Loss of hedgehog signaling leads to vascular stability defects. (A–D) igu^fo10a mutants present with a curled body axis and hemorrhage (hem) at 53 hpf (B), as compared to wild type siblings (A). igu^tm79a/igu^fo10a compound heterozygous embryos (C) and igu^tm79a mutant embryos (D) both present with hemorrhage at 53 hpf. (E–G) While wild type siblings do not show any signs of hemorrhage at 48 hpf (E), hemorrhages are present in the head of igu^fo10a mutants (F and G). (H and I) Electropherograms of dzip1 sequence from wild type (H) and igu^fo10a mutants (I) showing the presence of a Q339X mutation. (J) Hemorrhage rate after treatment with cyclopamine for developmental windows. Solid bars indicate the treatment interval with the percent of embryos displaying hemorrhage at 53 hpf given on the right side. Red bars indicate a treatment interval that results in maximal hemorrhage rates, while blue bars indicate intervals that produce minimal hemorrhage. The yellow bar indicates DMSO controls treated for the entire interval. The shaded area represents the most sensitive interval to cyclopamine treatment. Scale bar in (A)–(G) is 500 μm.

igu^fo10a mutants have cell attachment and vessel caliber defects at 48 hpf. (A–D) TEM images of a single dorsal aortae in the head region of wild type (A and C) and igu^fo10a mutant embryos (B and D) at 48 hpf. In wild type embryos (A), the dorsal aorta is lateral to the notochord (nc) and is comprised of endothelial cells (ec) and a lumen (lum) tightly surrounded by perivascular support cells (*). Conversely, in igu^fo10a mutants (B) the perivascular cells make few contacts with endothelial cells. Tight junctions (arrows) are readily identified in both wild type (C) and igu^fo10a mutants (D). Boxes in (A) and (B) show area of magnification in (C) and (D). (E and F) Confocal Z-stack projections of wild type (E) and igu^fo10a mutants (F) at 48 hpf reveals that the size of the dorsal aortae (indicated by arrow) in igu^fo10a mutants is significantly smaller than that found in wild type. Red lines in (E) and (F) indicate plane of sectioning in (A)–(D). Scale bar in (A) and (B) is 2 μm, (C) and (D) is 300 nm, and (E) and (F) is 100 μm.

Artery and vein marker expression is normal in igu^fo10a mutants Indistinguishable expression of the artery marker efnb2a is observed in cross-sections of the lateral dorsal aortae (red arrows) of the hindbrain and the dorsal aorta (red bracket) of the trunk in both wild type and igu^fo10a mutants at 28 hpf. Indistinguishable expression of the venous marker flt4 is observed in cross-sections of veins in the hindbrain (blue arrows) and posterior cardinal vein of the trunk (blue bracket) at 28 hpf. Scale bar is 30 μm.

Loss of Angiopoietin1–Tie signaling in igu^fo10a mutants results in hemorrhage. (A–F) (A and B) In wild type embryos at 28 hpf, angpt1 is expressed in the developing heart (h), hypochord (hc), and ventral mesenchyme (me) as shown in whole mount (A) and in cross-section (B). (D and E) In igu^fo10a mutants of the same age, angpt1 expression persists in the heart and hypochord, but is substantially down-regulated in ventral mesenchyme, as shown in whole mount (D) and in cross-section (E). The dashed line in (A) and (D) marks the level of the section in (B) and (E), respectively. (C and F) Expression of angpt2 is unchanged in igu^fo10a mutant embryos (F) compared to wild type sibs (C) at 28 hpf. (G–I) Injection of 3.2 ng of angpt1^{e2i2 MO} (H) or 3.9 ng of angpt2^{e2i2 MO} (I) into wild type embryos results in defective heart development, pericardial edema (H), and lack of circulation when compared to uninjected control embryos at 48 hpf (G). (J) Transgenic over-expression of angpt2 under the β-actin promoter results in hemorrhage (hem) in wild type embryos at 53 hpf. Scale bar is 500 μm in (A), (C), (D), and (F)–(J) and 50 μm in (B) and (E).

Multiple vessel origins for hemorrhages in igu^fo10 mutants at 48- 52 hpf
Cross sections of Isolectin B4 staining of blood in the hind brain in the vicinity of the ear (A-C), or in the posterior hindbrain/ anterior trunk (D-F) show hemorrhages in diverse tissue locations of igu mutants. (A-B) Hemorrhages are seen dorsal to the dorsal aorta (DA), but exterior to the brain in (B), as compared to blood confined to the dorsal aorta in wild-type (A). (C) Hemorrhages are also occasionally seen in the vicinity of the cerebellar central arteries (CCtA) within the brain, leading to accumulation of blood in the hindbrain ventricle. (D) More posteriorly, robust circulation can be observed in the anterior cardinal vein of a wild type embryo. (E) A hemorrhage in the somitic tissue is observed in one igu mutant, (F), while a hemorrhage deriving from an intersomitic vessel (ISV) of the trunk is also seen.

Blood moves dorsally after hemorrhage in the head
A single igu mutant embryo was imaged at 43 hpf at the onset of hemorrhage, at 53 hpf and at 67 hpf. As time progresses, blood pools dorsally away from the initial site of hemorrhage.

Endothelial tight junctions are formed normally in igu^fo10 mutants
(A) The presence of endothelial cell tight junctions (red arrows) was analyzed by transmission electron microscopy in igu^fo10 mutants and wild type embryos at 48 hpf. No difference was seen in the morphology of junctions between wild type and igu^fo10a mutants. (B) Tight junction length was measured in both wild type and igufo10a mutant embryos and shows no significant difference in tight junction length.

Vessel patterning in the head and trunk of igu embryos appears normal
53 hpf wild type and igu mutant embryos were imaged in a lateral view on a Tg(fli1a:neGFP)^y7 background. Patterning of both the head and trunk appear normal.

Artery-vein identity is normal in igu^fo10a mutants at 28 hpf. (A-B) vegfa is unchanged in the ventral head and somites (som) in both wild type (A) and igu^fo10a mutants (B). (C-D) notch3 expression in the dorsal aorta (da) is unchanged in both wild type (C) and igu^fo10a mutants (D). (E-F) flt4 expression is restricted to the posterior cardinal vein (pcv) and intersegmental vessels (isv) in both wild type (E) and igu^fo10a mutants (F). Scale bar is 600 μm.

angpt1 is dramatically downregulated in smo^nv122 mutants. (A, B) angpt1 is strongly expressed in ventral mesenchyme (me) in wild type embryos (A) at 28 hpf, but is almost absent in smo^nv122 mutants (B) at the same age. Scale bars are 500 μm.

Morpholino efficiency and off-target cell death measurements in angpt1^e2i2 and angpt2^e2i2 morphants. (A) RT-PCR of a fragment surrounding the exon 2-intron 2 boundary from angpt1^e2i2 morphants injected with 3.2 ng of angpt1^{e2i2 MO} showing a shift in size from the predominant isoform (arrowhead) to that of higher and lower molecular weights (asterisk) in comparison to uninjected control embryos. (B) RT-PCR of a fragment surrounding the exon 2-intron 2 boundary from angpt2^e2i2 morphants injected with 1.25 ng angpt2^{e2i2 MO} (the dose used to rescue hemorrhage in igu^fo10a mutants). Morpholino injection results in a shift from the wild type band in uninjected controls (arrowhead) to the presence of five additional bands of differing sizes representing different mis-spliced transcripts (asterisk) and reduced presence of a wild type sized band. (C-E) Off-target morpholino-induced cell death assessment using acridine orange. Epifluorescent images of uninjected embryos (C) incubated in acridine orange displaying basal levels of apoptosis. Angpt2^e2i2 (D) and angpt2^ATG (E) morphants injected with 1.25 ng or 1.3 ng of morpholino, respectively, display increased levels of apoptosis in the head region. This increase in apoptosis does not likely influence our rescue results since we would expect increased neural death to exacerbate hemorrhage, not rescue it. In A and B, “M” denotes New England Biolabs 100bp ladder. Scale bars in C-E are 500 μm.