UBIAD1-mediated vitamin K2 synthesis is required for vascular endothelial cell survival and development

Abstract

Multi-organ animals, such as vertebrates, require the development of a closed vascular system to ensure the delivery of nutrients to, and the transport of waste from, their organs. As a result, an organized vascular network that is optimal for tissue perfusion is created through not only the generation of new blood vessels but also the remodeling and maintenance of endothelial cells via apoptotic and cell survival pathways. Here, we show that UBIAD1, a vitamin K2/menaquinone-4 biosynthetic enzyme, functions cell-autonomously to regulate endothelial cell survival and maintain vascular homeostasis. From a recent vascular transgene-assisted zebrafish forward genetic screen, we have identified a ubiad1 mutant, reddish/reh, which exhibits cardiac edema as well as cranial hemorrhages and vascular degeneration owing to defects in endothelial cell survival. These findings are further bolstered by the expression of UBIAD1 in human umbilical vein endothelial cells and human heart tissue, as well as the rescue of the reh cardiac and vascular phenotypes with either zebrafish or human UBIAD1. Furthermore, we have discovered that vitamin K2, which is synthesized by UBIAD1, can also rescue the reh vascular phenotype but not the reh cardiac phenotype. Additionally, warfarin-treated zebrafish, which have decreased active vitamin K, display similar vascular degeneration as reh mutants, but exhibit normal cardiac function. Overall, these findings reveal an essential role for UBIAD1-generated vitamin K2 to maintain endothelial cell survival and overall vascular homeostasis; however, an alternative UBIAD1/vitamin K-independent pathway may regulate cardiac function.

Fig. 1.

reddish^s587 mutants exhibit cranial hemorrhage, degenerating vessels and pericardial edema. (A-F) Bright-field micrographs of wild-type (wt) and reh^s587 mutant (reh^-/-) embryos at (A-C) 48 hpf and (D-F) 72 hpf (A,C,D,F). Black arrows indicate cranial hemorrhage in reh mutant. Red arrowheads indicate pericardial edema. (G-N) Fluorescence micrographs of (G,I,K,M) wild-type (wt) and (H,J,L,N) reh^s587 mutants (reh^-/-) in Tg(gata1:dsRed);Tg(kdrl:GFP) background at (G,H) 36, (I,J) 48, (K,L) 56 and (M,N) 72 hpf. Cranial hemorrhage as detected by extravasation of Tg(gata1:dsRed) labeled blood in the head (white arrows) occurs in reh^s587 mutants as early as (J) 48 hpf and increases during development (L,N). reh^-/- intersegmental vessels appear to degenerate by 72 hpf (N, white arrowheads). (O,P) Tg(kdrl:GFP) (O) wild-type (wt) and (P) reh^-/- intersegmental vessels confirm this degeneration in reh^-/- mutants at 72 hpf (white arrowheads). Red arrowhead in N indicates pericardial edema.

Fig. 2.

reh encodes UBIAD1. (A) Genetic map of the reh region. Numbers below SSLP markers indicate recombination events out of 1944 diploid embryos examined. The mapped reh critical region contains one BAC and one genomic scaffold. (B) Sequencing of ubiad1 cDNA revealed a T to A change at base pair 422 in the s587 mutant allele, resulting in a Leu-to-Gln substitution at residue 65. Black asterisk indicates the location of the mutation. (C,D) Eighty-seven percent of wild-type (wt) embryos injected with a ubiad1 splice morpholino exhibited cranial hemorrhage (black arrows) and mild cardiac edema (red arrowheads), but wild-type embryos injected with control morpholino displayed no discernible cardiovascular phenotypes. (E) Wild-type zebrafish ubiad1 (WT zfubiad1) mRNA, human UBIAD1 (WT hUBIAD1) mRNA and zebrafish kdrl:ubiad1 rescued the reh cardiac and vascular phenotype. However, neither zebrafish ubiad1^s587 (MT zfubiad1) mRNA nor zebrafish cmlc2:ubiad1 could rescue either reh cardiovascular mutant phenotypes.

Fig. 3.

Vascular integrity and endothelial survival are compromised in reh mutants due to increased apoptosis. (A-D) Confocal projections of 48 and 72 hpf Tg(kdrl:gfp);Tg(gata1:dsRed) (A,C) wild-type or (B,D) reh mutant zebrafish. Yellow arrow indicates extravasation of Tg(gata1:dsRed)-labeled blood. (E-L) Confocal sections of 48 and 72 hpf Tg(kdrl:gfp) wild-type (wt) and reh mutant zebrafish that were TUNEL stained (red) reveal that (F,H,J,L) reh endothelial cells exhibit increased apoptosis when compared with (E,G,I,K) wild-type endothelial cells. (I-L) Enlargements of boxed areas in E-H, respectively, show (J) increased cell death in not only reh endothelial cells (white arrows) but also in (L) cells surrounding these reh endothelial cells (white arrowheads) when compared with wild-type vasculature. (M,N) The number of apoptotic cells observed per high-power field for each condition. Mean+s.e.m. Student’s t-test, ^*P<0.05 (n=15 reh and wild-type zebrafish).

Fig. 4.

Loss of ubiad1 function results in decreased cranial vasculature and endothelial cells due to increased endothelial karyorrhexis. (A-D) Confocal projections of 48 and 72 hpf Tg(fli1a:nEGFP);Tg(kdrl:cherry-ras) (A,C) control morpholino (MO) and (B,D) ubiad1 MO-injected zebrafish reveal that there are not only fewer reh endothelial cells but there is also an increase in nuclear fragmentation (karyorrhexis) in ubiad1 morpholino-injected zebrafish when compared with control-injected animals. Arrowheads indicate missing cranial vessels. (E) The number of endothelial nuclei is reduced in ubiad1 morpholino-injected zebrafish when compared with control morpholino-injected zebrafish. The number of endothelial nuclei observed per high-power field at 72 hpf. (F) An increase in the number of endothelial cells undergoing nuclear fragmentation was observed in ubiad1 morpholino-injected zebrafish when compared with control morpholino-injected zebrafish. The number of endothelial nuclei undergoing karyorrhexis per high-power field from 48 to 72 hpf. Mean+s.e.m. Student’s t-test, ^*P<0.05 (n=5). (G-H″″) Time-lapse imaging of boxed areas in C (G-G″″) and D (H-H″″) shows that endothelial nuclei from (H-H″″) ubiad1 morpholino-injected but not (G-G″″) control morpholino-injected zebrafish exhibit increased nuclear fragmentation, resulting in subsequent endothelial cell death. Yellow arrows indicate endothelial nuclei undergoing fragmentation.

Fig. 5.

UBIAD1 regulates menaquinone/vitamin K2 metabolism to maintain cranial vasculature. (A,B) Confocal projections of (A) DMSO- and (B) warfarin-treated reh;Tg(gata1:dsRed);Tg(kdrl:GFP) zebrafish reveal that warfarin treatment of zebrafish at 24 hpf leads to cranial hemorrhages (white arrow) by 72 hpf. (C,D) PK (C) and MK-4 (D) treatment at 36 hpf can rescue this warfarin-induced cranial hemorrhaging and vascular defect. (E-H) Confocal projections of PK- and MK-4-treated reh;Tg(gata1:dsRed);Tg(kdrl:GFP) mutant zebrafish showed that (H) MK-4 but not (F) PK treatment at 36 hpf can rescue the reh vascular phenotype by 72 hpf. However, (E) PK and (G) MK-4 treatment had no discernible effect on wild-type zebrafish. Arrows in F indicate cranial hemorrhages. (I) MK-4 or PK treatment rescued the warfarin-induced cranial hemorrhage; however, MK-4 but not PK treatment rescued the reh cranial hemorrhage. (J) The relationship of vitamin K2/MK-4, vitamin K1/PK, UBIAD1 and warfarin.