The transcription factor EPAS-1/hypoxia-inducible factor 2alpha plays an important role in vascular remodeling

Abstract

We have studied the role of the basic helix-loop-helix-PAS transcription factor EPAS-1/hypoxia-inducible factor 2alpha in vascular development by gene targeting. In ICR/129 Sv outbred background, more than half of the mutants displayed varying degrees of vascular disorganization, typically in the yolk sac, and died in utero between embryonic day (E)9.5 and E13.5. In mutant embryos directly derived from EPAS-1(-/-) embryonic stem cells (hence in 129 Sv background), all embryos developed severe vascular defects both in the yolk sac and embryo proper and died between E9.5 and E12.5. Normal blood vessels were formed by vasculogenesis but they either fused improperly or failed to assemble into larger vessels later during development. Our results suggest that EPAS-1 plays an important role at postvasculogenesis stages and is required for the remodeling of the primary vascular network into a mature hierarchy pattern.

Figure 1

Targeting of the EPAS-1 gene. (A) Targeting construct. A 1.9-kb EcoRI (R)/NaeI (N) fragment was used as the 5′ arm and a 6.2-kb XmnI (X) fragment as the 3′ arm. Homologous recombination results in the replacement of most of exon 2 (E2) and a region of intron after it (≈ 1.0 kb) by IRES-lacZ and PGK-Neo sequences. EcoRV (RV) digestion generates a 6.5-kb fragment for the wild-type allele and a 4.8-kb fragment for the targeted allele, both of which hybridize with the external probe (bar). (B) A Southern blot of DNA samples from embryos resulting from heterozygotes mating. (C) Genotyping by PCR. (D) Reverse transcription–PCR analysis of total RNA of E11.5 embryos for the expression of EPAS-1. The amount of template used and PCR efficiency for each reaction was normalized by reverse transcription–PCR for β-actin.

Figure 2

Hemorrhaging in EPAS-1^−/− mutants. (A and B) E11.5. (A) Normal morphology of a mouse conceptus. (B) The yolk sac membrane appears pale, most likely because of hemorrhaging significantly before dissection, rather than failure in hematopoiesis. This is supported by data shown in D and F, both photographed at the time of active hemorrhaging. (G–J) Embryos dissected at E10.0. The littermate shown in I and J appears to have been arrested at an earlier stage, approximately at E9.5. Interestingly, pools of blood cells (white arrow) were found to be deposited at the bottom of the yolk sac cavity, indicating leakage of blood vessels. (I) Photographed immediately after the original bottom side was flipped up. (J) Photographed 15 min after flipping (black arrow shows the reference point). The blood pool had diffused, indicating that these cells were indeed outside blood vessels. (Bars in A, B, E, and F, 2 mm; C, D, and G–J, 600 μm.)

Figure 3

Immunohistochemical staining of normal and mutant yolk sac membranes (all at E11.5) with anti-CD31 antibody. (A) Normal yolk sac membrane, showing a highly organized vascular tree pattern. (B) The vascular pattern from a severely hemorrhaging mutant differs from the normal control only slightly when viewed at the whole-mount level. (C–F) Histological sections of anti-CD31-stained specimens. (D and F) In most severe mutants, about a third of endothelial cells form extended endothelial sheets without proper lumen structures. (G and H) Moderately hemorrhaging mutants had no significant vascular disorganization, when viewed both at the whole-mount level and in histological sections. (Bars: A, B, and G, 600 μm; C–F and H,100 μm.)

Figure 4

Subtle vascular disorganization and hemorrhaging at E9.5. (A and B) Whole-mount anti-CD31 staining of yolk sac membranes, where B was taken from one of the embryos similar to that shown in Fig. 2 I and J. Sporadic lesions were found in histological sections (D and F). The very long endothelial linings shown in D were never found in sections from normal controls. (E and F) In areas where active remodeling was occurring to form larger vessels, large openings such as shown in F, were occasionally found in mutants only, which allowed blood cells to leak out (arrow). Note that there is no sign of tissue tearing or cracking in nearby areas, arguing that the opening was not an artifact of poor specimen preparation. (Bars: A and B, 500 μm; C–F, 50 μm.)

Figure 5

Severe vascular disorganization in the yolk sac membranes of mutant embryos in 129 Sv strain background. (A) E9.5. About 20% of the E9.5 mutants showed such pattern. (B) E11.5. Vascular patterns derived from +/− ES cell lines were identical to those shown in Figs. 3A and 4A and are not duplicated here. (C and D) Histological sections. About half of the E9.5 yolk sac membranes demonstrated disorganization. (C) Elongated endothelial linings and enlarged vascular lumen (compare with Fig. 4C). (D) Lumen enlargement appeared to be caused by merging of adjacent vessels. Arrows indicate sites of vascular merging. (Bars: A and B, 500 μm; C, 150 μm; and D, 75 μm.)

Figure 6

Vascular defects in the embryo proper in 129 Sv background (E11.5). (A) An area of perineural vasculature in +/− embryos, showing an organized hierarchy, where large vessels branch into smaller ones. (B) Perineural vessels in −/− embryos failed to evolve into a vascular tree consisting of large and small vessels. (C and D) A similar defect is also observed in the developing eyes. (Bars: 300 μm.)