rasa1-related arteriovenous malformation is driven by aberrant venous signalling.

Arteriovenous malformations (AVMs) develop where abnormal endothelial signalling allows direct connections between arteries and veins. Mutations in RASA1, a Ras GTPase activating protein, lead to AVMs in humans and, as we show, in zebrafish rasa1 mutants. rasa1 mutants develop cavernous AVMs that subsume part of the dorsal aorta and multiple veins in the caudal venous plexus (CVP) - a venous vascular bed. The AVMs progressively enlarge and fill with slow-flowing blood. We show that the AVM results in both higher minimum and maximum flow velocities, resulting in increased pulsatility in the aorta and decreased pulsatility in the vein. These hemodynamic changes correlate with reduced expression of the flow-responsive transcription factor klf2a. Remodelling of the CVP is impaired with an excess of intraluminal pillars, which is a sign of incomplete intussusceptive angiogenesis. Mechanistically, we show that the AVM arises from ectopic activation of MEK/ERK in the vein of rasa1 mutants, and that cell size is also increased in the vein. Blocking MEK/ERK signalling prevents AVM initiation in mutants. Alterations in venous MEK/ERK therefore drive the initiation of rasa1 AVMs.

Stroke-associated intergenic variants modulate a human FOXF2 transcriptional enhancer.

SNPs associated with human stroke risk have been identified in the intergenic region between Forkhead family transcription factors FOXF2 and FOXQ1, but we lack a mechanism for the association. FoxF2 is expressed in vascular mural pericytes and is important for maintaining pericyte number and stabilizing small vessels in zebrafish. The stroke-associated SNPs are located in a previously unknown transcriptional enhancer for FOXF2, functional in human cells and zebrafish. We identify critical enhancer regions for FOXF2 gene expression, including binding sites occupied by transcription factors ETS1, RBPJ, and CTCF. rs74564934, a stroke-associated SNP adjacent to the ETS1 binding site, decreases enhancer function, as does mutation of RPBJ sites. rs74564934 is significantly associated with the increased risk of any stroke, ischemic stroke, small vessel stroke, and elevated white matter hyperintensity burden in humans. Foxf2 has a conserved function cross-species and is expressed in vascular mural pericytes of the vessel wall. Thus, stroke-associated SNPs modulate enhancer activity and expression of a regulator of vascular stabilization, FOXF2, thereby modulating stroke risk.

SAM domain variants of EPHB4 associated with aberrant signaling are linked to lymphatic-related fetal hydrops and facial dysmorphology.

Variants in EPHB4 (Ephrin type B receptor 4), a transmembrane tyrosine kinase receptor, have been identified in individuals with various vascular anomalies including Capillary Malformation-Arteriovenous Malformation syndrome 2 and lymphatic-related (non-immune) fetal hydrops (LRHF). Here, we identify two novel variants in EPHB4 that disrupt the SAM domain in two unrelated individuals. Proband 1 presented within the LRHF phenotypic spectrum with hydrops, and proband 2 presented with large nuchal translucency prenatally that spontaneously resolved in addition to dysmorphic features on exam postnatally. These are the first disease associated variants identified that do not disrupt EPHB4 protein expression or tyrosine-kinase activity. We identify that EPHB4 SAM domain disruptions can lead to aberrant downstream signaling, with a loss of the SAM domain resulting in elevated MAPK signaling in proband 1, and a missense variant within the SAM domain resulting in increased cell proliferation in proband 2. This data highlights that a functional SAM domain is required for proper EPHB4 function and vascular development.

The selector genes midline and H15 control ventral leg pattern by both inhibiting Dpp signaling and specifying ventral fate.

mid and H15 encode Tbx20 transcription factors that specify ventral pattern in the Drosophila leg. We find that there are at least two pathways for mid and H15 specification of ventral fate. In the first pathway, mid and H15 negatively regulate Dpp, the dorsal signal in leg development. mid and H15 block the dorsalizing effects of Dpp signaling in the ventral leg. In loss- and gain-of-function experiments in imaginal discs, we show that mid and H15 block the accumulation of phospho-Mad, the activated form of the Drosophila pSmad1/5 homolog. In a second pathway, we find mid and H15 must also directly promote ventral fate because simultaneously blocking Dpp signaling in mid H15 mutants does not rescue the ventral to dorsal transformation in most ventral leg structures. We show that mid and H15 act as transcriptional repressors in ventral leg development. The two genes repress the Dpp target gene Dad, the laterally expressed gene Upd, and the mid VLE enhancer. This repression depends on the eh1 domain, a binding site for the Groucho co-repressor, and is likely direct because Mid localizes to target gene enhancers in PCR-ChIP assays. A mid allele mutant for the repressing domain (eh1), mideh1, was found to be compromised in gain-of-function assays and in rescue of mid H15 loss-of-function. We propose that mid and H15 specify ventral fate through inhibition of Dpp signaling and through coordinating the repression of genes in the ventral leg.

foxc1 is required for embryonic head vascular smooth muscle differentiation in zebrafish.

Vascular smooth muscle of the head derives from neural crest, but developmental mechanisms and early transcriptional drivers of the vSMC lineage are not well characterized. We find that in early development, the transcription factor foxc1b is expressed in mesenchymal cells that associate with the vascular endothelium. Using timelapse imaging, we observe that foxc1b expressing mesenchymal cells differentiate into acta2 expressing vascular mural cells. We show that in zebrafish, while foxc1b is co-expressed in acta2 positive smooth muscle cells that associate with large diameter vessels, it is not co-expressed in capillaries where pdgfrß positive pericytes are located. In addition to being an early marker of the lineage, foxc1 is essential for vSMC differentiation; we find that foxc1 loss of function mutants have defective vSMC differentiation and that early genetic ablation of foxc1b or acta2 expressing populations blocks vSMC differentiation. Furthermore, foxc1 is expressed upstream of acta2 and is required for acta2 expression in vSMCs. Using RNA-Seq we determine an enriched intersectional gene expression profile using dual expression of foxc1b and acta2 to identify novel vSMC markers. Taken together, our data suggests that foxc1 is a marker of vSMCs and plays a critical functional role in promoting their differentiation.

The LIM-homeodomain transcription factor Islet2a promotes angioblast migration.

Angioblasts of the developing vascular system require many signaling inputs to initiate their migration, proliferation and differentiation into endothelial cells. What is less studied is which intrinsic cell factors interpret these extrinsic signals. Here, we show the Lim homeodomain transcription factor islet2a (isl2a) is expressed in the lateral posterior mesoderm prior to angioblast migration. isl2a deficient angioblasts show disorganized migration to the midline to form axial vessels and fail to spread around the tailbud of the embryo. Isl2a morphants have fewer vein cells and decreased vein marker expression. We demonstrate that isl2a is required cell autonomously in angioblasts to promote their incorporation into the vein, and is permissive for vein identity. Knockout of isl2a results in decreased migration and proliferation of angioblasts during intersegmental artery growth. Since Notch signaling controls both artery-vein identity and tip-stalk cell formation, we explored the interaction of isl2a and Notch. We find that isl2a expression is negatively regulated by Notch activity, and that isl2a positively regulates flt4, a VEGF-C receptor repressed by Notch during angiogenesis. Thus Isl2a may act as an intermediate between Notch signaling and genetic programs controlling angioblast number and migration, placing it as a novel transcriptional regulator of early angiogenesis.

The Tbx20 homolog Midline represses wingless in conjunction with Groucho during the maintenance of segment polarity.

The regulation of the segment polarity gene wingless is essential for the correct patterning of the Drosophila ectoderm. We have previously shown that the asymmetric activation of wingless downstream of Hedghog-signaling depends on the T-box transcription factors, midline and H15. Hedgehog activates wingless anterior to the Hedgehog domain. midline/H15 are responsible in part for repressing wingless in cells posterior to the Hedgehog expressing cells. Here, we show that Midline binds the Groucho co-repressor directly via the engrailed homology-1 domain and requires an intact engrailed-homology-1 domain to repress wingless. In contrast, the regulation of Serrate, a second target of midline repression, is not dependent on the engrailed-homology-1 domain. Furthermore, we identify a midline responsive region of the wingless cis-regulatory region and show that Midline binds to sequences within this region. Mutating these sequences in transgenic reporter constructs results in ectopic reporter expression in the midline-expression domain, consistent with wingless being a direct target of Midline repression.

Tinman is a direct activator of midline in the Drosophila dorsal vessel.

Heart development requires a conserved core of transcription factors comprised of Nkx2.5, GATA and T-box family transcription factors. In Drosophila melanogaster, the Nkx2.5 gene tinman acts upstream of many cardiac genes including the Tbx20 homolog midline, a critical regulator of heart development in both flies and vertebrates. By testing genomic fragments containing clusters of consensus Tinman-binding sites, we identified a 4.3 kb fragment 5' of midline that directs reporter expression in all midline-expressing heart cells and a 1.7 kb subfragment that drives reporter expression in mid-expressing heart cells that maintain tin expression. Both fragments direct reporter gene expression in response to tinman in transgenic embryos and in transient transfection assays in Drosophila S2 cells. Mutation of two Tinman binding sites (Tin1 and Tin2) reduces or abolishes cardiac expression in derivatives of the 1.7 kb fragment. We conclude that Tin is a direct regulator of midline in fly heart development.

midline represses Dpp signaling and target gene expression in Drosophila ventral leg development.

Ventral leg patterning in Drosophila is controlled by the expression of the redundant T-box Transcription factors midline (mid) and H15. Here, we show that mid represses the Dpp-activated gene Daughters against decapentaplegic (Dad) through a consensus T-box binding element (TBE) site in the minimal enhancer, Dad13. Mutating the Dad13 DNA sequence results in an increased and broadening of Dad expression. We also demonstrate that the engrailed-homology-1 domain of Mid is critical for regulating the levels of phospho-Mad, a transducer of Dpp-signaling. However, we find that mid does not affect all Dpp-target genes as we demonstrate that brinker (brk) expression is unresponsive to mid. This study further illuminates the interplay between mechanisms involved in determination of cellular fate and the varied roles of mid.

Functional similarity of Knirps CtBP-dependent and CtBP-independent transcriptional repressor activities.

Short-range transcriptional repressors are locally acting factors that play important roles in developmental gene expression in Drosophila. To effect repression, Knirps and other short-range repressors bind the CtBP corepressor, but these repressors also function via CtBP-independent pathways. Possible mechanistic differences between CtBP-dependent and -independent repression activities are poorly understood. The distinct activities might provide qualitatively different activities necessary in different promoter contexts, or they might combine to give quantitatively different effects. We analyze separately the CtBP-dependent and CtBP-independent domains of Knirps previously characterized in the embryo to determine possible functional distinctions of the two repression activities. Both domains are active in cell culture and are dependent on the same residues required for activity in the embryo. The domains have similar properties with respect to distance-dependent repression and resistance to inhibition by the deacetylase inhibitor trichostatin A. In tests of repressor-activator specificity, the extent of repression was related not to the chemical nature of the activation domain but to the total activation potential. This result indicates that the balance of competing activation and repression signals is decisive in determining the effectiveness of repressors on genetic switches, suggesting that multiple repression activities are utilized to provide quantitatively, rather than qualitatively, distinct outputs.

The expression of the T-box selector gene midline in the leg imaginal disc is controlled by both transcriptional regulation and cell lineage.

The Drosophila Tbx20 homologs midline and H15 act as selector genes for ventral fate in Drosophila legs. midline and H15 expression defines the ventral domain of the leg and the two genes are necessary and sufficient for the development of ventral fate. Ventral-specific expression of midline and H15 is activated by Wingless (Wg) and repressed by Decapentaplegic (Dpp). Here we identify VLE, a 5 kb enhancer that drives ventral specific expression in the leg disc that is very similar to midline expression. Subdivision of VLE identifies two regions that mediate both activation and repression and third region that only mediates repression. Loss- and gain-of-function genetic mosaic analysis shows that the activating and repressing regions respond to Wg and Dpp signaling respectively. All three repression regions depend on the activity of Mothers-against-decapentaplegic, a Drosophila r-Smad that mediates Dpp signaling, and respond to ectopic expression of the Dpp target genes optomoter-blind and Dorsocross 3. However, only one repression region is responsive to loss of schnurri, a co-repressor required for direct repression by Dpp-signaling. Thus, Dpp signaling restricts midline expression through both direct repression and through the activation of downstream repressors. We also find that midline and H15 expression are both subject to cross-repression and feedback inhibition. Finally, a lineage analysis indicates that ventral midline-expressing cells and dorsal omb-expressing cells do not mix during development. Together this data indicates that the ventral-specific expression of midline results from both transcriptional regulation and from a lack of cell-mixing between dorsal and ventral cells.

In vitro site selection of a consensus binding site for the Drosophila melanogaster Tbx20 homolog midline.

We employed in vitro site selection to identify a consensus binding sequence for the Drosophila melanogaster Tbx20 T-box transcription factor homolog Midline. We purified a bacterially expressed T-box DNA binding domain of Midline, and used it in four rounds of precipitation and polymerase-chain-reaction based amplification. We cloned and sequenced 54 random oligonucleotides selected by Midline. Electromobility shift-assays confirmed that 27 of these could bind the Midline T-box. Sequence alignment of these 27 clones suggests that Midline binds as a monomer to a consensus sequence that contains an AGGTGT core. Thus, the Midline consensus binding site we define in this study is similar to that defined for vertebrate Tbx20, but differs from a previously reported Midline binding sequence derived through site selection.