Many animals have a network of blood vessels that supplies oxygen and nutrients to every part of the body. Each organ contains a unique pattern of blood vessels; some have lots of densely packed vessels, while others have fewer vessels that are more widely spaced.

New blood vessels typically form by sprouting from the side of pre-existing vessels. This involves the endothelial cells that line the inner wall of blood vessels moving outwards to create a sprout that is made up of ‘tip cells’ and ‘stalk cells’. Tip cells are found at the front of the growing vessels and encourage the formation of new sprouts, while the stalk cells trail behind and elongate the sprout.

Two signaling pathways that involve two proteins called VEGF and Notch interact with each other to control which cells become tip cells and which become stalk cells. Cells with higher levels of VEGF signaling will become tip cells. These cells also activate Notch signaling, which in turn blocks VEGF signaling in their neighboring cells. This feedback mechanism enables a new sprout to form by forcing cells present around a newly formed tip cell to become stalk cells. However, it was still not understood how the different organs develop blood vessel networks with different densities.

In 2011, researchers revealed that two other proteins, Semaphorin3E and its receptor Plexin-D1, are expressed in tip cells in the back of the eye in mice and control the VEGF/Notch signaling pathway. Now Kur et al. – including some of the researchers involved in the 2011 work – have used a combination of predictive computer simulations and experimental approaches to understand this interaction in more detail.

The analysis showed that Semaphorin3E and Plexin-D1 speed up VEGF/Notch signaling, which causes new tip cells to form at a faster rate, and results in a more densely packed network of blood vessels. For example, in mice that lack Semaphorin3E and Plexin-D1, VEGF/Notch signaling was slower and new tip cells formed more slowly, which resulted in the blood vessel network at the back of the mice’s eyes being less dense.

Kur et al. propose that different organs have different ‘molecular metronomes’ that control the pace of VEGF/Notch signaling. A fast acting metronome would yield a dense network, while a slower one would form a less dense network. This helps to explain how diverse densities of blood vessel networks are formed in different organs. This work may aid efforts to develop therapeutic approaches for controlling the development of new blood vessels in cancers and other diseases.

The unique vascular topology of different organs exists for organ-specific functions. Different vascular network densities determine the specific amount of oxygen and nutrients to be delivered to each host tissue. Development of new vascular networks depends upon two types of specialized endothelial cells that work together: (1) The endothelial 'tip cell', which is located at the front of a growing vessel and guides its extension by sensing and responding to environmental cues, analogous to the axonal growth cone (Gerhardt et al., 2003; Kurz et al., 1996). (2) 'Stalk cells', which trail behind the tip cell and elongate the sprout. Tip and stalk cell identities are primarily controlled by the Dll4-Notch lateral inhibition pathway, which is activated in endothelial cells in response to VEGF from the local environment. VEGF-induced Dll4 activates Notch1 on the neighboring cell, leading to the down-regulation of VEGF receptor levels (Figure 1B) (Hellström et al., 2007; Leslie et al., 2007; Lobov et al., 2007; Suchting et al., 2007),(Benedito et al., 2009; Phng et al., 2009; Roca and Adams, 2007). Thus, lateral inhibition between endothelial cells generates an alternating pattern of active tip cells (Dll4 high) and inhibited stalk cells (Dll4 low). Moreover, tip cell selection is a dynamic process, and tip cell identity is transient (Arima et al., 2011; Jakobsson et al., 2010). Reiteration of tip cell selection, sprout extension, and connection of neighboring sprouts (anastomosis) is the basis for building a sophisticated vascular network (Adams and Eichmann, 2010; Carmeliet, 2000). Although this VEGF/Notch signaling pathway has been well studied and found to be conserved among different vascular beds and species, how this central pattern generator is modified by various target tissue-specific signals to yield diverse network topologies is not known.

Figure 1

Download asset Open asset

Here, we propose a general principle of how collective cell behavior determines the diverse densities of different networks: the generation of vascular topologies depends heavily on the temporal regulation of tip cell selection. Integrated simulations predict that as cell neighborhoods change, due to anastomosis or cell rearrangement events, lateral inhibition patterns will necessarily be disrupted, requiring continual re-selection of new tip cells (Bentley et al., 2009; 2014a). In fact, mouse genetics experiments demonstrated that tip cell numbers are positively correlated with the branching points of the network (Hellström et al., 2007; Kim et al., 2011). Therefore, the length of time it takes to establish (and re-establish) the alternating pattern of tip and stalk cells may be a missing, critical determinant of vascular topology (Bentley et al., 2014b; 2014c). Here, we took an integrated approach combining computational modeling, mouse genetics, and in vivo endothelial cell tracking to determine whether tip/stalk patterning can be temporally modulated to generate different topologies. We hypothesize that the frequency of tip cell selection determines the length of linear extension vs. branching, thus dictating the density of the network.

To begin to test this hypothesis, it is crucial to analyze dynamic single cell behavior and collective movement in the context of network formation (Arima et al., 2011; Jakobsson et al., 2010). Previously, we used static analyses of the postnatal mouse retina as a model to understand how neural signals shape vascular topology (Kim et al., 2011). We discovered that retina ganglion cell-derived Semaphorin3E (Sema3E) and its receptor Plexin-D1, which is expressed in endothelial cells at the front of actively sprouting blood vessels, control the VEGF/Notch pathway via a feedback mechanism. Mice lacking either Sema3E or Plexin-D1 exhibited an uneven vascular growth front and a reduction of tip cells that resulted in a less branched network compared to their wildtype littermate controls (Kim et al., 2011) (Figure 1A). However, it is not clear how this phenotype is generated: specifically, how the Sema3E-Plexin-D1 feedback mechanism regulates VEGF/Notch signaling at a dynamic cellular level, and whether changes in temporal modulation of this pathway lead to the overall vascular topology phenotype.

To begin to understand how Sema3E-Plexin-D1 signaling modifies vascular topology formation in a dynamic, spatiotemporal manner, we took advantage of an existing agent-based computational model (the 'MemAgent-Spring Model' or MSM) that simulates the cellular processes during tip cell selection making explicit the time it takes for gene expression (e.g. transcription/translation) changes to occur (Figure 1B,C – note time delay parameters D1 and D2) (Bentley et al., 2008; 2009). The MSM has been tested against numerous independent experimental data sets and validated as predictive of new mechanisms in vivo/in vitro (Bentley et al., 2014a; Guarani et al., 2011; Jakobsson et al., 2010). To now simulate tip cell selection in the context of Sema3E-Plexin-D1 crosstalk signaling with VEGF/Notch signaling (Fukushima et al., 2011; Kim et al., 2011) the MSM model was extended by adding four new parameters (Figure 1B, Video 1–5), with sensitivity analyses and calibration simulations performed, which include modulation of the existing parameter (δ) representing the induction level of Dll4 by VEGFR-2 activation (See methods section). These four new parameters represent the time delay for induction of Plexin-D1 by VEGF (D3), the time delay (D4) and strength (s) of the reduction of Dll4 levels in response to Sema3E-Plexin-D1 signaling, based on the experimental data previously shown (Kim et al., 2011), as well as the degradation rate of Plexin-D1 (r3). Loss of Sema3E-PlexinD1 signaling was simulated by setting all Plexin-D1 levels to zero. Loss of function simulations recapitulated the two prominent features of Plxnd1^-/- and Sema3e^-/- mutant retinal vasculature remarkably closely: In vivo, mutant vascular networks exhibit fewer tip cells and 1.5–2 fold reduction in branching points, as well as an uneven growth front (compare Figure 1A with 1C [red boxes]) (Kim et al., 2011). Furthermore, the dynamic nature of the simulations provided a novel insight to explain how this phenotype is generated. Simulations predict that higher Dll4 levels in the cells generates a shift towards synchronized Dll4 fluctuations overtime in contiguous cells, as they collectively battle more strongly via lateral inhibition negative feedback, causing an overall delay in amplification of differences needed to select the alternating pattern of tip and stalk cells (Figure 1C–E, Video 6,7). Therefore, computational simulations suggest that Sema3E-Plexin-D1 signaling enhances the speed and frequency of tip cell selection and thus increases the density of retinal vascular networks.

Video 1

Download asset

Video 2

Download asset

Video 3

Download asset

Video 4

Download asset

Video 5

Download asset

Video 6

Download asset

Video 7

Download asset

We next examined single cell behavior in a mosaic environment, because lateral inhibition requires collective coordination, and the relative, not absolute, levels of Dll4 expression among neighboring cells determine the outcome of the tip cell selection process (Jakobsson et al., 2010). A Plxnd1^-/- cell will behave differently in competition with a second Plxnd1^-/- cell than in competition with a wildtype cell. Thus, we investigated whether the cell autonomous function of Sema3E-Plexin-D1 signaling in the direct competition process between two neighboring endothelial cells can drive changes to the tip cell selection dynamics. In a simulated mosaic vessel, where cells lacking Sema3E-Plexin-D1 signaling are intermingled with normal cells, the contribution of cells lacking Sema3E-Plexin-D1 signaling to the tip cell population was predicted to increase to 68–74% (robustly across a range of parameter values: δ =4–6, s=2–4) in comparison to control cells (Figure 2A,B, Video 8), with an increased speed of patterning compared to a vessel entirely lacking Sema3E-Plexin-D1 signaling. This result suggests that normally, Plexin-D1 cell-autonomously suppresses the tip cell phenotype.

Figure 2

Download asset Open asset

Video 8

Download asset

To test the direct competition of Plxnd1^-/- cells and wildtype cells in vivo, we performed mosaic analysis using mice with tamoxifen inducible loss of Plxnd1 expression in the vasculature (Cdh5-Cre-ER^T2; Plxnd1^flox/flox). In the sprouting front of the wildtype retina, cells expressing Plexin-D1 were equally distributed at both tip cell and the adjacent stalk cell positions (Figure 2C, left panel, Figure 2D), showing they have no preference for either position. However, at approximately 45% of Plxnd1^-/- mosaicism, 75% of the Plxnd1^-/- cells became tip cells (Figure 2C, middle panel, Figure 2D) indicating that these cells do have a competitive advantage over Plexin-D1 expressing wildtype cells during tip cell selection. In contrast, in control experiments using mice with tamoxifen inducible expression of GFP (Cdh5-Cre-ER^T2; Z/EG⁺), GFP positive cells showed no preference for either position (Figure 2C, right panel, Figure 2D), demonstrating that preferential tip cell occupancy of mutant cells is due to lack of Sema3E-Plexin-D1 signaling. Taken together, in silico prediction and in vivo mosaic retinal analyses demonstrate that Sema3E-Plexin-D1 signaling suppresses the tip cell phenotype in a cell-autonomous manner. Interestingly, this cell-autonomous effect at single cell level contrasts with the effect of Sema3E-Plexin-D1 signaling at the collective level, where more new tip cells are selected in the presence than in the absence of Sema3E-Plexin-D1 signaling (Figure 1). This finding highlights the differences between collective and cell autonomous behaviors, and the challenge of intuiting one from the other.

Having confirmed that the calibrated set of parameters is valid to model contributions of the Sema3E-Plexin-D1 pathway to VEGF/Notch patterning and to predict new in vivo data of a static nature, we next employed the model to predict the effect of Sema3E-Plexin-D1 signaling on the rate of tip cell selection during the fuller dynamic network formation processes in which tip and stalk cell identities are constantly re-defined during cell rearrangement/position switching. Here, we used a new 'MSM-CPM' model that has been previously extended, parameterized and experimentally validated to simulate tip/stalk patterning together with cellular rearrangements within a vascular sprout (Bentley et al., 2014a). In this model, a cell can move within the adhered collective of the sprout powered by multiple local junctional movements. By incorporating the newly calibrated Sema3E-PlexinD1 signaling extension the MSM-CPM model we simulated cell rearrangements in vascular sprouts in the presence or absence of Sema3E-Plexin-D1 signaling (Figure 3A,B, Video 9,10). Simulations predicted that in the presence of Sema3E-Plexin-D1 signaling, the tip cell is repeatedly overtaken by another cell. In contrast, in the absence of Sema3E-Plexin-D1 signaling, tip cell overtaking frequency is reduced by factor of 1.26; a given tip cell occupies the tip cell position for a longer time (Figure 3C). Given the MSM is a qualitative not quantitative model we also performed simulations over a range of parameters and found that if the strength of Dll4 upregulation by VEGF (δ) is increased then this delay in overtaking is further exaggerated; a given tip cell occupies the tip cell position longer (Figure 3C).

Figure 3

Download asset Open asset

Video 9

Download asset

Video 10

Download asset

To test experimentally if tip cell selection and overtaking rates are slowed down in the absence of Sema3E-Plexin-D1 signaling, we next performed live endothelial cell tracking to follow the behavior of individual cells in an actively forming vascular network. Previously, live imaging with single cell resolution has mainly been described in sprouting assays using aortic rings (Arima et al., 2011) or embryoid bodies (EBs) generated from ES cells (Jakobsson et al., 2010). However, those assays primarily give rise to simple linear sprouts with less frequent branching than observed during retinal angiogenesis. To better mimic in vivo temporal and spatial events within a highly branched network, we developed a new ex vivo system, using explants from embryonic lungs. Whole embryonic lungs (E12.0) were embedded within a collagen matrix in a tissue culture dish containing medium supplemented with recombinant human VEGF. After one day, endothelial sprouts grow out of the explant (Figure 4—figure supplement 1A) with tip cells extending filopodia into the collagen matrix (Figure 4—figure supplement 1B). We observed branching events at various positions of the sprouts (Figure 4B), as well as anastomosis (Figure 4—figure supplement 1C), showing that all steps of the in vivo angiogenic sprouting process are recapitulated in our system even though there is no gradient of VEGF signaling. Lung explants endogenously express Sema3e and Plxnd1 (Figure 4—figure supplement 1D). We could also detect Plexin-D1 and Dll4 in sprouting endothelial cells (Figure 4—figure supplement 1E,F). We next performed long-term live imaging of wildtype and Plxnd1^-/- explants and analyzed tip cell selection frequency. In contrast to the computational model where a tip cell selection event can only occur via a switch between two cells at the tip of the sprout (Figure 3A), in the live imaging assay, tip cell selection events occur in the following distinct categories (Figure 4A,B): a positional switch between a tip and a stalk cell in a linear sprout (switch), the selection of a new tip cell at the front of the sprout (branch, type I) or at more proximal sites (branch, type II). Single cells were tracked manually using a nuclear live stain. Consistent with the computational model prediction, in vivo live imaging data indeed showed a significant delay in the selection of new tip cells in Plxnd1^-/- vascular sprouts compared to wildtype sprouts. The overall appearance of new tip cells, i.e. the tip cell selection frequency, was slowed down by factor 1.5 in the mutants (Figure 4C,D, Video 11,12). When analyzing the different categories separately, the number of events per total imaging time was significantly reduced in the categories 'switch' and 'branch, type II' in the absence of Sema3E-Plexin-D1 signaling (Figure 4E).

Figure 4 with 3 supplements see all

Download asset Open asset

Video 11

Download asset

Video 12

Download asset

Finally, to directly test experimentally whether the reduced tip cell selection rate observed by live imaging in our Plxnd1^-/- lung explant indeed leads to a less branched network over time, we analyzed the topology of the lung explant under the same culture conditions as the live imaging paradigm. Using a computational method (Figure 4—figure supplement 2 and detailed description in the material and methods section) to analyze the number of branching points in an unbiased way we found a significant reduction in branching points of the network (Figure 4F,G). These data further demonstrate that a slowed tip cell selection rate results in a less branched vascular network.

Together, the ex vivo and in silico results demonstrate that Sema3E-Plexin-D1 signaling modulate the pace of tip cell selection. In the absence of Sema3E-Plexin-D1 signaling, the rate of tip cell selection is reduced, which overall leads to longer linear sprout extension with less frequent branching substantially influencing the architecture of the growing vasculature and resulting in a less dense network, as seen in the Plxnd1 and Sema3e mutant lung explant as well as in retina (Kim et al., 2011).

During angiogenesis, the topology of the network is shaped essentially through the dynamic process of stalk cells turning into a new tip cell (tip cell selection), which is dependent on the Delta-Notch lateral inhibition pathway, a widely used machinery to regulate aspects of development that require temporal control by a 'molecular clock'. For example, during somitogenesis, Delta-Notch oscillations determine the frequency of new somite formation ('the somite clock') (Aulehla and Pourquié, 2008). In endothelial cells, we propose that modulation of any of the components of the central pattern generator will result in an altered pace of Delta-notch oscillations and an altered vascular patterning. Deceleration of the Delta-notch feedback loop (by a 'slow molecular metronome') will lead to the selection of fewer tip cells within a certain time frame and thus to the formation of a less dense network with bigger pore sizes. Acceleration of the selection process (by a 'fast molecular metronome') would result in an overly dense network (Figure 4—figure supplement 3). However, complex nonlinear feedback dynamics are often hard to intuit, and further careful simulation integrated with experimentation will be required to fully elucidate the temporal modulations and topological outcomes possible.

In this work, we describe how Sema3E-Plexin-D1 signaling can modify vascular density by impinging on the central pattern generator VEGF/Notch signaling. Our computational modeling predictions, mouse genetics mosaic analysis, and live imaging of individual cell dynamics in actively forming blood vessel networks and computational quantification of branching points show that the lack of Sema3E-Plexin-D1 signaling slows down the rate of tip cell selection and rearrangement, resulting in a less branched vascular network. Therefore, the Sema3E-Plexin-D1 pathway represents a 'faster molecular metronome' that results in a relatively dense network. These data suggest that temporal regulation of this critical, iterative aspect of network formation could be a general mechanism, and additional temporal regulators with varying pace (fast vs. slow) may exist to sculpt vascular topology in different tissues. Furthermore, our findings may provide insights into our understanding of morphogenesis in general, and aid in efforts to develop therapeutic approaches for tissue engineering and control of tumor progression and vascular diseases.

1. 1. Adams RH
   2. Eichmann A

   (2010) Axon Guidance Molecules in Vascular Patterning

   Cold Spring Harbor Perspectives in Biology 2:a001875.

   https://doi.org/10.1101/cshperspect.a001875

   • Google Scholar
2. 1. Arima S
   2. Nishiyama K
   3. Ko T
   4. Arima Y
   5. Hakozaki Y
   6. Sugihara K
   7. Koseki H
   8. Uchijima Y
   9. Kurihara Y
   10. Kurihara H

   (2011) Angiogenic morphogenesis driven by dynamic and heterogeneous collective endothelial cell movement

   Development 138:4763–4776.

   https://doi.org/10.1242/dev.068023

   • Google Scholar
3. 1. Aulehla A
   2. Pourquié O

   (2008) Oscillating signaling pathways during embryonic development

   Current Opinion in Cell Biology 20:632–637.

   https://doi.org/10.1016/j.ceb.2008.09.002

   • Google Scholar
4. 1. Benedito R
   2. Roca C
   3. Sörensen I
   4. Adams S
   5. Gossler A
   6. Fruttiger M
   7. Adams RH

   (2009) The Notch Ligands Dll4 and Jagged1 Have Opposing Effects on Angiogenesis

   Cell 137:1124–1135.

   https://doi.org/10.1016/j.cell.2009.03.025

   • Google Scholar
5. 1. Bentley K
   2. Franco CA
   3. Philippides A
   4. Blanco R
   5. Dierkes M
   6. Gebala V
   7. Stanchi F
   8. Jones M
   9. Aspalter IM
   10. Cagna G
   11. Weström S
   12. Claesson-Welsh L
   13. Vestweber D
   14. Gerhardt H

   (2014a) The role of differential VE-cadherin dynamics in cell rearrangement during angiogenesis

   Nature Cell Biology 16:309–321.

   https://doi.org/10.1038/ncb2926

   • Google Scholar
6. 1. Bentley K
   2. Gerhardt H
   3. Bates PA

   (2008) Agent-based simulation of notch-mediated tip cell selection in angiogenic sprout initialisation

   Journal of Theoretical Biology 250:25–36.

   https://doi.org/10.1016/j.jtbi.2007.09.015

   • Google Scholar
7. Conference

   1. Bentley K
   2. Harrington KI
   3. Regan ER

   (2014b)

   In Can active perseption generate bistability? Heterogeneous collective dynamics and vascular patterning

   ALIFE 4th International Conference on the Synthesis and Simulation of Living Systems. pp. 328–335.

   • Google Scholar
8. 1. Bentley K
   2. Mariggi G
   3. Gerhardt H
   4. Bates PA

   (2009) Tipping the Balance: Robustness of Tip Cell Selection, Migration and Fusion in Angiogenesis

   PLoS Computational Biology 5:e1000549.

   https://doi.org/10.1371/journal.pcbi.1000549

   • Google Scholar
9. 1. Bentley K
   2. Philippides A
   3. Ravasz Regan E

   (2014c) Do Endothelial Cells Dream of Eclectic Shape?

   Developmental Cell 29:146–158.

   https://doi.org/10.1016/j.devcel.2014.03.019

   • Google Scholar
10. 1. Carmeliet P

    (2000) Mechanisms of angiogenesis and arteriogenesis

    Nature Medicine 6:389–395.

    https://doi.org/10.1038/74651

    • Google Scholar
11. 1. Costa LdaF
    2. Cesar RM, Jr

    (2009)

    Shape Classification and Analysis: Theory and Practice, CRC Press, Inc

    Shape Classification and Analysis: Theory and Practice, CRC Press, Inc.

    • Google Scholar
12. 1. Fukushima Y
    2. Okada M
    3. Kataoka H
    4. Hirashima M
    5. Yoshida Y
    6. Mann F
    7. Gomi F
    8. Nishida K
    9. Nishikawa S-I
    10. Uemura A

    (2011) Sema3E-PlexinD1 signaling selectively suppresses disoriented angiogenesis in ischemic retinopathy in mice

    Journal of Clinical Investigation 121:1974–1985.

    https://doi.org/10.1172/JCI44900

    • Google Scholar
13. 1. Gerhardt H
    2. Golding M
    3. Fruttiger M
    4. Ruhrberg C
    5. Lundkvist A
    6. Abramsson A
    7. Jeltsch M
    8. Mitchell C
    9. Alitalo K
    10. Shima D
    11. Betsholtz C

    (2003) VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia

    The Journal of Cell Biology 161:1163–1177.

    https://doi.org/10.1083/jcb.200302047

    • Google Scholar
14. 1. Giudicelli F
    2. Lewis J

    (2004) The vertebrate segmentation clock

    Current Opinion in Genetics & Development 14:407–414.

    https://doi.org/10.1016/j.gde.2004.06.014

    • Google Scholar
15. 1. Graner F
    2. Glazier JA

    (1992) Simulation of biological cell sorting using a two-dimensional extended Potts model

    Physical Review Letters 69:2013–2016.

    https://doi.org/10.1103/PhysRevLett.69.2013

    • Google Scholar
16. 1. Gu C
    2. Yoshida Y
    3. Livet J
    4. Reimert DV
    5. Mann F
    6. Merte J
    7. Henderson CE
    8. Jessell TM
    9. Kolodkin AL
    10. Ginty DD

    (2005) Semaphorin 3E and plexin-D1 control vascular pattern independently of neuropilins

    Science 307:265–268.

    https://doi.org/10.1126/science.1105416

    • Google Scholar
17. 1. Guarani V
    2. Deflorian G
    3. Franco CA
    4. Krüger M
    5. Phng L-K
    6. Bentley K
    7. Toussaint L
    8. Dequiedt F
    9. Mostoslavsky R
    10. Schmidt MHH
    11. Zimmermann B
    12. Brandes RP
    13. Mione M
    14. Westphal CH
    15. Braun T
    16. Zeiher AM
    17. Gerhardt H
    18. Dimmeler S
    19. Potente M

    (2011) Acetylation-dependent regulation of endothelial Notch signalling by the SIRT1 deacetylase

    Nature 473:234–238.

    https://doi.org/10.1038/nature09917

    • Google Scholar
18. 1. Hellström M
    2. Phng L-K
    3. Hofmann JJ
    4. Wallgard E
    5. Coultas L
    6. Lindblom P
    7. Alva J
    8. Nilsson A-K
    9. Karlsson L
    10. Gaiano N
    11. Yoon K
    12. Rossant J
    13. Iruela-Arispe ML
    14. Kalén M
    15. Gerhardt H
    16. Betsholtz C

    (2007) Dll4 signalling through Notch1 regulates formation of tip cells during angiogenesis

    Nature 445:776–780.

    https://doi.org/10.1038/nature05571

    • Google Scholar
19. 1. Jakobsson L
    2. Franco CA
    3. Bentley K
    4. Collins RT
    5. Ponsioen B
    6. Aspalter IM
    7. Rosewell I
    8. Busse M
    9. Thurston G
    10. Medvinsky A
    11. Schulte-Merker S
    12. Gerhardt H

    (2010) Endothelial cells dynamically compete for the tip cell position during angiogenic sprouting

    Nature Cell Biology 12:943–953.

    https://doi.org/10.1038/ncb2103

    • Google Scholar
20. 1. Jakobsson L
    2. Kreuger J
    3. Holmborn K
    4. Lundin L
    5. Eriksson I
    6. Kjellén L
    7. Claesson-Welsh L

    (2006) Heparan Sulfate in trans Potentiates VEGFR-Mediated Angiogenesis

    Developmental Cell 10:625–634.

    https://doi.org/10.1016/j.devcel.2006.03.009

    • Google Scholar
21. 1. Kim J
    2. Oh W-J
    3. Gaiano N
    4. Yoshida Y
    5. Gu C

    (2011) Semaphorin 3E-Plexin-D1 signaling regulates VEGF function in developmental angiogenesis via a feedback mechanism

    Genes & Development 25:1399–1411.

    https://doi.org/10.1101/gad.2042011

    • Google Scholar
22. 1. Kurz H
    2. Gärtner T
    3. Eggli PS
    4. Christ B

    (1996) First Blood Vessels in the Avian Neural Tube Are Formed by a Combination of Dorsal Angioblast Immigration and Ventral Sprouting of Endothelial Cells

    Developmental Biology 173:133–147.

    https://doi.org/10.1006/dbio.1996.0012

    • Google Scholar
23. 1. Leslie JD
    2. Ariza-McNaughton L
    3. Bermange AL
    4. McAdow R
    5. Johnson SL
    6. Lewis J

    (2007) Endothelial signalling by the Notch ligand Delta-like 4 restricts angiogenesis

    Development 134:839–844.

    https://doi.org/10.1242/dev.003244

    • Google Scholar
24. 1. Lobov IB
    2. Renard RA
    3. Papadopoulos N
    4. Gale NW
    5. Thurston G
    6. Yancopoulos GD
    7. Wiegand SJ

    (2007) Delta-like ligand 4 (Dll4) is induced by VEGF as a negative regulator of angiogenic sprouting

    Proceedings of the National Academy of Sciences of the United States of America 104:3219–3224.

    https://doi.org/10.1073/pnas.0611206104

    • Google Scholar
25. 1. Monvoisin A
    2. Alva JA
    3. Hofmann JJ
    4. Zovein AC
    5. Lane TF
    6. Iruela-Arispe ML

    (2006) VE-cadherin-CreERT2 transgenic mouse: a model for inducible recombination in the endothelium

    Developmental Dynamics 235:3413–3422.

    https://doi.org/10.1002/dvdy.20982

    • Google Scholar
26. 1. Novak A
    2. Guo C
    3. Yang W
    4. Nagy A
    5. Lobe CG

    (2000) 3.0.CO;2-G" class="reference__title">Z/EG, a double reporter mouse line that expresses enhanced green fluorescent protein upon cre-mediated excision

    Genesis 28:147–155.

    https://doi.org/10.1002/1526-968X(200011/12)28:3/4<147::AID-GENE90>3.0.CO;2-G

    • Google Scholar
27. 1. Palàgyi K
    2. Kuba A

    (1998) A 3D 6-subiteration thinning algorithm for extracting medial lines

    Pattern Recognition Letters 19:613–627.

    https://doi.org/10.1016/S0167-8655(98)00031-2

    • Google Scholar
28. 1. Phng L-K
    2. Potente M
    3. Leslie JD
    4. Babbage J
    5. Nyqvist D
    6. Lobov I
    7. Ondr JK
    8. Rao S
    9. Lang RA
    10. Thurston G
    11. Gerhardt H

    (2009) Nrarp Coordinates Endothelial Notch and Wnt Signaling to Control Vessel Density in Angiogenesis

    Developmental Cell 16:70–82.

    https://doi.org/10.1016/j.devcel.2008.12.009

    • Google Scholar
29. 1. Roca C
    2. Adams RH

    (2007) Regulation of vascular morphogenesis by Notch signaling

    Genes & Development 21:2511–2524.

    https://doi.org/10.1101/gad.1589207

    • Google Scholar
30. Book

    1. Stockman GC
    2. Shapiro LG

    (2001)

    Computer Vision

    Prentice Hall.

    • Google Scholar
31. 1. Suchting S
    2. Freitas C
    3. le Noble F
    4. Benedito R
    5. Breant C
    6. Duarte A
    7. Eichmann A

    (2007) The Notch ligand Delta-like 4 negatively regulates endothelial tip cell formation and vessel branching

    Proceedings of the National Academy of Sciences of the United States of America 104:3225–3230.

    https://doi.org/10.1073/pnas.0611177104

    • Google Scholar
32. 1. Tsai W-H

    (1985) Moment-preserving thresolding: A new approach

    Computer Vision, Graphics, and Image Processing 29:377–393.

    https://doi.org/10.1016/0734-189X(85)90133-1

    • Google Scholar
33. 1. Zhang Y
    2. Singh MK
    3. Degenhardt KR
    4. Lu MM
    5. Bennett J
    6. Yoshida Y
    7. Epstein JA

    (2009) Tie2Cre-mediated inactivation of plexinD1 results in congenital heart, vascular and skeletal defects

    Developmental Biology 325:82–93.

    https://doi.org/10.1016/j.ydbio.2008.09.031

    • Google Scholar

Author details

1. Esther Kur

   Department of Neurobiology, Harvard Medical School, Boston, United States

   Contribution

   EK, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

2. Jiha Kim

   Department of Neurobiology, Harvard Medical School, Boston, United States

   Contribution

   JK, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

3. Aleksandra Tata

   Department of Neurobiology, Harvard Medical School, Boston, United States

   Contribution

   AT, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

4. Cesar H Comin

   Instituto de Física de São Carlos, University of Sao Paulo, Sao Carlos, Brazil

   Contribution

   CHC, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

5. Kyle I Harrington

   Center for Vascular Biology Research, Department of Pathology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, United States

   Contribution

   KIH, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

6. Luciano da F Costa

   Instituto de Física de São Carlos, University of Sao Paulo, Sao Carlos, Brazil

   Contribution

   LdFC, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

7. Katie Bentley

   Center for Vascular Biology Research, Department of Pathology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, United States

   Contribution

   KB, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   kbentley@bidmc.harvard.edu

   Competing interests

   The authors declare that no competing interests exist.

8. Chenghua Gu

   Department of Neurobiology, Harvard Medical School, Boston, United States

   Contribution

   CG, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   Chenghua_Gu@hms.harvard.edu

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0002-4212-7232

• 2,554

  views

• 621

  downloads

• 20

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.