Interplay among Etsrp/ER71, Scl, and Alk8 signaling controls endothelial and myeloid cell formation

Abstract

Vascular endothelial and myeloid cells have been proposed to originate from a common precursor cell, the hemangioblast. The mechanism of endothelial and myeloid cell specification and differentiation is poorly understood. We have previously described the endothelial-specific zebrafish Ets1-related protein (Etsrp), which was both necessary and sufficient to initiate vasculogenesis in the zebrafish embryos. Here we identify human Etv2/ER71 and mouse ER71 proteins as functional orthologs of Etsrp. Overexpression of mouse ER71 and Etsrp caused strong expansion of hemangioblast and vascular endothelial lineages in a zebrafish embryo. In addition, we show that etsrp is also required for the formation of myeloid but not erythroid cells. In the absence of etsrp function, the number of granulocytes and macrophages is greatly reduced. Etsrp overexpression causes expansion of both myeloid and vascular endothelial lineages. Analysis of mosaic embryos indicates that etsrp functions cell autonomously in inducing myeloid lineage. We further demonstrate that the choice of endothelial versus myeloid fate depends on a combinatorial effect of etsrp, scl, and alk8 genes.

Figure 1

Etsrp is a functional ortholog of the mammalian ER71 subfamily. (A) Alignment of zebrafish Etsrp, medaka, stickleback, mouse, and human ER71 amino acid sequences. Identical and similar amino acids are labeled in red and blue, respectively. Etsrp and hER71 share 71% homology within the ETS DNA-binding domain (underlined in gray). GenBank accession numbers used for the analysis are as follows: human ER71 (O00321), mouse ER71 (NP_031985), and zebrafish ETSRP (AAY89037). Medaka ER71 (ENSORLP00000019929) and stickleback ER71 (ENSGACP00000016315) are Ensembl predictions. (B) Phylogenetic analysis of zebrafish Etsrp and its closest human, mouse, frog, and fish homologs. The phylogenetic tree is built using the Neighbor Joining method. Length of horizontal branches is proportional to the evolutionary distance between the protein molecules. GenBank accession numbers used for the analysis are as follows: human ER71 (O00321), mouse ER71 (NP_031985), zebrafish ETSRP (AAY89037), human ETS1 (NP_005229), mouse ETS1 (NP_035938), frog ETS1 (NP_001081621), zebrafish ETS1a (NP_001017558), human ETS2 (NP_005230), mouse ETS2 (NP_035939), frog ETS2 (NP_001081007), zebrafish ETS2 (NP_001018874), human FLI1 (NP_002008), mouse FLI1 (NP_032052), zebrafish FLI1a (NP_571423), and zebrafish FLI1b (NP_001008780). Medaka ER71 (ENSORLP00000019929), stickleback ER71 (ENSGACP00000016315), fugu ETS1 (SINFRUP00000163510), and medaka ETS1 (ENSORLP00000016939) are Ensembl predictions. NTI Vector (Invitrogen) has been used to build the alignment and the phylogenetic tree. (C) Chromosomal location of the zebrafish etsrp, medaka, stickleback, mouse, and human ER71 genes. Numbers alongside the chromosomal regions of interest correspond to the actual physical distances (Mb). Etsrp/ER71 genes are highlighted in red. (D-J) Mouse ER71 and zebrafish Etsrp overexpression causes ectopic expression of hemangioblast marker scl and vascular endothelial marker flk1. flk1-GFP transgenic embryos were injected with 75 pg of either Etsrp or mER71 DNA at the 1-cell stage and analyzed at 8- to 10-somite stages. Relative to uninjected embryos (D,G), both Etsrp (E) and mER71 (F,H) result in the ectopic induction of scl when examined by in situ hybridization. Flk1 expression was also induced by ER71 injections as revealed by ectopic GFP expression (J) relative to uninjected controls (I). Panels G to J are lateral views with anterior to the left. Images were taken using Zeiss CV11 stereomicroscope, Axiocam color camera (Zeiss, model 412-312) and Openlab 4.0 software (Improvision, Waltham, MA). Magnification: 12× (D-F); 60× (G-J).

Figure 2

Etsrp is necessary and sufficient for the formation of myeloid cells. (A-F) Knockdown of Etsrp results in the nearly complete absence of myeloid cells as analyzed by in situ hybridization. Anterior is to the left. (A,C,E) Control uninjected embryos. (B,D,F) Etsrp morphants, injected with 12 ng to 15 ng of etsrp MO1 and etsrp MO2 in a 1:1 mixture. (A,B) L-plastin (lcp1) expression at 24 hours postfertilization (hpf). Note that lcp1-expressing macrophages are nearly completely absent in etsrp morphants (B). (C,D) mpx expression at 24 hpf. Note that mpx-expressing neutrophils are nearly completely absent in etsrp morphants (D). (E,F) pu.1 expression at the 16-somite stage. Embryos have been flat-mounted with the yolk removed. Note that the anterior myeloid-specific pu.1 expression is severely reduced in panel F, whereas posterior erythroid-specific expression is not significantly affected. (G-J) Etsrp RNA overexpression induces ectopic myeloid cell formation. (G,I) Control uninjected embryos. (H,J) etsrp RNA-overexpressing embryos. (G,H) pu.1 expression at the 16-somite stage, anterior view (G), ventro-lateral view (H). Note the strong expansion of pu.1-expressing cells, some of which are located ectopically (H, ). (I,J) lcp1 expression at 24 hpf. Note the increase in the number of lcp1-expressing macrophages in panel J. (K-N) Etsrp RNA with missing MO-binding sites can restore pu.1 expression in etsrp morphants. Embryos are at the 8-somite stage; (K,L) anterior view; (M,N) anterior-ventral view. (K) Control uninjected embryo; (L) 10 ng etsrp MO2-injected embryo; (M) 100 pg etsrp RNA-injected embryo; (N) embryo coinjected with 10 ng etsrp MO2 and 100 pg etsrp RNA. Images were taken using Axioskop2 and 10×/0.30 NA dry objective (Zeiss) (A-F; I,J) or CV11 stereomicroscope (Zeiss) (G,H,K-N), Axiocam color camera (Zeiss, model 412-312) and Openlab 4.0 software (Improvision). Magnification: 100× (A-D); 60× (E-N).

Figure 3

Etsrp affects both hematopoiesis and vasculogenesis in the anterior but not the posterior region. (A-D) Posterior ectopic pu.1 expression is independent of etsrp function as analyzed by in situ hybridization at 22 hpf. (A) Control uninjected embryo; (B) embryo injected with etsrp MOs; (C) gata1 MO-injected embryo; (D) gata1 MO and etsrp MO coinjected embryo. Note that ectopic pu.1 expressed in the erythroid cells in gata1 morphants (C) is unaffected in the double gata1/etsrp morphants (D). Anterior myeloid-specific pu.1 expression (, A,C) is missing in etsrp morphants (B,D). (E-H) Anterior ectopic gata1 expression is dependent on etsrp function as analyzed by the in situ hybridization at 20 hpf to 21 hpf. (E) Control uninjected embryo; (F) embryo injected with etsrp MOs; (G) pu.1 MO-injected embryo; (H) etsrp and pu.1 MOs coinjected embryo. Note that the ectopic myeloid-specific anterior gata1 expression in pu.1 morphants (, G) is absent in double etsrp/pu.1 morphants (H). Posterior erythroid gata1 expression is not affected in etsrp morphants. Images were taken using Axioskop2 and 5×/0.15 NA dry objective, Axiocam color camera (Zeiss, model 412-312) and Openlab 4.0 software (Improvision). Magnification: 75×.

Figure 4

Analysis of interaction between etsrp and scl and alk8 signaling pathways. (A-H) scl RNA can rescue myeloid but not vascular cell formation in etsrp morphants as evident from pu.1 and flk1 expression analysis at the 10-somite stage. (A-D) pu.1 expression, anterior view; (E-H) flk1 expression, dorsal view, anterior to the left. (A,E) Control uninjected embryos; (B,F) embryos injected with etsrp MOs; (C,G) scl RNA-injected embryos; (D,H) embryos coinjected with etsrp MOs and scl RNA. Note that the myeloid-specific pu.1 expression is restored in etsrp MO and scl RNA coinjected embryos (D). Also note that the vascular-specific flk1 expression is absent in etsrp MO and scl RNA coinjected embryos in the same experiment (H). (I-P) etsrp RNA can rescue vascular but not myeloid cell formation in scl morphants as evident from pu.1 and flk1 expression analysis at the 8-somite stage. (I-L) pu.1 expression as analyzed by in situ hybridization, anterior view; (M-P) GFP fluorescence in flk1-GFP transgenic embryos, dorsal view, anterior is to the left. (I,M) Control uninjected embryo; (J,N) scl MO-injected embryo; (K,O) etsrp RNA-injected embryo; (L,P) scl MO and etsrp RNA coinjected embryo. Note that etsrp RNA fails to rescue myeloid-specific pu.1 expression in scl morphants (L). Etsrp RNA can restore vascular flk1-GFP expression in the same experiment (P). Flk1-GFP fluorescence in wild-type embryos and scl morphants (M,N) is much weaker and not apparent under the same exposure. (Q-T) etsrp RNA fails to rescue pu.1 expression in alk8 morphants. pu.1 expression analyzed at the 10-somite stage, anterior view, except for panel S, which is anterior-ventral. (Q) Control uninjected embryo; (R) alk8 MO-injected embryo; (S) etsrp RNA-injected embryo; (T) alk8 MO and etsrp RNA coinjected embryo. (U-X) etsrp and constitutively active CA-alk8 RNA synergize in inducing pu.1 expression at the 14-somite stage. Anterior-ventral views except for panel X, which is the ventrolateral view, anterior is to the top. (U) Control uninjected embryo; (V) CA-alk8 RNA-injected embryo; (W) etsrp RNA-injected embryo; (X) CA-alk8 RNA and etsrp RNA coinjected embryo. Images were taken using CV11 stereomicroscope (Zeiss), Axiocam color camera (Zeiss, model 412-312) and Openlab 4.0 software (Improvision). Magnification: 60×.

Figure 5

Analysis of etsrp, pu.1, and scl expression by the 2-color in situ hybridization. Flat-mounted embryos, anterior is to the left. (A, B) Etsrp (red) and pu.1 (blue) expression in the anterior region of a flat-mounted embryo at the 6-somite stage. Panel B is a higher magnification of panel A. Note that pu.1-expressing cells lie immediately adjacent to etsrp-expressing cells but expression of the 2 markers does not overlap. (C,D) Etsrp (blue) and scl (red) expression in the anterior (C) and posterior (D) regions of a flat-mounted embryo at the 6-somite stage. Note that the 2 markers completely overlap in panel C while in panel D the trunk region contains only etsrp-expressing cells (); scl expression partially overlaps with etsrp in the tail region where scl is restricted to erythroid cells during later stages ( and the right ). (E,F) Etsrp RNA expression expands into the myeloid region in etsrp morphants injected with Etsrp translation-blocking MOs. Etsrp expression in control uninjected embryos (E) and etsrp morphants (F) at the 9-somite stage. Note the more intense and expanded etsrp expression in panel F. (G-J) scl RNA restores pu.1 expression in etsrp morphants with pu.1 and etsrp-expressing cells intermingled. Two-color in situ hybridization analysis for pu.1 (blue) and etsrp (red) expression at the 9- to 10-somite stage. Only the anterior part of an embryo is shown. (G) Control uninjected embryo; (H) etsrp MO-injected embryo; (I) scl RNA-injected embryo; (J) etsrp MOs and scl RNA coinjected embryo. Etsrp staining is very weak in the control embryos because of the short staining time, which was the same for all experimental batches. Note that etsrp morphants in panel H have absent pu.1 expression and strongly up-regulated and expanded etsrp expression. scl RNA-injected embryos (I) display up-regulated etsrp expression. pu.1-expressing cells are intermingled with etsrp-expressing cells in panel J but they do not overlap. Images were taken using Axioskop2 and 10×/0.30 NA (A; C-J) or 20×/0.50 NA (B) dry objectives (Zeiss), Axiocam color camera (Zeiss, model 412-312) and Openlab 4.0 software (Improvision). Magnification: 100× (A,C,D,G-J); 200× (B); 75× (E,F).

Figure 6

Etsrp-expressing precursor cells give rise to both vascular endothelial and myeloid lineages. (A) Diagram of transplantation experiment. Donor embryos were injected at the 1-cell stage with etsrp RNA and TRITC-dextran (B-D) or FITC-dextran (E-H), while recipient embryos were injected with 7.5 ng etsrp MO1/MO2 mixture. Cells were transplanted at the beginning of epiboly. (B-D) flk1-GFP–expressing cells are a subset of etsrp RNA/TRITC-labeled cells. The embryo is at the 8-somite stage, lateral view, anterior to the left. (B) TRITC-filter image. Only transplanted cells are visible. (C) GFP-filter image. Only flk1-GFP–expressing cells are visible. (D) Overlay of the TRITC, GFP, and transmitted light DIC images. Cells where GFP and TRITC fluorescence overlaps are in yellow ( point to some of these cells). Note that every GFP-expressing cell has also TRITC fluorescence. (E-H) pu.1-expressing cells originate from etsrp-expressing cells, transplanted from etsrp RNA-overexpressing embryos into etsrp morphants. Etsrp RNA was coinjected with fluorescein-labeled dextran; pu.1 expression and fluorescein presence was analyzed by 2-color fluorescent (E-G) or conventional (H) in situ hybridization at the 8- to 10-somite stages. (E-G) Anterior-lateral view of the same embryo, dorsal is up. (E) FITC-filter image. Only transplanted cells are visible. (F) pu.1 expression as detected by tyramide-Cy3 amplification, visualized through the rhodamine channel filter. Note the ectopically located pu.1-expressing cells () and 3 remaining endogenous pu.1-expressing cells that are located bilaterally within the anterior lateral mesoderm (). (G) Overlay of FITC, Cy3, and transmitted light images. Note that all 3 ectopic pu.1-expressing cells contain FITC label () while the endogenous pu.1 cells do not (). (H) A posterior region from an embryo containing multiple pu.1 and fluorescein-positive cells. Embryo has been flat-mounted to show dorsal, lateral, and ventral tissues. (1) Endogenous pu.1-expressing cells in the posterior lateral mesoderm. (2) Fluorescein-labeled transplanted cells. (3-5) Double pu.1 and fluorescein-positive cells. Average color for each cell group is shown in the boxes below the figure (“Methods”). Images were taken using Axioplan2 and 10×/0.30 NA (A;C-G) (Zeiss), Axiocam color camera (Zeiss, model 412-312) (H) or monochrome C4742-95 camera (B-G) (Hamamatsu Photonics, Hamamatsu City, Japan) and Openlab 4.0 software (Improvision). Magnification: 100× (B-G); 300× (H).

Figure 7

Proposed models for etsrp function within the anterior lateral mesoderm. (A) Etsrp induces scl in the hemangioblast cells, which give rise to both vascular endothelial and myeloid precursors. As hemangioblasts divide, etsrp expression becomes restricted to endothelial cells where it is both necessary and sufficient for vasculogenesis. (B) Alternatively, there are 2 separate pools of etsrp-expressing endothelial and myeloid precursors. As the cells initiate myeloid marker expression, they down-regulate etsrp expression.