FoxH1 negatively modulates flk1 gene expression and vascular formation in zebrafish

Abstract

Flk1 is the major receptor for VEGF on endothelial cells. During embryogenesis, flk1 is required for both vasculogenesis and angiogenesis and abnormally elevated flk1 expression is often associated with pathological conditions in adults. While the biological function of flk1 has been studied extensively, very little is known about how the flk1 gene is regulated at the transcriptional level. Our transgenic study led to the identification of a flk1 endothelial enhancer positioned approximately 5 kb upstream of the flk1 translation initiation site. Binding sites for FoxH1, scl, ets and gata factors are found in the zebrafish flk1 endothelial enhancer, as well as in upstream sequences of mouse flk1 and human kdr genes, suggesting that the regulatory machinery for flk1/kdr is conserved from fish to mammals. The roles of scl, ets and gata factors in hemangioblasts have been well defined, but the significance of FoxH1 in vessel formation has not been explored previously. Here we show that FoxH1 binds to the flk1 endothelial enhancer in vitro and functions as a repressor for flk1 transcription in cultured cells. Consistent with these findings, the expression level of flk1 is elevated in embryos lacking both maternal and zygotic FoxH1. We further show that overexpression of FoxH1 has a negative effect on vascular formation that can be counteracted by the down-regulation of smad2 activity in zebrafish embryos. Taken together, our data provide the first evidence that flk1 is a direct target of FoxH1 and that FoxH1 is involved in vessel formation in zebrafish.

Fig.1

GFP expression in TG(flk1:GFP)^la116 embryos. (A-D) GFP signals driven by the flk(6.4)-GFP transgene can be detected in TG(flk1:GFP)^la116 embryos as early as the 8-somite stage. flk1 expression is detected in three patches of cells by in situ hybridization (A). A similar pattern is seen in TG(flk1:GFP)^la116 embryos (B). Lateral views are shown in panels A and B and the dorsal views are shown in panels C and D. Anterior to the left. (E-H) GFP expression pattern of TG(flk1:GFP)^la116 embryos at 24hpf (F) resembles the flk1 pattern detected by in situ hybridization (E). Higher magnification images of the head and trunk are shown in G and H, respectively. (I-L) GFP expression pattern of TG(flk1:GFP)^la116 embryos at 48hpf (J) resembles the flk1 pattern detected by in situ hybridization (I). Higher magnification images of the head and trunk are shown in K and L, respectively. (M) Confocal image of GFP expression in endothelial cells in the brain and brachial arches in 3-day-old TG(flk1:GFP)^la116 embryos. (N-P) GFP expression remains active in adult TG(flk1:GFP)^la116 fish. Images show GFP signals in endothelial cells on the skin (N), in the gill (O) and gut (P).

Fig.2

Deletion analysis of the flk1 regulatory region identifies a highly conserved element that is necessary for endothelial expression. (A) Schematic diagram of the deletion constructs of flk1-GFP reporter. Linearized DNA of each construct was injected in wild type zebrafish embryos at the 1-cell stage. GFP expression in injected embryos was analyzed after 1 day of development. The transient endothelial expression directed by each construct is summarized by a plus (endothelial expression) or a minus (no detectable endothelial expression) to the right of the line representing each construct. * marks the transcription initiation site of flk1. (B-G) Transient GFP expression in endothelial cells of 1-day-old embryos injected with flk(−6.4)-GFP (B), flk(−5.0)-GFP (C), flk(−4.3)-GFP (D), flk(−5.0,−4.3/−1.5)-GFP (E), flk(−5.0,−3.5/−1.5)-GFP (F) or flk(−4.3,−3.5/−1.5)-GFP (G).

Fig.3

Transgenic analysis of the flk1 regulatory region identifies a 1.5kb minimal endothelial specific enhancer. (A-B) GFP expression patterns in the brain (A) and trunk (B) of 2-day-old TG(flk1:GFP)^la116 embryos. (C-D) GFP expression patterns of 2-dayold embryos carrying germ line integrated flk(−5.0,−3.5/1.5)-GPF transgene resembles the patterns observed in TG(flk1:GFP)^la116. (E-F) Embryos carrying germ line integrated flk(−4.3,−3.5/1.5)-GFP have GFP expression in endothelial cells (arrowhead) as well as neural tissues. Arrows point to GFP positive cells in forebrain in E and to GFP positive cells in neural tube in F.

Fig.4

Cross-species comparison of flk1 5'-flanking sequences. (A) Sequence of flk1 enhancer from −3990 to −3571. (B) Schematic diagrams of 5'-flanking regions of zebrafish, mouse, and human flk1 genes. Consensus sequences of FoxH1, scl, ets and GATA factor binding sites are highlighted in green, red, yellow and blue, respectively in (A) and represented with dots with the same color scheme in (B). Translation initiation sites are marked by arrows. The hatched bar indicates the zebrafish flk1 upstream sequence from −5.0kb to −3.5kb and the grey bar indicates the region from −5.0kb to −4.3kb.

Fig.5

FoxH1 regulates flk1 expression at the level of transcription. (A) FoxH1-GST fusion protein were purified and used in a gel shift assay with a radiolabeled probe (flk-FoxH1) corresponding to the flk1 enhancer. Unlabeled probe and unlabeled flk-FoxH1-mt, oligonucletides carrying a G to C substitution in the consensus sequences of the FoxH1 binding site were used as competitors in the gel shift assay at 50-, 100- or 200-fold molar excess. A representative experiment is shown. All assays were performed at least three times with comparable results. (B) HEK293T cells were transiently transfected with the indicated luciferase reporters and an expression vector encoding FoxH1. Values are relative to the luciferase activity of cells transfected with pGL3-flkP alone. Results of an average of four independent experiments are shown.

Fig.6

FoxH1 modulates vascular formation in zebrafish. (A) flk1 is expressed in the developing vasculature of embryos at 32hpf (left). The expression level of flk1 is elevated in MZsur embryos (right). (B) Injection of FoxH1 mRNA disrupts zebrafish vascular formation. Embryos were analyzed at the 18-somite stage. (C) Injection of mRNA encoding the FAST-Eng chimeric protein disrupts vascular formation (right), resembling the phenotype of FoxH1 overexpression. Embryos were analyzed at the 23-somite stage. (D) Injection of mRNA encoding FAST-VP16 chimeric protein does not disrupt the GFP expression pattern in TG(flk1:GFP)^la116 embryos. Embryos were analyzed at the 18-somite stage. Arrows point to patches of endothelial cells missing in FoxH1 or FAST-Eng mRNA injected embryo. (E) Lateral view of the un-injected (right), smad2MO and FoxH1 mRNA co-injected (middle) and smad2MO injected (left) TG(flk1:GFP)^la116 embryos at 24hpf. (F) Graph represents relative GFP intensities within the region of interests (indicated by the yellow and red boxes in panel E).