaging of live animals and clearly showed the persistence of the HSPCs/MEPs in the beating zebrafish heart (Supplementary Movies 1–4).Fig. 8 Schematic model showing the heart as a resident tissue for and a contributor of hematopoietic stem and progenitor cells.

The endocardium in red and the myocardium in gray. Endocardial cells undergo EHT (flat magenta cells), leading to the formation of HSPCs (round magenta cells in the endocardium). Additional HSPCs originating from the DA, particularly in the posterior DA region, attach to and stably integrate into the endocardium (green round cells). HSPCs originating from the endocardium and the DA contribute to the HSPC population in the CHT. Figure 8 created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license.

A notable limitation of our study pertains to a lack of endocardial-specific Cre line in zebrafish for definitive lineage tracing of endocardial-derived hematopoietic cells. We utilized the photoconvertible Kaede line to the best of our technical abilities and uncovered the spatiotemporal contribution of endothelial cells in the heart and the posterior DA region to cardiac-residing cells. However, we could not combine other transgenic lines labeling hematopoietic lineages with photoconverted Kaede-red+/Kaede-green animals because of color restrictions in fluorescence or perform immunostaining with photoconverted Kaede-red+ animals due to strong background fluorescence of fixed samples. Based on the single-cell sequencing data and the low contributions of endocardial-derived cells in the thymus, the resident tissue for immune cells, our results suggest that endocardial-residing/-derived HSPCs might be biased towards the erythrocyte/platelet fate. With the current limitations, we could not distinguish whether environmental signals from the endocardium result in a lineage bias of any hematopoietic cells residing in the heart, or whether hematopoietic cells derived from the endocardium are intrinsically fated to become erythrocytes or platelets. Future investigations with improved lineage tracing tools are needed to address this question, and to uncover the precise hematopoietic identities of endocardial-derived cells. In addition, the establishment of endocardial-specific lineage tracing tools will help to investigate the capability of the endocardium to support long-term quiescence and proliferation of HSPCs throughout the zebrafish lifetime.

Although the endocardium appears to exhibit some characteristics suggestive of an HSPC niche, such as the prevalence of EdU+ HSPCs attached to the endocardium (Fig. 1K, L) and the high vcam1b expression in the EdCs that retain HSPCs (Fig. 5B–B”), it is also important to note several significant differences between known HSPC niches, particularly the CHT, and the endocardium. The endocardium is constantly under mechanical pressure from cardiac contraction and blood flow. In contrast, the endothelial cells in the CHT are fenestrated, and form pockets for the HSPCs protected against blood flow. In addition, whereas the CHT contains stromal cells that provide signaling to the HSPCs, the endocardium does not possess stromal cell populations, and is mostly in contact with the myocardium. Whether the myocardial tissue provides signaling cues to maintain the hematopoietic cells is unknown, but the myocardial trabecular network does not appear to provide structural support for the HSPCs (Fig. 6A–E). Lastly, our data indicate that, unlike the heart, the CHT does not respond to the lack of primitive erythrocytes caused by gata1 knockdown (Fig. 7A–K), suggesting that the endocardium and the CHT play distinct functions under different physiological conditions. Thus, the endocardium provides a unique model to study how hematopoietic cells attach and proliferate under constant pressure. We are currently investigating molecular similarities and differences between the endocardium and the CHT.

Studies of EHT in the DA indicate that HSPC budding depends on blood flow-induced shear stress69. Similarly, multiple developmental processes during cardiac development are regulated by proper cardiac contraction/blood flow-induced biomechanical forces64,71–75. Interestingly, we found that reducing cardiac contraction did not significantly affect the attachment of hematopoietic cells to the endocardium (Fig. 6F–J), suggesting that hematopoietic cell maintenance in the heart is at least partly independent of shear stress. This feature could provide one of the distinct characteristics between endocardial- and DA-derived HSPC production and might point to the role of the endocardium as a reserve hematopoietic tissue that is not influenced by mechanical forces.

Finally, whether endocardial-derived hematopoietic cells play distinct functions compared with hematopoietic cells from the DA or other vascular sources remains to be investigated. Our results suggest that the endocardium can give rise to a minor population of hematopoietic cells (~10% of Tg(cd41:GFP)+ cells in the CHT) that might have different cellular characteristics than other endothelial sources. Interestingly, we observed a higher number of endocardial-residing hematopoietic cells when primitive erythropoiesis is abrogated, suggesting that the endocardium might compensate for the loss of red blood cells by producing HSPCs/MEPs or promoting their proliferation to contribute to circulation. In addition, the endocardial-residing hematopoietic cells may play a local function during homeostasis or injury. Consistent with this hypothesis, Shigeta et al.39. have reported that the endocardial-derived valve macrophages exhibit higher phagocytic activities, which is important to clear cellular debris in the tissue. In conclusion, we provide evidence that the endocardium contributes to hematopoiesis as both a tissue source and a niche, providing a unique model to study HSPCs under mechanical pressure.

Methods

Zebrafish handling and lines

Zebrafish were handled at the University of Münster, where the ethical guidelines outlined by the state of North Rhine-Westphalia were adhered to and the handling was conducted under the supervision of the veterinary authorities of the city of Münster, at the Max Planck Institute for Heart and Lung Research as approved by the Animal Protection Committee of the Regierungspräsidium Darmstadt, or at the University of Wisconsin-Madison in accordance with their Institutional Animal Care and Use Committee guidelines. All procedures performed on animals conform to the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals for scientific purposes.

Embryos were raised in a solution composed of 0.33 Danieau’s medium, containing the following concentrations: 17.4 mM NaCl, 0.21 mM KCl, 0.12 mM MgSO4·H2O, and 0.18 mM Ca(NO3)2, while being maintained at a stable temperature of 28 °C. The following strains of zebrafish were maintained under standard conditions as previously described76: Tg(kdrl:nls-mCherry)is4Tg,77, Tg(kdrl:GFP)s843,56, Tg(kdrl:BFP-CAAX)mu293Tg,78, Tg(gata1:DsRed)sd2,79, Tg(mpx:GFP)i114Tg,80, Tg(itga2b/cd41:GFP)la2,79, Tg(Runx1:mCherry)22, Tg(fli1a:Gal4)ubs4; Tg(UAS:Kaede)rk8Tg,81, and nrg2amn0237Gt,66.

The itga4 and vcam1b mutant lines were established at the University of Wisconsin-Madison via injection of CRISPR-Cas9 ribonucleoprotein complex in single-cell stage embryos. The following sgRNA for itga4 and vcam1b, respectively, was used: 5’-TAAAAGCTGGAATTGGCCAC-3’ and 5’-GCGCGGCAGCTCCAGCGTGT-3’. This created a 13 base-pair deletion in exon 2 of the itga4 gene resulting in a frameshift mutation, with the deleted sequence as follows: 5’-CCTGTGGCCAATT-3’, and a 7 base-pair deletion in the exon 4 of the vcam1b gene resulting in an early stop codon at amino acid 34, with the deleted sequences as follows: 5’-AATTTAA-3’. itga4 and vcam1b mutant carriers were genotyped by PCR amplification and Sanger sequencing. The mutant embryos were raised in a solution composed of 1x E3 medium containing 5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, and 0.33 mM MgSO4 and supplemented with methylene blue to prevent fungal growth, while being maintained at a stable temperature of 28 °C.

Confocal microscopy and image analysis

Confocal microscopes were employed for capturing images of arrested hearts. Embryos and larvae, up to 120 hpf, were embedded in 1% low-melting agarose supplemented with 0.2% Tricaine for immobilization. 7 and 10 dpf larvae were fixed in 4% paraformaldehyde for 2 h at room temperature prior to imaging, and then mounted in 1% low-melting agarose. Except for itga4 and vcam1b mutants, images of stopped hearts and the CHT were acquired using a Zeiss LSM710 or Zeiss LSM880 confocal laser microscope with a 20x or 40x magnification dipping lens. For imaging stopped hearts in itga4 and vcam1b mutant embryos, images were acquired using a Nikon A1 HD25 confocal laser microscope with a 25x magnification dipping lens set at 1.6x zoom. Maximal intensity projections were uniformly generated with identical settings applied to all samples. Image processing and subsequent quantitative analyses were conducted using Fiji (National Institutes of Health, USA) software tools and Imaris (Bitplane, UK). Brightness and contrast were adjusted with Imaris and Fiji. For imaging beating hearts, videos were acquired using a Zeiss Cell Observer spinning disk microscope at 100 frames per second with the 40X water immersion lens.

Cardiac light sheet imaging and analysis

72 hpf Tg(cd41:GFP)+, Tg(kdrl:nls-mCherry)+ zebrafish embryos were mounted in 1.5% low-melting agarose (#A9414; Sigma) in FEP tubes (Pro Liquid). The inner and outer diameters of the FEP tubes were 0.8 mm and 1.2 mm, respectively. To immobilize the embryos while maintaining its physiological heartbeat, tricaine (MS-222) was supplemented to the agarose and the imaging chamber at a final concentration of 133 mg/l. This setup was used for time-lapse imaging in a custom-built light sheet microscope, featuring a Nikon 16×0.8 NA (CFI75 LWD 16X W) detection objective for high-resolution imaging. The system included a Hamamatsu Image Splitter (#A12801-01; W-View GEMINI) for simultaneous dual-channel fluorescence imaging and a pco.edge 4.2 CLHS sCMOS camera configured to capture images at 200 frames per second (fps) with a 5 ms exposure time per frame. Emission filters used were a BP525/25 bandpass filter for the green channel and an LP600 long-pass filter for the red channel. Excitation of fluorescent probes was conducted with a TOPTICA iChrome CLE laser system. Movie stacks of 10 z-planes in the central location of the heart were recorded for 1.5 s in each plane with 1 µm z-spacing. Time-lapse movies with a 30-minute interval were recorded for 7 h. Recorded movie stacks were synchronized as previously described82. Subsequently, cells expressing Tg(cd41:GFP) were manually visualized and tracked throughout the dataset.

Localization analysis of cd41:GFP+ cells in cardiac regions

For the precise determination of cd41:GFP+ cell localization within cardiac tissues and their respective regions, Fiji software (National Institutes of Health, USA) was employed. A manual inspection of each z-stack derived from confocal images was undertaken. Subsequently, the heart was segmented into eight regions of interest, ensuring consistency across all analyzed samples. GFP-expressing cells with low intensity were enumerated within each predefined area. The resulting data were visually represented as a heatmap, displaying the percentage of cells within each region relative to the total cd41:GFP+ cell count across all regions.

Analysis of endocardial and cd41:GFP+ surface area co-localization

A co-surface analysis was conducted to quantify the extent of co-localized tissue between the endocardium (Tg(kdrl:BFP-CAAX) expression) and cd41:GFP+ cell populations. Imaris software (Bitplane, UK) was utilized for this purpose, facilitating the creation of surface renderings for the endocardial and cd41:GFP+ cell populations. Subsequently, the shared surface area in µm³ and percentage of the total surface area was determined.

Cell collection and flow cytometry

To isolate EdCs for the single-cell RNA sequencing experiment, hearts from 48- and 72-hpf Tg(kdrl:nls-mCherry)+; Tg(cd41:GFP)+ larvae were manually dissected in DMEM + 10% FBS. Hearts were then centrifuged for 60 s at 3000 rpm, washed with 1 mL Hanks’ Balanced Salt Solution and dissociated into single cells by incubating in 100 μl Enzyme 1 and 5 μl Enzyme 2 (Pierce Cardiomyocyte Dissociation Kit, Thermo Fisher Scientific, Cat#88281) for 20 min at 300 rpm in a 30 °C shaker83. The samples were centrifuged for 5 min at 3000 rpm, the supernatant was discarded, and fresh medium was added to the dissociated cells and passed through 40 μM-filter polystyrene 5 ml tubes. Negative controls of non-fluorescent hearts or single-color fluorescent hearts were prepared to define the sorting gates. Cells were sorted using the BD FACSAria™ III (BD Biosciences), equipped with a 100 μM nozzle tip and an instrument pressure setting of 20 psi. Live cells were selected by exclusion of DAPI positive signal using a 30 mW 405 nm excitation paired with 450/50 nm band pass filter. Single positive Tg(kdrl:nls-mCherry)+ cells and double-positive Tg(kdrl:nls-mCherry)+; Tg(cd41:GFP)+ cells were sorted and collected into tubes with DMEM + 10% FBS. Cells from both collected samples were immediately combined and processed for the single-cell RNA sequencing. mCherry fluorescence was measured with 50 mW 561 nm excitation paired with 610/20 nm band pass filter. GFP fluorescence was measured with 50 mW 488 nm excitation paired with 530/30 band pass filter. Sorted cell data was recorded and analyzed on BD FACSDiva v8.0.1 software.

Single-cell RNA sequencing

For single-cell RNA sequencing, FACS-sorted cell suspensions were counted with Moxi cell counter and diluted according to the manufacturer’s protocol to obtain a total of 2,660 cells at 48 hpf and 4,125 cells at 72 hpf. Each sample was run separately on a lane in Chromium controller with Chromium Next GEM Single Cell 3ʹ Reagent Kits v3,1 (10xGenomics). scRNA-seq library preparation was done using standard protocol, and sequencing was done on Nextseq2000. scRNA-seq analysis was performed as previously described84 and final data visualization was done by CellxGene package (doi:10.5281/zenodo.3235020).

The RNA-sequencing data generated in this study have been deposited in the NCBI GEO database under accession code GSE269378.

Photoconversion of endothelial cells

Tg(fli:Gal4)+, Tg(UAS:Kaede)+ larvae with the photoconvertible protein Kaede specifically expressed in endothelial cells were used for photoconversion. Specific vascular regions, which were the endocardium or the DA, were selected for photoconversion using the Zeiss LSM710 confocal microscope. The photoconversion was accomplished through five scans of the 405-nanometer wavelength laser at 100% power. Z-stacks of Kaede-green and Kaede-red channels with the larval hearts before and after photoconversion were obtained to ensure spatially restricted photoconversion. Within the DA, we defined the anterior portion as the anterior-most part of the DA until just before the first intersegmental vessel (ISV) sprout (Fig. 4A), the “middle region” as the DA region adjacent to the first 12 ISV sprouts (Fig. 4D), and the “posterior region” as the posterior-most DA region from the 13th ISV sprout until the tail (Fig. 4G). Subsequently, the photoconverted embryos were retrieved and grown at 28 °C until subsequent imaging.

Morpholino-mediated knockdown

Knockdown experiments were conducted by injecting specific antisense oligonucleotide morpholinos (MO) acquired from Gene Tools, LLC. The MO solution, comprising a total volume of 1 nl, was injected into one-cell-stage embryos at the designated concentration:

gata1 MO68: 5′-CTGCAAGTGTAGTATTGAAGATGTC-3′, final concentration used was 400 µM.

itga4 MO63: 5′-CTCCAGTGAGTTGTGATGAAATCAT-3′, final concentration used was 100 µM.

vcam1b MO63: 5′-GTCAGCAAGAAAGCTGTATCCATGT-3′, final concentration used was 400 µM.

tnnt2a MO85: 5′-CATGTTTGCTCTGATCTGACACGC-3′, final concentration used was 100 µM for high and 10 µM for low concentration injections, respectively.

EdU labeling

The Click-iT reaction for EdU staining was executed per the manufacturer’s instructions (ThermoFisher #C10337). The samples were fixed in 4% paraformaldehyde, and incubated in Click-iT EdU Alexa Fluor 647 for 30 min. They were also incubated in anti-eGFP primary antibody (1:500 dilution, Thermo Fisher A-11122, polyclonal rabbit) overnight at 4 °C, and then Alexa Fluor-488 secondary antibody (1:500 dilution, Thermo Fisher A32723, goat secondary antibody) for 2 h at room temperature prior to imaging. Imaging was accomplished utilizing a Zeiss 710 confocal laser microscope, employing a 40x water immersion objective. Quantitative analysis of proliferative Tg(cd41:GFP)+ cells was conducted using the Imaris software (Bitplane, UK). Cells were classified as proliferative when they exhibited dual labeling with Tg(cd41:GFP) expression and EdU. The proportion of EdU-positive cells was calculated as the percentage of proliferative cells relative to the total count of Tg(cd41:GFP)+ cells.

Fluorescent in-situ hybridization of myb

The following primers were used to create the in-situ probes for myb:

Forward: GCTGCTAAAGTCAGCCCAAC

Reverse: TAATACGACTCACTATAGGGGTTTAATCGTGCCGACCACT

The PCR template was prepared using these primers and cDNA from 48 hpf embryos. The probe was then prepared by incubating the PCR template with the T7 RNA polymerase (Promega), 10X DIG NTP labeling mixture (Roche), transcription buffer (Promega), and RNase inhibitor (Promega) at 37 °C for 2 h and purified using the Zymo RNA Purification Kit. The purified RNA probe was then diluted in hybridization buffer in a concentration of 1 µg/µL. Fluorescent in situ hybridization was performed on 72 hpf Tg(kdrl:GFP)+ larvae. Larvae were fixed in 4% paraformaldehyde for 3 h at room temperature, then washed with RNase-free PBS/0.1%Tween for 1 h. They were permeabilized in serial methanol dilutions for 5 min each and incubated overnight in 100% methanol at -20 °C. They were rehydrated in PBS/Tween, incubated in 1 µg/mL Proteinase K for 30 min, and fixed again in 4% paraformaldehyde for 20 min. After washing with PBS/Tween, the larvae were incubated in hybridization buffer (Hyb buffer; 50% formamide, 5X SSC, 0.1% Tween-20, 50 µg/mL heparin, 500 µg/mL tRNA) at 65 °C for ≥2 h, and then in DIG-labeled RNA probe for myb (1 µg/mL in Hyb buffer) at 65 °C overnight. After serial dilutions with Hyb buffer/SSC, the larvae were incubated in anti-tyramide-Alexa 568 (Thermo Fisher Scientific) in TSA buffer (100 mM Borate pH 8.5, 0.1% Tween-20, 2% Dextran Sulfate, 0.003% H2O2, 450 µg/mL 4-iodophenol) for 30 min at room temperature. The larvae were subsequently incubated overnight at 4 °C in anti-eGFP primary antibody (1:500 dilution, Thermo Fisher A-11122, polyclonal rabbit), and then Alexa Fluor-488 secondary antibody (1:500 dilution, Thermo Fisher A32723, goat secondary antibody) for 2 h at room temperature prior to imaging.

RNAscope assay of vcam1b

Zebrafish embryos (Tg(kdrl:GFP)) at 73 h post-fertilization (hpf) were fixed overnight at 4 °C in 4% paraformaldehyde (PFA). Following fixation, embryos were rinsed with PBST (0.1% Tween-20 in PBS) and stored in 100% methanol at -20 °C. For rehydration, embryos underwent a series of ten 10-minute washes at room temperature (RT) with decreasing concentrations of methanol (75%, 50%, and 25%) in PBST, followed by a 10-minute wash in PBST. The RNAscope™ Multiplex Fluorescent Reagent Kit v2 (Advanced Cell Diagnostics, Cat. No. 323100) was used according to the manufacturer’s instructions with several modifications. Embryos were incubated in 3% H2O2 for 10 min at RT, followed by washes with PBST, and then treated with RNAscope Protease III (Cat. No. 322337) for 5 min at RT and subsequently washed three times with 0.01% PBST (0.01% Tween-20 in PBS). A prehybridization step was included by incubating the embryos at 40 °C for 2 h in RNAscope probe diluent (Cat. No. 300041). The vcam1b-C3 probe (Cat. No. 445681-C3) was then applied, and embryos were incubated overnight at 40 °C. All subsequent wash steps were performed with 0.01% Tween-20 in 0.2X SSCT, three times for 15 min each. Post-probe incubation, a postfix treatment with 4% PFA was conducted for 20 min at RT. All 40 °C incubation steps were performed in a water bath. Following the RNAscope protocol, embryos were washed three times with PBST for 5 min each. Blocking was performed for 1 h in a blocking solution (2% BSA in PBST). Embryos were then incubated overnight at 4 °C with a chicken polyclonal anti-GFP antibody (1:500; Invitrogen, Cat. No. A10262). After primary antibody incubation, embryos were washed six times for 30 min each with PBST. Secondary antibody incubation was carried out overnight at 4 °C using a goat anti-chicken Alexa Fluor 488 antibody (1:1000; Invitrogen, Cat. No. A11039). Post-secondary antibody incubation, embryos were washed six times for 30 min each with PBST. Nuclear staining was performed using DAPI. Before imaging, embryos were washed with 0.01% Tween-20 in PBS and mounted in 1% low melting point agarose (Invitrogen, Cat. No. 16520) in PBS for confocal imaging.

Ex vivo cardiac culture

Cardiac extraction followed an established large-scale extraction protocol as described86. Leibovitz’s L1-5 culture medium was enriched with 10% fetal bovine serum by a previously outlined protocol61. All procedural steps were conducted at room temperature, and the culture medium was pre-warmed to 28.5 °C. Subsequently, hearts were flushed from the filtration system into a petri dish coated with agarose and cultured at 28.5 °C for 24 h. Before confocal imaging, hearts were fixed with 4% PFA for 15 min and mounted in 1% low-melting agarose.

Heartbeat quantification and reduction assay

To reduce heartbeat, we used low concentrations of tnnt2a MO, Tricaine (MS-222) or Blebbistatin treatment. For the tnnt2a knockdown, one-cell-stage embryos were injected with a low concentration of tnnt2a MO (0.1 ng/ul). Tricaine or Blebbistatin was added at a final concentration of 0.04% (w/v) or 1 µM, respectively, to the egg water from 48 to 74 hpf. To quantify their heart rates, live zebrafish embryos were observed under a stereomicroscope, and the heart rate was counted as the number of ventricular contractions per minute. Each condition was analyzed in triplicate with a minimum of 9 embryos per replicate.

Treatment with PD drug (ErbB2 inhibitor)

Tg(cd41:GFP), Tg(kdrl:nls-mCherry) embryos were dechorionated and incubated in egg water supplemented with 10 µM PD168393 (dissolved in DMSO) from 50 to 72 hpf at 28 °C, and then immediately imaged. Control embryos were incubated in egg water containing an equal amount of the respective solvent (DMSO).

Statistics and reproducibility

Unless otherwise specified, all experiments in this research study were conducted using a minimum of three independent biological replicates. Statistical analyses were conducted employing Graph Pad PRISM software. To compare two groups with normally distributed data, an unpaired two-tailed Student’s t-test was employed. For comparisons involving more than two groups, we utilized a one-way ANOVA with Tukey’s multiple comparison test; in the presence of two categorical independent variables, a two-way ANOVA with Sidak’s multiple comparison test was applied. A significance level of ≤ 0.05 was employed for all experiments (*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001).

Data from all experiments listed in Figs. 2–7, Figs. S2–8, S10–12 were obtained from at least three independent biological replicates. One exception is the vcam1b RNAscope assay in Fig. 5B, which was performed in two independent biological replicates using two different genotypes.

Schematic model illustration

The model in Fig. 8 was drawn by free hand in Adobe Illustrator, from a template of Biorender images (licensed to D.B.).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Movie 1

Supplementary Movie 2

Supplementary Movie 3

Supplementary Movie 4

Supplementary Movie 5

Supplementary Movie 6

Supplementary Movie 7

Supplementary Movie 8

Reporting Summary

Source data

Source Data

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-024-51920-7.

Acknowledgements

This work started in the laboratory of DYRS at the MPI-HLR in Bad Nauheim and, from 2022, continued in the laboratory of F.G. at the University of Münster. We thank Marlies Rossmann, Erez Raz, Arica Beisaw, Leah Biggs, Stefan Schulte-Merker, and Christian Mosimann for discussions and/or critical reading of the manuscript, and Sara Wickstrøm, Leah Biggs, and Radhan Ramadass for help with image analysis and quantification. Confocal imaging support was provided by the University of Wisconsin-Madison Optical Imaging Core. This work was supported by funds from the Add-on Fellowship of the Joachim Herz Foundation and the German National Academy of Sciences Leopoldina fellowship (LPDS 2021-01) to D.B., the Alexander von Humboldt Foundation (Humboldt professorship), and the DFG (Deutsches Forschungsgemeinschaft; Excellence Strategy Project EXC 2067/ 1-390729940) to J.H., the National Institutes of Health National Heart, Lung, and Blood Institute (R01HL174965, R01HL142998, R56HL142998), an American Society of Hematology Bridge Grant Award, and the Department of Cell and Regenerative Biology (University of Wisconsin-Madison) to O.J.T., the Max Planck Society and the European Research Council (AdG 101021349-TAaGC) to D.Y.R.S., and the DFG (Projects SFB1348B12 and SF1450N04), the Cells-in-Motion Interfaculty Center and the Faculty of Medicine (University of Münster; the Innovative Medical Research grant (I-GU122208)) to F.G.

Author contributions

F.G. initiated the study in the group of D.Y.R.S. D.B. and F.G. developed the project and designed the research basis. D.B., A.V.H., L.S., N.M.W., and F.G. performed the majority of the experiments. P.M., A.G., A.A., O.S.L., and M.W. assisted with experiments and data analysis. K.K. provided technical assistance with the flow cytometry, and S.G. assisted with the single-cell RNA sequencing run and analysis. J.H., O.J.T., D.Y.R.S., and F.G. supervised and acquired funding for the project. F.G. and D.B. wrote the first draft of the manuscript with substantial input from D.Y.R.S. and O.J.T. All authors edited subsequent drafts of the manuscript.

Peer review

Peer review information

Nature Communications thanks Bin Zhou and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

Open Access funding enabled and organized by Projekt DEAL.

Data availability

The RNA-sequencing data generated in this study have been deposited in the NCBI GEO database under accession code GSE269378. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Amulya V. Hejjaji, Lena Steuter
==== Refs
References

1. Chen MJ Erythroid/myeloid progenitors and hematopoietic stem cells originate from distinct populations of endothelial cells Cell Stem Cell 2011 9 541 552 10.1016/j.stem.2011.10.003 22136929
Chen, M. J. et al. Erythroid/myeloid progenitors and hematopoietic stem cells originate from distinct populations of endothelial cells. Cell Stem Cell 9, 541–552 (2011).22136929 10.1016/j.stem.2011.10.003
2. Bertrand, J. Y. et al. Haematopoietic stem cells derive directly from aortic endothelium during development. Nature 464, 108–111 (2010).
3. Bertrand, J. Y. et al. Definitive hematopoiesis initiates through a committed erythromyeloid progenitor in the zebrafish embryo. Development 134, 4147–4156 (2007).
4. Medvinsky A Dzierzak E Definitive hematopoiesis is autonomously initiated by the AGM region Cell 1996 86 897 906 10.1016/S0092-8674(00)80165-8 8808625
Medvinsky, A. & Dzierzak, E. Definitive hematopoiesis is autonomously initiated by the AGM region. Cell 86, 897–906 (1996).8808625 10.1016/S0092-8674(00)80165-8
5. Matsuoka S Generation of definitive hematopoietic stem cells from murine early yolk sac and paraaortic splanchnopleures by aorta-gonad-mesonephros region–derived stromal cells Blood 2001 98 6 12 10.1182/blood.V98.1.6 11418454
Matsuoka, S. et al. Generation of definitive hematopoietic stem cells from murine early yolk sac and paraaortic splanchnopleures by aorta-gonad-mesonephros region–derived stromal cells. Blood 98, 6–12 (2001).11418454 10.1182/blood.V98.1.6
6. Canu G Ruhrberg C First blood: the endothelial origins of hematopoietic progenitors Angiogenesis 2021 24 199 211 10.1007/s10456-021-09783-9 33783643
Canu, G. & Ruhrberg, C. First blood: the endothelial origins of hematopoietic progenitors. Angiogenesis 24, 199–211 (2021).33783643 10.1007/s10456-021-09783-9
7. Heck AM Ishida T Hadland B Location, location, location: how vascular specialization influences hematopoietic fates during development Front. Cell Dev. Biol. 2020 8 1 21 10.3389/fcell.2020.602617 32117956
Heck, A. M., Ishida, T. & Hadland, B. Location, location, location: how vascular specialization influences hematopoietic fates during development. Front. Cell Dev. Biol. 8, 1–21 (2020).32117956 10.3389/fcell.2020.602617
8. Gritz E Hirschi KK Specification and function of hemogenic endothelium during embryogenesis Cell. Mol. Life Sci. 2016 73 1547 1567 10.1007/s00018-016-2134-0 26849156
Gritz, E. & Hirschi, K. K. Specification and function of hemogenic endothelium during embryogenesis. Cell. Mol. Life Sci. 73, 1547–1567 (2016).26849156 10.1007/s00018-016-2134-0
9. Ng, Y. Y., Baert, M. R. M., de Haas, E. F. E., Pike-Overzet, K. & Staal, F. J. T. In Genetic Modification of Hematopoietic Stem Cells: Methods and Protocols (ed. Baum, C.) 13–21 (Humana Press, 2009).
10. Rodriguez-Fraticelli AE Clonal analysis of lineage fate in native haematopoiesis Nature 2018 553 212 216 10.1038/nature25168 29323290
Rodriguez-Fraticelli, A. E. et al. Clonal analysis of lineage fate in native haematopoiesis. Nature 553, 212–216 (2018).29323290 10.1038/nature25168
11. Busch K Fundamental properties of unperturbed haematopoiesis from stem cells in vivo Nature 2015 518 542 546 10.1038/nature14242 25686605
Busch, K. et al. Fundamental properties of unperturbed haematopoiesis from stem cells in vivo. Nature 518, 542–546 (2015).25686605 10.1038/nature14242
12. Velten L Human haematopoietic stem cell lineage commitment is a continuous process Nat. Cell Biol. 2017 19 271 281 10.1038/ncb3493 28319093
Velten, L. et al. Human haematopoietic stem cell lineage commitment is a continuous process. Nat. Cell Biol. 19, 271–281 (2017).28319093 10.1038/ncb3493
13. Xavier-Ferrucio J Krause DS Concise review: bipotent megakaryocytic-erythroid progenitors: concepts and controversies Stem Cells 2018 36 1138 1145 10.1002/stem.2834 29658164
Xavier-Ferrucio, J. & Krause, D. S. Concise review: bipotent megakaryocytic-erythroid progenitors: concepts and controversies. Stem Cells 36, 1138–1145 (2018).29658164 10.1002/stem.2834
14. Kissa K Live imaging of emerging hematopoietic stem cells and early thymus colonization Blood 2008 111 1147 1156 10.1182/blood-2007-07-099499 17934068
Kissa, K. et al. Live imaging of emerging hematopoietic stem cells and early thymus colonization. Blood 111, 1147–1156 (2008).17934068 10.1182/blood-2007-07-099499
15. Lefrançais E The lung is a site of platelet biogenesis and a reservoir for haematopoietic progenitors Nature 2017 544 105 109 10.1038/nature21706 28329764
Lefrançais, E. et al. The lung is a site of platelet biogenesis and a reservoir for haematopoietic progenitors. Nature 544, 105–109 (2017).28329764 10.1038/nature21706
16. Yeung AK De novo hematopoiesis from the fetal lung Blood Adv. 2023 7 6898 6912 10.1182/bloodadvances.2022008347 37729429
Yeung, A. K. et al. De novo hematopoiesis from the fetal lung. Blood Adv. 7, 6898–6912 (2023).37729429 10.1182/bloodadvances.2022008347
17. Li Z Mouse embryonic head as a site for hematopoietic stem cell development Cell Stem Cell 2012 11 663 675 10.1016/j.stem.2012.07.004 23122290
Li, Z. et al. Mouse embryonic head as a site for hematopoietic stem cell development. Cell Stem Cell 11, 663–675 (2012).23122290 10.1016/j.stem.2012.07.004
18. Ding L Saunders TL Enikolopov G Morrison SJ Endothelial and perivascular cells maintain haematopoietic stem cells Nature 2012 481 457 462 10.1038/nature10783 22281595
Ding, L., Saunders, T. L., Enikolopov, G. & Morrison, S. J. Endothelial and perivascular cells maintain haematopoietic stem cells. Nature 481, 457–462 (2012).22281595 10.1038/nature10783
19. Kusumbe AP Ramasamy SK Adams RH Coupling of angiogenesis and osteogenesis by a specific vessel subtype in bone Nature 2014 507 323 328 10.1038/nature13145 24646994
Kusumbe, A. P., Ramasamy, S. K. & Adams, R. H. Coupling of angiogenesis and osteogenesis by a specific vessel subtype in bone. Nature 507, 323–328 (2014).24646994 10.1038/nature13145
20. Kobayashi H Angiocrine factors from Akt-activated endothelial cells balance self-renewal and differentiation of haematopoietic stem cells Nat. Cell Biol. 2010 12 1046 1056 10.1038/ncb2108 20972423
Kobayashi, H. et al. Angiocrine factors from Akt-activated endothelial cells balance self-renewal and differentiation of haematopoietic stem cells. Nat. Cell Biol. 12, 1046–1056 (2010).20972423 10.1038/ncb2108
21. Hooper AT Engraftment and reconstitution of hematopoiesis is dependent on VEGFR2-mediated regeneration of sinusoidal endothelial cells Cell Stem Cell 2009 4 263 274 10.1016/j.stem.2009.01.006 19265665
Hooper, A. T. et al. Engraftment and reconstitution of hematopoiesis is dependent on VEGFR2-mediated regeneration of sinusoidal endothelial cells. Cell Stem Cell 4, 263–274 (2009).19265665 10.1016/j.stem.2009.01.006
22. Tamplin OJ Hematopoietic stem cell arrival triggers dynamic remodeling of the perivascular niche Cell 2015 160 241 252 10.1016/j.cell.2014.12.032 25594182
Tamplin, O. J. et al. Hematopoietic stem cell arrival triggers dynamic remodeling of the perivascular niche. Cell 160, 241–252 (2015).25594182 10.1016/j.cell.2014.12.032
23. Kumar S Geiger H HSC niche biology and HSC expansion ex vivo Trends Mol. Med. 2017 23 799 819 10.1016/j.molmed.2017.07.003 28801069
Kumar, S. & Geiger, H. HSC niche biology and HSC expansion ex vivo. Trends Mol. Med. 23, 799–819 (2017).28801069 10.1016/j.molmed.2017.07.003
24. Blaser BW Zon LI Making HSCs in vitro: don’t forget the hemogenic endothelium Blood 2018 10.1182/blood-2018-04-784140 30089629
Blaser, B. W. & Zon, L. I. Making HSCs in vitro: don’t forget the hemogenic endothelium. Blood10.1182/blood-2018-04-784140 (2018).30089629 10.1182/blood-2018-04-784140
25. Bussmann J Bakkers J Schulte-Merker S Early endocardial morphogenesis requires Scl/Tal1 PLoS Genet. 2007 3 e140 10.1371/journal.pgen.0030140 17722983
Bussmann, J., Bakkers, J. & Schulte-Merker, S. Early endocardial morphogenesis requires Scl/Tal1. PLoS Genet. 3, e140 (2007).17722983 10.1371/journal.pgen.0030140
26. Van Handel B Scl represses cardiomyogenesis in prospective hemogenic endothelium and endocardium Cell 2012 150 590 605 10.1016/j.cell.2012.06.026 22863011
Van Handel, B. et al. Scl represses cardiomyogenesis in prospective hemogenic endothelium and endocardium. Cell 150, 590–605 (2012).22863011 10.1016/j.cell.2012.06.026
27. Miao Y Intrinsic endocardial defects contribute to hypoplastic left heart syndrome Cell Stem Cell 2020 27 574 589.e8 10.1016/j.stem.2020.07.015 32810435
Miao, Y. et al. Intrinsic endocardial defects contribute to hypoplastic left heart syndrome. Cell Stem Cell 27, 574–589.e8 (2020).32810435 10.1016/j.stem.2020.07.015
28. Mikryukov AA BMP10 signaling promotes the development of endocardial cells from human pluripotent stem cell-derived cardiovascular progenitors Cell Stem Cell 2021 28 96 111.e7 10.1016/j.stem.2020.10.003 33142114
Mikryukov, A. A. et al. BMP10 signaling promotes the development of endocardial cells from human pluripotent stem cell-derived cardiovascular progenitors. Cell Stem Cell 28, 96–111.e7 (2021).33142114 10.1016/j.stem.2020.10.003
29. Neri, T. et al. Human pre-valvular endocardial cells derived from pluripotent stem cells recapitulate cardiac pathophysiological valvulogenesis. Nat. Commun. 10, 1929 (2019).
30. Zhang, H., Lui, K. O. & Zhou, B. Endocardial cell plasticity in cardiac development, diseases and regeneration. Circ. Res. 122, 774–789 (2018).
31. Zhang, H. et al. Endocardium minimally contributes to coronary endothelium in the embryonic ventricular free walls. Circ. Res. 118, 1880–1893 (2016).
32. Red-Horse, K., Ueno, H., Weissman, I. L. & Krasnow, M. A. Coronary arteries form by developmental reprogramming of venous cells. Nature 464, 549–553 (2010).
33. Tang, J. et al. Extension of endocardium-derived vessels generate coronary arteries in neonates. Circ. Res. 130, 352–365 (2022).
34. Lu P Perinatal angiogenesis from pre-existing coronary vessels via DLL4–NOTCH1 signalling Nat. Cell Biol. 2021 23 967 977 10.1038/s41556-021-00747-1 34497373
Lu, P. et al. Perinatal angiogenesis from pre-existing coronary vessels via DLL4–NOTCH1 signalling. Nat. Cell Biol. 23, 967–977 (2021).34497373 10.1038/s41556-021-00747-1
35. Fioret BA Heimfeld JD Paik DT Hatzopoulos AK Endothelial cells contribute to generation of adult ventricular myocytes during cardiac homeostasis Cell Rep. 2014 8 229 241 10.1016/j.celrep.2014.06.004 25001281
Fioret, B. A., Heimfeld, J. D., Paik, D. T. & Hatzopoulos, A. K. Endothelial cells contribute to generation of adult ventricular myocytes during cardiac homeostasis. Cell Rep. 8, 229–241 (2014).25001281 10.1016/j.celrep.2014.06.004
36. Motoike T Markham DW Rossant J Sato TN Evidence for novel fate of Flk1+ progenitor: contribution to muscle lineage Genesis 2003 35 153 159 10.1002/gene.10175 12640619
Motoike, T., Markham, D. W., Rossant, J. & Sato, T. N. Evidence for novel fate of Flk1+ progenitor: contribution to muscle lineage. Genesis 35, 153–159 (2003).12640619 10.1002/gene.10175
37. Chen, Q. et al. Endothelial cells are progenitors of cardiac pericytes and vascular smooth muscle cells. Nat Commun. 7, 12422 (2016).
38. Nakano, H. et al. Haemogenic endocardium contributes to transient definitive haematopoiesis. Nat Commun. 4, 1564 (2013).
39. Shigeta, A. et al. Endocardially derived macrophages are essential for valvular remodeling. Dev. Cell. 48, p617–630.e3 (2019).
40. Hu Y Identification of cardiac hemo-vascular precursors and their requirement of sphingosine-1-phosphate receptor 1 for heart development Sci. Rep. 2017 7 45205 10.1038/srep45205 28338096
Hu, Y. et al. Identification of cardiac hemo-vascular precursors and their requirement of sphingosine-1-phosphate receptor 1 for heart development. Sci. Rep. 7, 45205 (2017).28338096 10.1038/srep45205
41. Liu N Notch and retinoic acid signals regulate macrophage formation from endocardium downstream of Nkx2-5 Nat. Commun. 2023 14 5398 10.1038/s41467-023-41039-6 37669937
Liu, N. et al. Notch and retinoic acid signals regulate macrophage formation from endocardium downstream of Nkx2-5. Nat. Commun. 14, 5398 (2023).37669937 10.1038/s41467-023-41039-6
42. Zamir L Nkx2.5 marks angioblasts that contribute to hemogenic endothelium of the endocardium and dorsal aorta. Alitalo K, ed Elife 2017 6 e20994 10.7554/eLife.20994 28271994
Zamir, L. et al. Nkx2.5 marks angioblasts that contribute to hemogenic endothelium of the endocardium and dorsal aorta. Alitalo K, ed. Elife 6, e20994 (2017).28271994 10.7554/eLife.20994
43. Liu K Lineage tracing clarifies the cellular origin of tissue-resident macrophages in the developing heart J. Cell Biol. 2022 221 e202108093 10.1083/jcb.202108093 35482005
Liu, K. et al. Lineage tracing clarifies the cellular origin of tissue-resident macrophages in the developing heart. J. Cell Biol. 221, e202108093 (2022).35482005 10.1083/jcb.202108093
44. Liu K Intercellular genetic tracing of cardiac endothelium in the developing heart Dev. Cell 2023 58 1502 1512.e3 10.1016/j.devcel.2023.05.021 37348503
Liu, K. et al. Intercellular genetic tracing of cardiac endothelium in the developing heart. Dev. Cell 58, 1502–1512.e3 (2023).37348503 10.1016/j.devcel.2023.05.021
45. Staudt D Stainier D Uncovering the molecular and cellular mechanisms of heart development using the zebrafish Annu. Rev. Genet. 2012 46 397 418 10.1146/annurev-genet-110711-155646 22974299
Staudt, D. & Stainier, D. Uncovering the molecular and cellular mechanisms of heart development using the zebrafish. Annu. Rev. Genet. 46, 397–418 (2012).22974299 10.1146/annurev-genet-110711-155646
46. Wattrus SJ Zon LI Stem cell safe harbor: the hematopoietic stem cell niche in zebrafish Blood Adv. 2018 2 3063 3069 10.1182/bloodadvances.2018021725 30425071
Wattrus, S. J. & Zon, L. I. Stem cell safe harbor: the hematopoietic stem cell niche in zebrafish. Blood Adv. 2, 3063–3069 (2018).30425071 10.1182/bloodadvances.2018021725
47. Kissa K Herbomel P Blood stem cells emerge from aortic endothelium by a novel type of cell transition Nature 2010 464 112 115 10.1038/nature08761 20154732
Kissa, K. & Herbomel, P. Blood stem cells emerge from aortic endothelium by a novel type of cell transition. Nature 464, 112–115 (2010).20154732 10.1038/nature08761
48. Murayama E Tracing hematopoietic precursor migration to successive hematopoietic organs during zebrafish development Immunity 2006 25 963 975 10.1016/j.immuni.2006.10.015 17157041
Murayama, E. et al. Tracing hematopoietic precursor migration to successive hematopoietic organs during zebrafish development. Immunity 25, 963–975 (2006).17157041 10.1016/j.immuni.2006.10.015
49. Gurung S Restrepo NK Sumanas S Endocardium gives rise to blood cells in zebrafish embryos Cell Rep. 2024 43 113736 10.1016/j.celrep.2024.113736 38308842
Gurung, S., Restrepo, N. K. & Sumanas, S. Endocardium gives rise to blood cells in zebrafish embryos. Cell Rep. 43, 113736 (2024).38308842 10.1016/j.celrep.2024.113736
50. Paolini A Abdelilah-Seyfried S ScienceDirect The mechanobiology of zebrafish cardiac valve leaflet formation Curr. Opin. Cell Biol. 2018 55 52 58 10.1016/j.ceb.2018.05.007 30007126
Paolini, A. & Abdelilah-Seyfried, S. ScienceDirect The mechanobiology of zebrafish cardiac valve leaflet formation. Curr. Opin. Cell Biol. 55, 52–58 (2018).30007126 10.1016/j.ceb.2018.05.007
51. Haack, T. & Abdelilah-Seyfried, S. The force within: endocardial development, mechanotransduction and signalling during cardiac morphogenesis. Development. 143, 373–386 (2016).
52. da Silva AR egr3 is a mechanosensitive transcription factor gene required for cardiac valve morphogenesis Sci. Adv. 2024 10 eadl0633 10.1126/sciadv.adl0633 38748804
da Silva, A. R. et al. egr3 is a mechanosensitive transcription factor gene required for cardiac valve morphogenesis. Sci. Adv. 10, eadl0633 (2024).38748804 10.1126/sciadv.adl0633
53. Gunawan, F., Gentile, A., Gauvrit, S., Stainier, D. & Bensimon-Brito, A. Nfatc1 promotes interstitial cell formation during cardiac valve development in zebrafish. Circ. Res. 126, 968–984 (2020).
54. Gunawan F Priya R Stainier DYR Sculpting the heart: Cellular mechanisms shaping valves and trabeculae Curr. Opin. Cell. Biol. 2021 73 26 34 10.1016/j.ceb.2021.04.009 34147705
Gunawan, F., Priya, R. & Stainier, D. Y. R. Sculpting the heart: Cellular mechanisms shaping valves and trabeculae. Curr. Opin. Cell. Biol. 73, 26–34 (2021).34147705 10.1016/j.ceb.2021.04.009
55. Gunawan, F. et al. Focal adhesions are essential to drive zebrafish heart valve morphogenesis. J. Cell Biol. 218, 1039–1054 (2019).
56. Beis D Genetic and cellular analyses of zebrafish atrioventricular cushion and valve development Development 2005 132 4193 4204 10.1242/dev.01970 16107477
Beis, D. et al. Genetic and cellular analyses of zebrafish atrioventricular cushion and valve development. Development 132, 4193–4204 (2005).16107477 10.1242/dev.01970
57. Xia J A single-cell resolution developmental atlas of hematopoietic stem and progenitor cell expansion in zebrafish Proc. Natl Acad. Sci. 2021 118 e2015748118 10.1073/pnas.2015748118 33785593
Xia, J. et al. A single-cell resolution developmental atlas of hematopoietic stem and progenitor cell expansion in zebrafish. Proc. Natl Acad. Sci. 118, e2015748118 (2021).33785593 10.1073/pnas.2015748118
58. Meng, P., Wu, L., Lin, Q. & Zhang, Y. Zebrafish for thrombocytopoiesis- and hemostasis-related researches and disorders. Blood Sci. 2, 44–49 (2020).
59. North TE Runx1 expression marks long-term repopulating hematopoietic stem cells in the midgestation mouse embryo Immunity 2002 16 661 672 10.1016/S1074-7613(02)00296-0 12049718
North, T. E. et al. Runx1 expression marks long-term repopulating hematopoietic stem cells in the midgestation mouse embryo. Immunity 16, 661–672 (2002).12049718 10.1016/S1074-7613(02)00296-0
60. Soza-Ried C Hess I Netuschil N Schorpp M Boehm T Essential role of c-myb in definitive hematopoiesis is evolutionarily conserved Proc. Natl Acad. Sci. 2010 107 17304 17308 10.1073/pnas.1004640107 20823231
Soza-Ried, C., Hess, I., Netuschil, N., Schorpp, M. & Boehm, T. Essential role of c-myb in definitive hematopoiesis is evolutionarily conserved. Proc. Natl Acad. Sci. 107, 17304–17308 (2010).20823231 10.1073/pnas.1004640107
61. Noël ES A Nodal-independent and tissue-intrinsic mechanism controls heart-looping chirality Nat. Commun. 2013 4 2754 10.1038/ncomms3754 24212328
Noël, E. S. et al. A Nodal-independent and tissue-intrinsic mechanism controls heart-looping chirality. Nat. Commun. 4, 2754 (2013).24212328 10.1038/ncomms3754
62. Imai K Selective transendothelial migration of hematopoietic progenitor cells: a role in homing of progenitor cells Blood 1999 93 149 156 10.1182/blood.V93.1.149 9864156
Imai, K. et al. Selective transendothelial migration of hematopoietic progenitor cells: a role in homing of progenitor cells. Blood 93, 149–156 (1999).9864156 10.1182/blood.V93.1.149
63. Li D VCAM-1+ macrophages guide the homing of HSPCs to a vascular niche Nature 2018 564 119 124 10.1038/s41586-018-0709-7 30455424
Li, D. et al. VCAM-1+ macrophages guide the homing of HSPCs to a vascular niche. Nature 564, 119–124 (2018).30455424 10.1038/s41586-018-0709-7
64. Priya R Tension heterogeneity directs form and fate to pattern the myocardial wall Nature 2020 588 130 134 10.1038/s41586-020-2946-9 33208950
Priya, R. et al. Tension heterogeneity directs form and fate to pattern the myocardial wall. Nature 588, 130–134 (2020).33208950 10.1038/s41586-020-2946-9
65. Vedula V A method to quantify mechanobiologic forces during zebrafish cardiac development using 4-D light sheet imaging and computational modeling PLoS Comput. Biol. 2017 13 1 24 10.1371/journal.pcbi.1005828
Vedula, V. et al. A method to quantify mechanobiologic forces during zebrafish cardiac development using 4-D light sheet imaging and computational modeling. PLoS Comput. Biol. 13, 1–24 (2017).10.1371/journal.pcbi.1005828
66. Rasouli, S. J. & Stainier, D. Y. R. Regulation of cardiomyocyte behavior in zebrafish trabeculation by Neuregulin 2a signaling. Nat. Commun. 8, 15281 (2017).
67. Liu, J. et al. A dual role for ErbB2 signaling in cardiac trabeculation. Development 137, 3867–3875 (2010).
68. Galloway JL Wingert RA Thisse C Thisse B Zon LI Loss of Gata1 but not Gata2 converts erythropoiesis to myelopoiesis in zebrafish embryos Dev. Cell 2005 8 109 116 10.1016/j.devcel.2004.12.001 15621534
Galloway, J. L., Wingert, R. A., Thisse, C., Thisse, B. & Zon, L. I. Loss of Gata1 but not Gata2 converts erythropoiesis to myelopoiesis in zebrafish embryos. Dev. Cell 8, 109–116 (2005).15621534 10.1016/j.devcel.2004.12.001
69. North TE Hematopoietic stem cell development is dependent on blood flow Cell 2009 137 736 748 10.1016/j.cell.2009.04.023 19450519
North, T. E. et al. Hematopoietic stem cell development is dependent on blood flow. Cell 137, 736–748 (2009).19450519 10.1016/j.cell.2009.04.023
70. Adamo L Biomechanical forces promote embryonic haematopoiesis Nature 2009 459 1131 1135 10.1038/nature08073 19440194
Adamo, L. et al. Biomechanical forces promote embryonic haematopoiesis. Nature 459, 1131–1135 (2009).19440194 10.1038/nature08073
71. Gentile A Mechanical forces remodel the cardiac extracellular matrix during zebrafish development Development 2024 151 dev202310 10.1242/dev.202310 38984541
Gentile, A. et al. Mechanical forces remodel the cardiac extracellular matrix during zebrafish development. Development 151, dev202310 (2024).38984541 10.1242/dev.202310
72. Gentile, A. et al. The EMT transcription factor Snai1 maintains myocardial wall integrity by repressing intermediate filament gene expression. Elife 10, e66143 (2021).
73. Fukui H Bioelectric signaling and the control of cardiac cell identity in response to mechanical forces Science 2021 374 351 354 10.1126/science.abc6229 34648325
Fukui, H. et al. Bioelectric signaling and the control of cardiac cell identity in response to mechanical forces. Science 374, 351–354 (2021).34648325 10.1126/science.abc6229
74. Bornhorst, D. et al. Biomechanical signaling within the developing zebrafish heart attunes endocardial growth to myocardial chamber dimensions. Nat. Commun. 10, 4113 (2019).
75. Fontana F Antagonistic activities of Vegfr3/Flt4 and Notch1b fine-tune mechanosensitive signaling during zebrafish cardiac valvulogenesis Cell Rep. 2020 32 107883 10.1016/j.celrep.2020.107883 32668254
Fontana, F. et al. Antagonistic activities of Vegfr3/Flt4 and Notch1b fine-tune mechanosensitive signaling during zebrafish cardiac valvulogenesis. Cell Rep. 32, 107883 (2020).32668254 10.1016/j.celrep.2020.107883
76. Westerfield M Doerry E Kirkpatrick AE Driever W Douglas SA An on-line database for zebrafish development and genetics research Semin. Cell Dev. Biol. 1997 8 477 488 10.1006/scdb.1997.0173 9441953
Westerfield, M., Doerry, E., Kirkpatrick, A. E., Driever, W. & Douglas, S. A. An on-line database for zebrafish development and genetics research. Semin. Cell Dev. Biol. 8, 477–488 (1997).9441953 10.1006/scdb.1997.0173
77. Wang, Y. et al. Moesin1 and Ve-cadherin are required in endothelial cells during in vivo tubulogenesis. Development 137, 3119–3128 (2010).
78. Matsuoka RL Radial glia regulate vascular patterning around the developing spinal cord. Bronner M, ed Elife 2016 5 e20253 10.7554/eLife.20253 27852438
Matsuoka, R. L. et al. Radial glia regulate vascular patterning around the developing spinal cord. Bronner M, ed. Elife 5, e20253 (2016).27852438 10.7554/eLife.20253
79. Traver D Transplantation and in vivo imaging of multilineage engraftment in zebrafish bloodless mutants Nat. Immunol. 2003 4 1238 1246 10.1038/ni1007 14608381
Traver, D. et al. Transplantation and in vivo imaging of multilineage engraftment in zebrafish bloodless mutants. Nat. Immunol. 4, 1238–1246 (2003).14608381 10.1038/ni1007
80. Mathias JR Resolution of inflammation by retrograde chemotaxis of neutrophils in transgenic zebrafish J. Leukoc. Biol. 2006 80 1281 1288 10.1189/jlb.0506346 16963624
Mathias, J. R. et al. Resolution of inflammation by retrograde chemotaxis of neutrophils in transgenic zebrafish. J. Leukoc. Biol. 80, 1281–1288 (2006).16963624 10.1189/jlb.0506346
81. Hatta, K., Tsujii, H. & Omura, T. Cell tracking using a photoconvertible fluorescent protein. Nat Protoc. 1, 960–967 (2006).
82. Mickoleit M High-resolution reconstruction of the beating zebrafish heart Nat. Methods 2014 11 919 922 10.1038/nmeth.3037 25042787
Mickoleit, M. et al. High-resolution reconstruction of the beating zebrafish heart. Nat. Methods 11, 919–922 (2014).25042787 10.1038/nmeth.3037
83. Boezio, G. L. M., et al. The developing epicardium regulates cardiac chamber morphogenesis by promoting cardiomyocyte growth. Dis, Model Mech. 16, dmm049571 (2022).
84. Mattonet K Endothelial versus pronephron fate decision is modulated by the transcription factors Cloche/Npas4l, Tal1, and Lmo2 Sci Adv 2023 8 eabn2082 10.1126/sciadv.abn2082
Mattonet, K. et al. Endothelial versus pronephron fate decision is modulated by the transcription factors Cloche/Npas4l, Tal1, and Lmo2. Sci Adv 8, eabn2082 (2023).10.1126/sciadv.abn2082
85. Sehnert, A. J. et al. Cardiac troponin T is essential in sarcomere assembly and cardiac contractility. Nat. Genet. 31, 106–110 (2002).
86. Lombardo, V. A. et al. Morphogenetic control of zebrafish cardiac looping by Bmp signaling. Development (Cambridge) 146, dev180091 (2019).
