1. Developmental Biology
Download icon

VEGF/VEGFR2 signaling regulates hippocampal axon branching during development

  1. Robert Luck
  2. Severino Urban
  3. Andromachi Karakatsani
  4. Eva Harde
  5. Sivakumar Sambandan
  6. LaShae Nicholson
  7. Silke Haverkamp
  8. Rebecca Mann
  9. Ana Martin-Villalba
  10. Erin Margaret Schuman
  11. Amparo Acker-Palmer
  12. Carmen Ruiz de Almodóvar  Is a corresponding author
  1. Biochemistry Center (BZH), University of Heidelberg, Germany
  2. Medicine Faculty Mannheim, Heidelberg University, Germany
  3. University of Frankfurt, Germany
  4. Max Planck Institute for Brain Research, Germany
  5. Buchmann Institute for Molecular Life Sciences (BMLS), University of Frankfurt, Germany
  6. German Cancer Research Center (DKFZ), Germany
Research Article
  • Cited 1
  • Views 476
  • Annotations
Cite this article as: eLife 2019;8:e49818 doi: 10.7554/eLife.49818

Abstract

Axon branching is crucial for proper formation of neuronal networks. Although originally identified as an angiogenic factor, VEGF also signals directly to neurons to regulate their development and function. Here we show that VEGF and its receptor VEGFR2 (also known as KDR or FLK1) are expressed in mouse hippocampal neurons during development, with VEGFR2 locally expressed in the CA3 region. Activation of VEGF/VEGFR2 signaling in isolated hippocampal neurons results in increased axon branching. Remarkably, inactivation of VEGFR2 also results in increased axon branching in vitro and in vivo. The increased CA3 axon branching is not productive as these axons are less mature and form less functional synapses with CA1 neurons. Mechanistically, while VEGF promotes the growth of formed branches without affecting filopodia formation, loss of VEGFR2 increases the number of filopodia and enhances the growth rate of new branches. Thus, a controlled VEGF/VEGFR2 signaling is required for proper CA3 hippocampal axon branching during mouse hippocampus development.

Introduction

The correct function of neuronal networks relies on the establishment of specific patterns of connectivity between axons and dendrites. In this respect, proper regulation of axon branching during development is crucial to ensure precise axonal connections with multiple synaptic targets. Hippocampus development starts in mice around embryonic day E14 and continues through the first postnatal weeks (Tole et al., 1997). During this process, CA3 pyramidal neurons project their collateral axons to other CA3 neurons (associational connections) and to the CA1 hippocampal field via the Schaffer collaterals (Witter, 2007). During the first two postnatal weeks, CA3 axon collaterals undergo a process of growth, branching, remodeling and maturation to establish their connective network (Gomez-Di Cesare et al., 1997). Although some molecular cues are described to regulate CA3 axon guidance (Skutella and Nitsch, 2001), the cues that regulate CA3 axon branching remain largely unknown.

De novo formation of axon branches, directly from the axon shaft (Gallo, 2011), comprises the major mechanism for establishing axon connectivity in the mammalian central nervous system (CNS) (Kalil and Dent, 2014). Target-derived cues such as extrinsic axon guidance cues (Netrin, Slits, Semaphorins or Ephrins), growth factors (Neurotrophins, BDNF, FGF-2, NGF) or morphogens (WNTs) can regulate axon branching by different dynamic strategies including promoting branch growth, repulsion, pruning or stabilization of branches (Kalil and Dent, 2014). Axon branching can also be regulated by activity-dependent mechanisms (Yamamoto and López-Bendito, 2012). In all cases, cytoskeleton reorganization is required for the initiation and growth of axon branches. Initiation of a new axon branch occurs when actin filaments accumulate along the axon (actin patches) and expand to membrane protrusions from which filopodia will emerge. Subsequently, microtubules will invade the filopodia thus giving rise to new axon branches (Gallo, 2011). Several of the molecular cues that regulate axon branching have been shown to regulate actin or microtubule dynamics. For example, Netrin-1 promotes axon branching in cortical neurons by inducing a rapid accumulation of actin filaments in filopodia (Dent et al., 2004). Also, asymmetric EGFR signaling has been shown to regulate actin dynamics and thereby axon branch pruning in Drosophila dorsal cluster neurons (Zschätzsch et al., 2014).

Vascular endothelial growth factor A (VEGFA, from here on termed VEGF) has been implicated in various neurodevelopmental processes including neurite outgrowth, neuronal survival and migration, as well as axon guidance (Carmeliet and de Almodovar, 2013; Erskine et al., 2011; Meissirel et al., 2011; Ruiz de Almodovar et al., 2010; Ruiz de Almodovar et al., 2011; Schwarz et al., 2004). Those direct effects on neurons are mediated by signaling via VEGFR2 (also known as KDR and FLK1) (Carmeliet and de Almodovar, 2013; Erskine et al., 2011; Meissirel et al., 2011; Ruiz de Almodovar et al., 2010; Ruiz de Almodovar et al., 2011; Schwarz et al., 2004) or via Neuropilin 1 (Erskine et al., 2011; Schwarz et al., 2004). Whether direct signaling of VEGF on neurons can regulate axon branching still remains unknown. Here we show that VEGF/VEGFR2 signaling regulates axon branching in CA3 hippocampal neurons. We find that VEGFR2 is expressed in CA3 hippocampal neurons during development and that VEGF is temporally and dynamically expressed in CA1-CA3 hippocampal neurons as well as in glial cells. We show that VEGFR2 is dynamically distributed along the axon and that VEGF stimulation increases VEGFR2 motility and localization towards actin-rich structures. We further show that CNS-specific VEGFR2 knockout mice display increased hippocampal axon branching in vivo, with branches that appear to be less mature and that form less functional synapses with CA1 neurons. Mechanistically, while VEGF stimulation results in increased axon branching by promoting the growth of newly formed branches in a Src Family Kinases (SFKs)-dependent manner, VEGFR2 inactivation leads to an increase in filopodia number that subsequently leads to increased branch formation.

Results

VEGFR2 and VEGF are expressed in the developing mouse hippocampus

Previous studies have demonstrated the expression of VEGF and its receptors in the adult murine hippocampus (Licht et al., 2010; Wang et al., 2005). To characterize their expression during hippocampal development, we performed in situ hybridization (ISH) at late embryonic (E18.5) and early postnatal (P4 and P8) stages. As expected, the mRNA encoding VEGFR2 was expressed in blood vessels (Figure 1A). In addition, we also detected VEGFR2 mRNA transcripts specifically in the CA3 hippocampal region throughout all developmental stages analyzed (Figure 1A). To further characterize the expression of VEGFR2, we took advantage of a transgenic knock-in mouse line in which GFP expression reliably reflects endogenous expression of VEGFR2 (Vegfr2-GFP, where exon 1 of the Kdr gene is replaced by GFP [Ema et al., 2006]). Immunostaining of postnatal brains at P4 and P8 with an antibody against GFP revealed specific labeling of the CA3 hippocampal region but not in the CA1, in addition to the strong labeling of endothelial cells (Figure 1B–1D, Figure 1—figure supplement 1A). These results indicate that expression of VEGFR2 mRNA is not only detected in endothelial cells but also in cells of the CA3 region. In order to determine the CA3 cell types that express VEGFR2 mRNA we co-immunostained brain sections from P8 of Vegfr2-GFP mice with the pan-neuronal marker NeuN and the interneuron-specific marker calretinin. Analysis revealed that VEGFR2 is specifically expressed in pyramidal neurons, but not in interneurons (Figure 1B–1D; Figure 1—figure supplement 1B). This spatiotemporal analysis showed that VEGFR2 mRNA is expressed in CA3 pyramidal neurons during the first two postnatal weeks. As a third approach to characterize VEGFR2 protein localization, we performed immunostainings using a specific anti-VEGFR2 antibody previously used to detect VEGFR2 in both neurons and blood vessels (Bellon et al., 2010). Consistently, we detected VEGFR2 in CA3 pyramidal neurons at P8 (Figure 1E). Finally, RT-PCR and Western blotting of lysates from isolated primary hippocampal neurons verified VEGFR2 expression in vitro (Figure 1—figure supplement 1D and E). This developmental time window of VEGFR2 expression coincides with the dynamic neurite growth and branching of CA3 neurons.

Figure 1 with 1 supplement see all
VEGFR2 is expressed by CA3 pyramidal neurons during hippocampal development.

(A) VEGFR2 mRNA ISH in the hippocampus at E18.5, P4 and P8. Blood vessels are indicated by red arrowheads. Scale bars 250 µm. (B) GFP and NeuN immunostaining of Vegfr2-GFP hippocampus at P8. Scale bar 250 µm. (C,D) High magnification images of the CA3 region (C) and the CA1 region (D) from insets in (B). Scale bar 50 µm. (E) VEGFR2 immunostaining in the CA3 region of P8 hippocampus. High magnification images of the CA3 region (inset in left panel) and CA1 region are shown on the right. Scale bars 100 µm and 25 µm, respectively.

VEGF (encoded by Vegfa in mouse) is the main ligand and activator of VEGFR2. To characterize the expression of VEGF we performed ISH for VEGF mRNA in the developing mouse hippocampus. We detected VEGF mRNA in the CA1, CA2 and CA3 hippocampal regions of all the embryonic and postnatal stages analyzed (Figure 2A). Additionally, VEGF mRNA was detected in the dentate gyrus (DG) during postnatal development (Figure 2A). Consistently, we detected VEGF protein in conditioned medium from cultured isolated primary hippocampal neurons (Figure 2—figure supplement 1A). At P8 VEGF mRNA expression was also detected in other cell types of the hippocampus, in addition to CA hippocampal neurons (Figure 2A). Recent publications identified endothelial cells as a cellular source of VEGF during CNS development (Barber et al., 2018; Li et al., 2013). To further identify the other cells types that express VEGF in the hippocampus, we combined ISH for VEGF mRNA with immunostainings for NeuN and the glial fibrillary acidic protein (GFAP), that specifically labels astrocytes. We observed that both pyramidal neurons and astrocytes of the developing CA hippocampal regions express VEGF mRNA (Figure 2B and C). Notably, the expression levels of VEGF mRNA decrease in the neuronal populations during postnatal development, but remain rather stable in the other cell types (Figure 2A–2C).

Figure 2 with 1 supplement see all
VEGF mRNA is expressed in different cell types during hippocampal development.

(A) ISH for VEGF mRNA in the hippocampus at E18.5, P4 and P8. Scale bars 250 µm. (B) ISH for VEGF mRNA combined with NeuN immunostaining at P8. High magnification images of the CA1 and CA3 region are shown. VEGF-positive pyramidal cell layer (PC) is delineated by dotted lines. Scale bar 50 µm. (C) VEGF mRNA ISH combined with GFAP immunostaining at P8. Insets show high magnification images of the CA3 region (Scale bar 10 µm). The pyramidal cell layer (PC) is delineated by a dotted line. Arrowheads indicate VEGF mRNA positive glial cells. Scale bar 50 µm.

Taken together, our analysis indicates that VEGFR2 is expressed in pyramidal neurons of the CA3 area of the developing hippocampus. Additionally, VEGF is expressed in different cell populations of the developing hippocampus. This expression pattern opened the question of whether direct VEGF signaling to VEGFR2 in hippocampal neurons could regulate their development.

VEGF induces axon branching in hippocampal neurons in vitro

As axon branching of CA3 hippocampal neurons coincides with the spatiotemporal expression of VEGFR2 within the first two postnatal weeks (Gomez-Di Cesare et al., 1997), we investigated whether VEGF/VEGFR2 signaling is important for axon branching. For this, we used cultures of primary hippocampal neurons and stimulated them with 100 ng/ml VEGF at day in vitro 1 (1 DIV). We subsequently analyzed the axonal morphology at 3 DIV (axons were defined as the longest neurite process [Dotti et al., 1988]). VEGF stimulation of hippocampal neurons resulted in a significant increase in both the number and length of axon branches, when compared with vehicle stimulated (control) neurons (Figure 3A–3C). The number and length of primary neurites (excluding the axon), as well as the axon length, were not affected upon VEGF stimulation (Figure 3—figure supplement 1A–1C). Next, we performed time-lapse video microscopy of 1 DIV hippocampal neurons to analyze the dynamics of axon branch formation. Analysis of the movies revealed that VEGF stimulation led to an increase in branching and net branch growth rate of axon branches, without affecting branch retraction (Figure 3D and E; Figure 3—figure supplement 1D and E; Figure 3—video 1 and Figure 3—video 2).

Figure 3 with 5 supplements see all
VEGF stimulation promotes the growth of axon branches in hippocampal neurons.

(A) Representative images of 3 DIV hippocampal neurons stimulated with or without 100 ng/ml VEGF for 48 hr and stained with beta-III-tubulin. Scale bars 50 µm. (B,C) Quantification of axonal branch number (B) and branch length (C). Data are represented as % of non-stimulated control. Mean ± SEM,>60 neurons from n = 4. ****p<0.0001; unpaired Student’s ttest. (D,E) 1 DIV hippocampal neurons were stimulated with 50 ng/ml VEGF or vehicle control and time-lapse movies were recorded over the course of 4 hr. The number of extending axon branches (D) was quantified over the course of the movies and the net growth rate of axon branch was calculated (E). Data represents mean ± SEM from 39 neurons, n = 3 independent experiments. ***p<0.001; ****p<0.0001; unpaired Student’s ttest. (F) Hippocampal neurons were transfected with mCherry-UtrCH and imaged in TIRF-mode at 3 DIV. Using this approach, actin patches, membrane bending protrusions and filopodia could be identified and differentiated. Time-lapse movies were recorded over the course of 2 min (actin patches) or 10 min (protrusions and filopodia) to study the dynamic formation of such events. Scale bar 2 µm. (G) The number of newly forming actin patches during 2 min per 10 µm axon segment were counted before and after VEGF stimulation (100 ng/ml). Data represents mean ± SEM from at least 11 neurons of 2 independent experiments. n.s. not significant; paired Student’s ttest. (H,I) The number and the size of newly forming protrusions and filopodia were analyzed during the course of 10 min before and after VEGF stimulation (100 ng/ml). Data represents mean ± SEM from at least 12 neurons of at least three independent experiments. n.s. not significant; paired Student’s ttest.

Axon branches develop from filopodia that stabilize and continue elongating (Kalil and Dent, 2014). The first step in filopodia formation is the focal accumulation of actin filaments, known as actin patches (Kalil and Dent, 2014). Actin patches form spontaneously, grow in size and eventually dissipate. Only a subset of actin patches stabilizes, giving rise to protrusions and subsequently filopodia before fully dissipating (Loudon et al., 2006). To study these events in vitro, we transfected primary hippocampal neurons with a plasmid containing the calponin homology domain of the F-actin binding protein Utrophin (Utr-CH) fused to mCherry (mCherry-UtrCH) in order to visualize F-actin in the axon (Burkel et al., 2007) and performed time-lapse movies over the course of 2 to 10 min. Analysis of axonal segments showed that F-actin nucleation is mainly localized in patches along the axon shaft and branches. These patches appear and dissipate in a dynamic manner (Figure 3—video 3). Analysis of mCherry-UtrCH dynamics also revealed that membrane protrusions (<2 µm in length) arise from actin patches and can either transition into filopodia (>2 µm in length) or regress and lead to F-actin dissipation (Figure 3F; Figure 3—video 4). Using this experimental set up and quantifying the different events we show that stimulation of hippocampal neurons with VEGF did not result in any significant alteration in the total number of newly formed actin nucleation events (Figure 3G). Similarly, within the 10 min interval, the percentage of the newly formed protrusions and filopodia, as well as their size, was not affected upon VEGF stimulation (Figure 3H and I).

Altogether, these results indicate that VEGF stimulation of hippocampal neurons in vitro promotes the growth of newly formed branches, but does not affect the initial step of filopodia formation.

VEGF induces VEGFR2 activation and motility and promotes axon branching in a Src family kinase dependent manner

To determine whether VEGF signals via VEGFR2 to control axon branching, we first determined the activation of VEGFR2 upon VEGF stimulation in our culture system. Indeed, VEGF stimulation of 1 DIV hippocampal neurons resulted in increased phosphorylation of VEGFR2 (Y1175; Figure 4—figure supplement 1A).

To better understand the mechanism via which VEGFR2 signaling controls axon branching, we investigated the specific localization of VEGFR2 along the axon and its position or motility upon VEGF stimulation. For this, we transfected 1 DIV hippocampal neurons with a plasmid encoding a GFP-tagged version of the human VEGFR2 (VEGFR2-GFP). Confocal microscopy analysis of hippocampal neurons at 3 DIV revealed localization of the recombinant receptor in the axonal growth cones, branching points and actin-rich regions (Figure 4—figure supplement 1B–1D). To further investigate VEGFR2 localization and dynamics as well as its correlation with actin protrusions and filopodia, we co-transfected hippocampal neurons with the VEGFR2-GFP and the mCherry-UtrCH plasmids. Analysis of double transfected neurons confirmed that VEGFR2 localizes to actin-rich membrane protrusions and filopodia (Figure 4A). To analyze the effect of VEGF on the motility and directed localization of VEGFR2, we stimulated hippocampal neurons with VEGF for 5 min and performed live imaging along the axon over the course of 60 s before and after VEGF stimulation. Analysis of VEGFR2-GFP kymographs (Figure 4—figure supplement 1E) showed that stimulation with VEGF increased the motility of VEGFR2 in the axonal segments of the hippocampal neurons (Figure 4B and C; Figure 4—video 1 and Figure 4—video 2). VEGF stimulation led to an enhanced movement of VEGFR2-GFP towards (converging) membrane protrusions and filopodia (Figure 4D). Consistently, blockage of the receptor using an anti-VEGFR2 functional blocking antibody abolished this motility (Figure 4—figure supplement 1F–1H; Figure 4—video 3 and Figure 4—video 4).

Figure 4 with 5 supplements see all
VEGF stimulation induces VEGFR2 motility towards actin-rich protrusions and filopodia and promotes axonal branching in a Src-dependent manner.

(A) Hippocampal neurons were double transfected with mCherry-UtrCH and VEGFR2-GFP plasmids. TIRF microscopy reveals the localization of VEGFR2-GFP punctae at actin nucleation sites (arrows), the base of membrane protrusions and filopodia (black arrowheads) and along axon segments with neither actin nucleation nor protrusion sites (open arrowheads). Scale bar 5 µm. (B–D) To analyze relative VEGFR2-GFP motility and directionality, kymographs of VEGFR2-GFP mobility were generated from time-lapse movies of VEGFR2-GFP and mCherry-UtrCH co-transfected hippocampal neurons (note that IgG treatment is present in these neurons as the movies and kymographs were performed at the same time as the movies and resultant kymographs shown in Figure 4—figure supplement 1 F-H, and thus they serve as the corresponding controls) Upper images shows mCherry-UtrCH status at t = 0 s, middle images shows VEGFR2-GFP at t = 0 s, lower panels shows the kymograph of VEGFR2-GFP over the course of 1 min (B). Relative VEGFR2-GFP motility at protrusions and filopodia (C) as well as the direction of VEGFR2-GFP movement (D) were analyzed over the course of 1 min before and 5 min after VEGF stimulation (100 ng/ml). Per condition,>13 neurons were analyzed of at least three independent experiments. n.s. not significant, *p<0.05; Chi-squared test. (E) Hippocampal neurons were stimulated with or without VEGF and fixed after 5 min. Immunostaining for p-Src was performed and the relative fluorescence was analyzed in the axonal growth cone (GC) or along the axon. Data are represented as mean ± SEM, from at least three independent experiments. *p<0.05; ***p<0.001; One-way ANOVA. (F,G) 1 DIV hippocampal neurons were treated with 1 µM PP2 (a widely used SFK inhibitor) or PP3 (control) for 1 hr prior stimulation with or without 100 ng/ml VEGF for 48 hr. Quantifications of axonal branch number (F) and branch length (G) are shown. Data are represented as % of non-stimulated control. Mean ± SEM of at least three independent experiments. n.s. not significant; **p<0.01; ****p<0.0001; Two-way ANOVA. (H,I) 1 DIV hippocampal neurons were pretreated with 1 µM PP2 or PP3 for 1 hr. After stimulation with 50 ng/ml VEGF or vehicle control time-lapse movies were recorded over the course of 4 hr. The number of extending axon branches is quantified over the course of the movies (H) and the net growth rate of axon branch was calculated (I). Mean ± SEM of at least three independent experiments. n.s. not significant; ***p<0.001; ****p<0.0001; Two-way ANOVA.

Together, these data show that VEGFR2 moves towards actin-rich structures upon VEGF stimulation and suggest that VEGFR2 might subsequently locally stabilize the engaged actin structure to form a new emerging axon branch.

In cerebellar granule cells and commissural neurons, where VEGFR2 is required for migration and axon guidance respectively, VEGF/VEGFR2-mediated effects occur via SFKs activation (Bellon et al., 2010; Chauvet et al., 2007; Meissirel et al., 2011; Ruiz de Almodovar et al., 2011). Moreover, SFKs can remodel the actin cytoskeleton (Brunton et al., 2004; Wang et al., 2011; Winograd-Katz et al., 2011). Thus, we next investigated whether VEGF-induced axon branching in hippocampal neurons may also require the activation of SFKs. To address this, we first determined the levels of SFKs activation upon VEGF stimulation of 1 DIV neurons. We immunostained control and VEGF treated hippocampal neurons with a specific antibody that detects the activated form of SFKs. Quantitative analysis of the fluorescence intensity showed an increased activation of SFKs at axonal growth cones (GC) as well as in the primary axon upon treatment with VEGF (Figure 4E).

We next explored whether SFKs activity is required for the VEGF-induced morphological changes in hippocampal neurons. To this end, we used 1 μM PP2 (a widely used SFK inhibitor) to block SFKs activity in cultured hippocampal neurons or its inactive analog, PP3, as a control. Analysis of the axon branch number and length revealed a significant increase in PP3-treated control neurons upon VEGF stimulation, which was completely abolished upon PP2 treatment (Figure 4F and G). The requirement of SFKs activity was further confirmed by monitoring branching dynamics using time-lapse video microscopy in 1 DIV hippocampal neurons. Blocking SFKs completely abrogated the VEGF-mediated increase in the number of axon branches and the branch growth rate (Figure 4H and I).

Thus, VEGF can induce axon branching via both the mobilization of VEGFR2 to actin-rich structures and the activation of SFKs.

Nervous system specific VEGFR2 knockout mice show increased axon branching and deficits in synapse development in vivo

To investigate whether VEGFR2 is crucial for the development of hippocampal neurons, we genetically deleted the gene encoding VEGFR2 (Kdr) in the nervous system by crossing Kdrlox/- mice (Haigh et al., 2003) with the Nestin-cre mouse line (Tronche et al., 1999). PCR analysis and Western blotting of lysates from isolated hippocampal neurons confirmed the absence of VEGFR2 mRNA and protein in hippocampal neurons of Nestin-cre+;Kdrlox/- animals compared to Nestin-cre-;Kdrlox/- littermates (from hereon termed Nes-cre;Kdrlox/- and control, respectively) (see accompanying manuscript Harde et al., 2019) We first characterized the cytoarchitecture of the entire hippocampus in Nes-cre;Kdrlox/- mice. We labeled brain sections of P10 Nes-cre;Kdrlox/- and control mice with TO-PRO-3 (labels all nuclei), NeuN (labels neuronal nuclei) and L1 (labels axonal tracks) and analyzed the gross morphology of the hippocampus, the number of hippocampal neurons, the size of the CA3 area and the axonal tracks. Morphometric analysis revealed that hippocampal anatomy was not affected in Nes-Cre;Kdrox/- mice compared to control littermates (Figure 5—figure supplement 1A–1E), consistent with what was previously described in Nes-cre;Kdrlox/lox adult mice (Hermann et al., 2013).

To characterize the hippocampus in Nes-cre;Kdrlox/- mice in detail, we analyzed the morphology of CA3 hippocampal neurons at a high resolution level by focusing on their axons and investigated whether VEGFR2 plays a role in axon branching of these neurons in vivo. We set up a technique to label single axons of CA3 pyramidal neurons in situ as it has been previously described for DRG neurons (Schmidt and Rathjen, 2011). We used horizontal thick vibratome sections of postnatal brains and placed small DiI crystals in the stratum pyramidale of the CA3 region (Figure 5—figure supplement 2A and B). Short incubation with the DiI crystals allowed the labeling of a small number of axons and the identification of single axon segments and their branches (Figure 5—figure supplement 2C–2E). Labeled CA3 axons were imaged and traced in the stratum radiatum of the CA1 region where they project to (Figure 5—figure supplement 2C–2E). To validate this approach, we first characterized axon branching of CA3 neurons in different postnatal developmental stages of wildtype mice. Consistent with previous published data (Gomez-Di Cesare et al., 1997), we observed a steady increase in the number of axon branches until P16, which subsequently declined and reduced by P30 (Figure 5—figure supplement 2F). Next, we determined whether axon branching was affected in the brains of Nes-cre;Kdrlox/- mice at P10. Our data revealed that in the absence of VEGFR2 there is a significant increase of axon branches of CA3 pyramidal neurons (Figure 5A and B). Notably, these branches were shorter compared to control littermates (Figure 5C).

Figure 5 with 4 supplements see all
Nes-cre;Kdrlox/-mice show defects in axon branching and synapse density in vivo.

(A) Representative tracings in the CA1 stratum radiatum of CA3 DiI labeled axons from P10 control and Nes-cre;Kdrlox/- mice. Selected individual axon traces are shown highlighted in different colors in the panels below. Scale bar 100 µm. (B,C) Quantification of the axon branch number per 100 µm axon length (B) as well as their length distribution (C). Data are represented as mean ± SEM from n = 15 control and n = 18 Nes-cre;Kdrlox/- pups of 6 independent litters (B,C). n.s. not significant; *p<0.05; **p<0.01; unpaired Student’s ttest. (D) Representative electron micrographs of the hippocampal CA1 stratum radiatum from P10 control and Nes-cre;Kdrlox/- mice. Red arrowheads indicate synapses. Scale bar 1 µm. (E) Quantification of synapse density in control and Nes-cre;Kdrlox/- mice. Synapse numbers from 8 fields of view of n = 4 mice from each genotype were analyzed. *p<0.05; unpaired Student’s ttest. (F,G) Average mEPSC frequency (F) and amplitude (G) in control and Nes-cre;Kdrlox/- mice. n.s. not significant; **p<0.01; unpaired Student’s ttest. (H) Representative mEPSC recordings in control and Nes-cre;Kdrlox/- mice.

To further examine whether the increase in the number of axon branches would correlate with an increase in the number of synaptic connections, we quantified the number of functional synapses in the stratum radiatum of the CA1 area using transmission electron microscopy. Synapses were identified based on the presence of the synaptic cleft, postsynaptic densities and presynaptic vesicles. We observed that the number of synapses was decreased in Nes-cre;Kdrlox/- mice when compared to their control littermates (Figure 5D and E). Consistently, whole-cell patch clamp recordings in CA1 neurons of Nes-cre;Kdrlox/- mice revealed a decrease in the frequency of miniature excitatory postsynaptic currents (mEPSCs) (Figure 5F and H). The amplitude of mEPSCs was unchanged, suggesting that the strength of newly formed CA3-CA1 synapses is similar in both control and Nes-cre;Kdrlox/- mice (Figure 5G).

Together, these results indicate that VEGFR2 is required for proper axon branching and synapse formation in CA3 pyramidal neurons. They further show that Nes-cre;Kdrlox/- mice develop an increased amount of axon branches that remain immature and do not form functional synapses within the CA1 stratum radiatum, suggesting that perhaps this increase in branching is built as a compensatory response.

Absence of VEGFR2 in hippocampal neurons results in newly formed axon branches due to increased filopodia number and growth rate

To better understand the mechanisms underlying the increased axon branching upon loss of VEGFR2, we isolated hippocampal neurons from Nes-cre;Kdrlox/- and control mice and analyzed their axonal morphology in vitro. Morphometric analysis of knockout neurons at 3 DIV revealed a significant increase in the number and length of axon branches (Figure 6A–6C). These results are in agreement with the axonal phenotype we observed in vivo in Nes-cre;Kdrlox/- animals and confirm that VEGFR2 is important for proper axonal development. To further explore whether this phenotype results from increased axon branch elongation or reduced retraction, we performed wide field time-lapse movies. We followed the dynamic branching process in hippocampal neurons from Nes-cre;Kdrlox/- mice or control neurons over the course of 4 hr. Quantitative analysis showed that loss of VEGFR2 led to an increase of the net growth and net growth rate of newly forming branches without affecting the number and rate of axonal branch retraction (Figure 6D and E; Figure 6—figure supplement 1A and B).

Figure 6 with 1 supplement see all
VEGFR2 deficiency in hippocampal neurons promotes axonal branching by increasing filopodia formation.

(A) Representative images of 3 DIV hippocampal neurons isolated from control and Nes-cre;Kdrlox/- embryos and stained beta-III tubulin. Axon branches are indicated by red arrowheads. Scale bars 50 µm. (B,C) Quantification of the axon branch number (B) and branch length (C) of control and Nes-cre;Kdrlox/- neurons at 3 DIV. Data are represented as % of control. Mean ± SEM,>120 neurons from n = 6. ****p<0.0001; unpaired Student’s ttest. (D,E) Time-lapse movies over the course of 4 hr were recorded from 1 DIV hippocampal neurons of control and Nes-cre;Kdrlox/- mice. The net axon branch growth was quantified over the course of the movies (D) and the growth rate of axon branch was calculated (E). Data represents mean ± SEM of at least three independent experiments. *p<0.05; unpaired Student’s ttest. (F) The number of newly forming actin patches during 2 min per 10 µm axon segment was counted in neurons of control and Nes-cre;Kdrlox/- mice. Data represents mean ± SEM of at least three independent experiments. n.s. not significant; unpaired Student’s ttest. (G,H) The number (G) and the size (H) of newly formed protrusions and filopodia was analyzed during the course of 10 min in neurons of control and Nes-cre;Kdrlox/- mice. Data represents mean ± SEM of at least three independent experiments. *p<0.05; **p<0.01; unpaired Student’s ttest.

Next, we determined the changes that occur at the level of actin patches and filopodia by analyzing mCherry-UtrCH transfected control and Nes-cre;Kdrlox/- hippocampal neurons. Using 2 min time-lapse TIRF movies we analyzed the dynamics of actin patch formation in hippocampal axons from Nes-cre;Kdrlox/- or control mice. We did not observe any changes in the formation of actin patches (Figure 6F). However, neurons from Nes-cre;Kdrlox/- mice showed a significantly increased percentage of filopodia formed in 10 min, as well as increased percentage of events larger than 2 µm (Figure 6G and H). Consistently with a functional signaling role of VEGFR2 in axon branching, similar results were obtained by blocking VEGFR2 signaling with an anti-VEGFR2 functional blocking antibody (α-VEGFR2) in WT isolated neurons (Figure 6—figure supplement 1A and C–K).

Our in vitro analysis shows that lack of VEGFR2 activation correlates with enhanced filopodia formation and increased net growth rate of axon branches, which together result in increased number of axon branches. Taken together, these in vitro results are consistent with the in vivo phenotype of Nes-cre;Kdrlox/- mice. In vivo, the increased axon branching is non-functional as axon branches fail to form synaptic contacts with the CA1 and present impaired functionality.

VEGF-mediated axon branching via VEGFR2 does not require Neuropilin 1 nor EphrinB2

Neuropilin 1 (Nrp1) is expressed in hippocampal neurons during development (Figure 5—figure supplement 3A; He and Tessier-Lavigne, 1997). As VEGF also binds Nrp1 and can induce signaling dependent or independent of VEGFR2 (Roth et al., 2016), we questioned whether Nrp1 is required for VEGF-induced axon branching. For this, we blocked Nrp1 in hippocampal neurons using an anti-Nrp1 neutralizing antibody (α-Nrp1) and analyzed axon branching. In the presence of α-Nrp1 VEGF was still able to induce axon branching (Figure 5—figure supplement 3B and C), thus indicating that Nrp1 is dispensable for VEGF-mediated effect on branching. Moreover, blockage of Nrp1 in the presence of α-VEGFR2 did not normalize the increased axon branching (Figure 5—figure supplement 3D and E), suggesting that in the absence of VEGFR2 Nrp1 is not a receptor via which VEGF could alternatively signal to regulate axon branching.

In the accompanying manuscript (Harde et al., 2019), we show that ephrinB2 is expressed in hippocampal neurons during development and that its expression is required for VEGFR2 internalization and VEGF-induced dendritic branching. We therefore tested whether VEGFR2 internalization is required for VEGF-induced axon branching and whether ephrinB2 would also play a role in regulating axon branching. Inhibition of VEGFR2 internalization with dynasore prevented the increased axon branching upon VEGF stimulation (Figure 5—figure supplement 4A and B), indicating that VEGFR2 internalization is indeed required for the response of the axon to VEGF. However, stimulation of primary ephrinB2 knockout (Efnb2-/-) hippocampal neurons with VEGF resulted in a similar increase in axon branching as control neurons (Figure 5—figure supplement 4C and D), suggesting that ephrinB2 is not regulating VEGFR2 function for its effects in axon branching. Consistently, analysis of CA3 hippocampal axon branching in vivo in compound mutant mice heterozygous for ephrinB2 and Kdr in neurons (Kdr-ephrinB2 compound mice; Nes-cre;Kdrlox/+;Efnb2lox/+; see accompanying manuscript) using the DiI labeling approach as described before did not show any difference when compared to control littermates (Figure 5—figure supplement 4E).

Altogether, these results show that VEGF/VEGFR2 regulation of hippocampal axon branching does not depend on other co-receptors such as Nrp1 and EphrinB2. They also highlight that the molecular mechanism of VEGF/VEGFR2 signaling, regulating hippocampal axon branching or hippocampal dendritogenesis and spine morphogenesis, are distinctly different.

Discussion

This study reports a novel molecular mechanism by which direct VEGF/VEGFR2 signaling on hippocampal pyramidal neurons regulates axon branching during development.

Our expression analysis shows that VEGF and VEGFR2 are regionally, dynamically and temporally expressed in the developing hippocampus at the time when axon arborization takes place. VEGF is expressed in CA neurons, glia and endothelial cells and its expression in neurons decreases during the first postnatal weeks. During late embryonic and early postnatal (first two weeks) hippocampus development, neuronal expression of VEGFR2 is restricted to CA3 neurons. These expression patterns are consistent with later postnatal and adult stages where VEGF expression becomes restricted to adult astrocytes and VEGFR2 is primarily found in vessels (Licht et al., 2011) and at the postsynaptic side of CA1 neurons (De Rossi et al., 2016).

We demonstrate that in hippocampal neurons VEGF activates VEGFR2 and SFKs to induce axon branching. While Nrp1 is also expressed in hippocampal neurons, its blockage does not inhibit VEGF-mediated increase in axon branching. Previous reports show that VEGF regulates cerebellar granule cell migration or commissural axon guidance in a SFK-dependent manner (Meissirel et al., 2011; Ruiz de Almodovar et al., 2011). Here we show that VEGF-induced axon branching requires VEGFR2 internalization and SFK activation. We found that VEGF stimulation induces activation of SFKs at growth cones and along the axon shaft. Thus, SFK activation seems to be a crucial step in VEGF-mediated effects in neurons. Upon VEGF stimulation, VEGFR2 motility and migration towards actin-rich protrusions and filopodia is enhanced, suggesting that localization of VEGFR2 in those cellular structures contributes to the final outcome of increased axon branching.

Interestingly, we observed that both stimulation with VEGF as well as the inactivation of VEGFR2 result in an increased number of axon branches (Figure 7). In all cases we see that branch growth is enhanced and that there are no differences in branch retraction. How can the activation and the loss of the same signaling receptor lead to a similar phenotype? We show that although the phenotypic outcome is the same, the cause for the increase in axon branch number differs between the activation and the loss of VEGFR2. In the case of VEGFR2 activation we find that VEGF stimulation does not change the percentage of forming filopodia but it does increase the number of branches (Figure 7). VEGF stimulation further increases the growth rate of axon branches. We propose, that VEGF induces axon branching by stabilizing and promoting the growth of branches once the filopodia have already been formed. In contrast, upon VEGFR2 loss, the percentage of forming filopodia increases, as well as the growth rate, overall resulting in an increased number of axon branches (Figure 7). One hypothesis could be that in absence of the receptor the axon branch cannot be properly stabilized and grow, and thus the axon branches that are formed remain shorter.

Working model for VEGF/VEGFR2 function in hippocampal axon branching.

VEGF stimulation induces axon branching by promoting the growth of newly formed branches without affecting filopodia formation. Loss-of-functional VEGFR2 also results in increased axon branching, but in this case from an increased number of filopodia and enhanced growth rate. However, in the loss of function situation, the generated branches are shorter and non-functional.

Similarly, disruption of VEGFR2 signaling in vivo (in Nes-cre;Kdrlox/- mice) results in axonal hyperbranching of CA3 hippocampal neurons. Despite the increase in axon branch number, we observed a reduced number of synapses in the CA1 area, indicating that the branches are not functional and rather immature. An increase in the number of immature non-functional axon branches has been described in DCN neurons in Drosophila upon inhibition of EGFR signaling (Zschätzsch et al., 2014). In our case, the overshooting of branches could occur as a compensatory response with the aim to restore proper axon branching.

The accompanying manuscript by Harde et al. (2019) shows that the VEGF/VEGFR2 signaling effect is not limited to the axon, but this ligand/receptor pair also regulates dendritic development and spine morphogenesis in hippocampal neurons. While in both compartments the internalization of VEGFR2 is required for VEGF signal transduction, the presence of EphrinB2 and its regulatory function seems to specifically contribute to dendritic development and to be dispensable for axonal branching. In line with this, we find that inactivation of VEGFR2 has an opposite effect in the axon and in the dendrites. While absence of VEGFR2 results in increased branching of hippocampal axons, it leads to a less branched dendritic network. A differential effect in dendrites and axons has been also described for other signaling molecules such as Sema3A, which promotes dendritic growth but restricts axon growth in cultured hippocampal neurons, or the leucine zipper kinase (DLK) pathway, which restrains dendritic growth but promotes axon growth (Wang et al., 2013; Wang et al., 2014). Therefore, VEGF/VEGFR2 signaling falls into the category of bimodal regulators that differentially direct axonal and dendritic development in an opposite manner.

A study in Drosophila elegantly showed that intrinsic asymmetry in EGFR localization and local activation of EGFR signaling within filopodia is required for proper axon branching of dorsal cluster neurons and that both EGFR activation and loss of function results in increase axon branching due to a defect in pruning (Zschätzsch et al., 2014). Our data provide for the first time evidence that in mammals similar processes can also modulate axon branching. In particular, we find that appropriate levels of VEGFR2 signaling in hippocampal neurons are required for proper axon branching by a mechanism that involves branch growth but not branch pruning. We show that upon VEGF stimulation and activation of VEGFR2 signaling, VEGFR2 motility increases within actin protrusions and filopodia, correlating with an increase in branch number. Two possibilities, non-exclusive, could explain the increased axon branching upon VEGFR2 deletion. On the one hand, the lack of VEGFR2 could lead to a decrease in the dynamics of protrusions turnover and thus increase the probability of a protrusion to become a filopodium. Our data support such a model as we observe that the absence of VEGFR2 increases the percentage of filopodia. On the other hand, a compensatory mechanism, yet unidentified, might become activated to overcome the inhibition of VEGFR2, resulting in higher branch number than in control conditions.

Materials and methods

Key resources table
Reagent type
(species) or resource
DesignationSource or referenceIdentifierAdditional information
Genetic reagent (Mus musculus)Kdrlox/loxHaigh et al., 2003
Genetic reagent (Mus musculus)Nestin-creTronche et al., 1999
Genetic reagent (Mus musculus)Vegfr2-GFPEma et al., 2006
Genetic reagent (Mus musculus)Efnb2lox/loxGrunwald et al., 2004
Antibodyrabbit anti-GFP polyclonal antibodyInvitrogenCat# A11122
RRID:AB_221569
1:200 IF
Antibodychicken anti-GFP polyclonal antibodyAvesCat# GFP-1020
RRID:AB_10000240
1:200 IF
Antibodyrabbit anti-GFAP polyclonal antibodyDakoCat# Z0334
RRID:AB_10013382
1:100 IF
Antibodymouse anti-NeuN monoclonal antibodyMilliporeCat# MAB377
RRID:AB_2298772
1:50 IF
Antibodyrat anti-L1 monoclonal antibodyMilliporeCat# MAB5272
RRID:AB_2133200
1:100 IF
Antibodyrabbit anti-Calretinin polyclonal antibodyMilliporeCat# MAB50541:500 IF
Antibodygoat anti-Nrp1 polyclonal antibodyR and DCat# AF566
RRID:AB_355445
2 µg/ml for functional blocking
Antibodygoat anti-VEGFR2
polyclonal antibody
R and DCat# AF644
RRID:AB_355500
1:15 IF
Antibodyrabbit anti-VEGFR2 monoclonal antibodyCell SignalingCat# 2479
RRID:AB_2212507
1:1000 WB
Antibodyrabbit anti-(phospho)VEGFR2 (Y1175)Cell SignalingCat# 2478
RRID:AB_331377
1:500 WB
Antibodyrabbit anti-(phospho)SFK (Y416) monoclonal antibodyInvitrogenCat# 44660G
RRID:AB_2533714
1:200 IF
Antibodymouse anti-beta-III-tubulin monoclonal antibodySigmaCat# T5076
RRID:AB_532291
1:150 IF
Antibodygoat anti-VE-Cadherin polyclonal antibodyR and DCat# AF1002
RRID:AB_2077789
1:1000 WB
Antibodydonkey anti-rabbit Alexa488 polyclonal antibodyJackson ImmunoresearchCat# 711-545-152
RRID:AB_2313584
1:1000 IF
Antibodydonkey anti-rabbit Alexa568 polyclonal antibodyLife TechnologiesCat# A10042
RRID:AB_2534017
1:1000 IF
Antibodydonkey anti-goat Alexa568 polyclonal antibodyMolecular ProbesCat# A11057
RRID:AB_142581
1:1000 IF
Antibodygoat anti-mouse Alexa568 polyclonal antibodyInvitrogenCat# A11031
RRID:AB_144696
1:1000 IF
Antibodygoat anti-rat Alexa568 polyclonal antibodyInvitrogenCat# A11077
RRID:AB_2534121
1:1000 IF
Antibodygoat anti-chicken Alexa488 polyclonal antibodyMolecular ProbesCat# A11039
RRID:AB_142924
1:1000 IF
Antibodydonkey anti-goat HRP polyclonal antibodyJackson ImmunoresearchCat# 705-035-147
RRID:AB_2313587
1:1000 IF
Antibodydonkey anti-rabbit HRP polyclonal antibodyJackson ImmunoresearchCat# 711-035-152
RRID:AB_10015282
1:1000 IF
Antibodydonkey anti-mouse HRP polyclonal antibodyJackson ImmunoresearchCat# 715-035-150
RRID:AB_2340770
1:1000 IF
Chemical compound, drugPhalloidin-TRITCSigmaCat# P1951
RRID:AB_2315148
1:400 IF
Chemical compound, drugdynasoreSelleckchemCat# S804780 µM for functional blocking
Chemical compound, drugDiI lipophilic tracer dyeMolecular ProbesCat# D-3911
Chemical compound, drugPP2CalbiochemCat# 5295731 µM for functional blocking
Chemical compound, drugPP3CalbiochemCat# 5295741 µM for functional blocking
Recombinant DNA reagentUtrCH-mCherry plasmidBurkel et al., 2007
Recombinant DNA reagenthVEGFR2-GFP fusion plasmid
Commercial assay or kitVEGF ELISA kitR and DCat# MMV00

Animals

Nervous system-specific VEGFR2 conditional knockout mice (Nes-cre;Kdrlox/-) were generated by crossing the previously described Kdrlox/- mice (Haigh et al., 2003) with Nestin-cre mice (Tronche et al., 1999). Vegfr2-GFP mice were obtained from Rüdiger Klein and previously described (Ema et al., 2006). Kdr-EphrinB2 double-heterozygous conditional compound mice (Nes-cre; Kdrlox/+;EfnbB2lox/+) were generated by crossing mice double-homozygous floxed for Kdr and ephrinB2 (Kdrlox/lox;Efnb2lox/lox) to mice bearing one copy of the Nestin-cre transgene (Nes-cre) (see further accompanying manuscript by Harde et al., 2019). Single-heterozygous mice, where either only one Kdr allele (Nes-cre;Kdrlox/+) or only one ephrinB2 allele (Nes-cre;Efnb2lox/+) was deleted, were generated in a similar manner as the compound mice (see further accompanying manuscript by Harde et al., 2019). All mice were bred in the C57BL/6 background, except the Vegrf2-GFP mice, which were maintained in the ICR/HaL background. Wildtype C57BL/6NRJ mice were obtained from Janvier Laboratories. All animal experiments were conducted in accordance with the German institutional guidelines and ethical committees (references refs: FU1090, T60/12, T79/13, T49/15, T69/14, T46/16, T36/17, T48/18, and I19/13).

Antibodies

Following primary antibodies were used for Immunofluorescence (IF) and Western blotting (WB): rabbit anti-GFP (1:200 IF, Invitrogen, A11122), chicken anti-GFP (1:200 IF, Aves, GFP-1020), rabbit anti-GFAP (1:100 IF, Dako, Z0334), mouse anti-NeuN (1:50 IF, Millipore, MAB377), rat anti-L1 (1:100 IF, Millipore, MAB5272), rabbit anti-Calretinin (1:500 IF, Millipore, MAB5054), goat anti-Nrp1 (2 µg/ml for function blocking, R and D, AF566), goat anti-VEGFR2 (1:15 IF, R and D, AF644), rabbit anti-VEGFR2 (1:1000 WB, Cell Signaling, 2479), rabbit anti-(phospho)VEGFR2 (Y1175) (1:500 WB, Cell Signaling, 2478), rabbit anti-(phospho/SFK (Y416) (1:200 IF, Invitrogen, 44660G), mouse anti-beta-III-tubulin (1:150 IF, Sigma, T5076), goat anti-VE-Cadherin (1:1000 WB, R and D, AF1002), Phalloidin (1:400 IF, Sigma, P1951). The following secondary antibodies were used: donkey anti-rabbit Alexa488 (Jackson Immunoresearch, 711-545-152), donkey anti-rabbit Alexa568 (Life Technologies, A10042), donkey anti-goat Alexa568 (Molecular Probes, A11057), goat anti-mouse Alexa568 (Invitrogen, A11031), goat anti-rat Alexa568 (Invitrogen, A11077), goat anti-chicken Alexa488 (Molecular Probes, A11039), donkey anti-goat HRP (Jackson Immunoresearch, 705–0350147), donkey anti-rabbit HRP (Jackson Immunoresearch, 711-035-152), donkey anti-mouse HRP (Jackson Immunoresearch, 715-035-150).

In situ hybridization and immunohistochemistry

Request a detailed protocol

Wildtype and Vegfr2-GFP pups at the indicated ages received an anesthetic overdose by intraperitoneal injection of ketamine (180 mg/kg of body weight; Ketavet) and xylazine (10 mg/kg of body weight; Rompun) and were transcardially perfused first with PBS followed by 4% PFA. Brains were collected in 4% PFA and fixed overnight at 4°C. For cryosections, fixed and washed brains were placed in 30% sucrose and subsequently frozen in NEG-50 (Richard-Allan Scientific). For in situ hybridization, 20 µm cryosections were hybridized with DIG-labeled RNA probes for VEGFR2 (Kdr), VEGF (Vegfa) and Nrp1 mRNA as already described (Ruiz de Almodovar et al., 2011). For IF, cryosections were blocked for 1 hr at RT with PBS containing 1% BSA, 2% normal goat or normal donkey serum (Dianova) and 0.3% Triton X-100. Primary antibodies in blocking solution were added overnight at 4°C. Secondary antibodies in blocking solution were added for 2 hr at room temperature (RT), followed by nuclear staining with TO-PRO3 (1:1000, Life technologies). For combination with IF, ISH was performed first. Following postfixation, sections were washed and blocked for 4 hr at RT with PBS containing 1% BSA, 5% normal goat serum and 0.3% Triton. Subsequently, GFAP (1:100) or NeuN (1:50) antibodies were added in blocking solution overnight at 4°C. Corresponding secondary antibodies were incubated for 2 hr at room temperature followed by nuclear staining with TO-PRO3 (1:1000, Life technologies). Immunostainings were examined using a Zeiss confocal microscope (LSM510).

Quantification of Phospho-SFK fluorescence

Request a detailed protocol

25.000 primary hippocampal neurons per coverslip were cultured for 24 hr and treated for 30 min with 80 µM dynasore or 0.1% DMSO. 5 min prior to stimulation with VEGF. At the same time, 1 mM sodium orthovanadate was added to the cultures. Neurons were stimulated for 5 min with 100 ng/ml VEGF and subsequently fixed in 4% PFA/4% sucrose supplemented with 1 mM sodium orthovanadate for 20 min at room temperature. Immunostaining with rabbit anti-(phospho) SFK (Y416) antibody followed by Alexa-488 conjugated secondary antibody was used to detect phosphorylated SFKs in the axonal growth cones and axons of the neurons. Phalloidin was used to visualize the actin cytoskeleton. Image stacks were acquired at a confocal microscope (Zeiss LSM510, 63x objective) and fluorescence intensities of maximal intensity projections were quantified using FIJI. At least 38 growth cones and axons in three independent experiments were analyzed and statistical differences were assessed by unpaired students t test.

Analysis of hippocampal gross morphology in vivo

Request a detailed protocol

20 μm thick cryosections from P10 pups were immune-stained for nuclei using TO-PRO3 (1:1000, Life technologies), neuronal nuclei using NeuN and for the axonal marker L1. Image stacks were obtained with a confocal microscope (Zeiss LSM510, 10x objective). To determine neuronal density, high magnification images of the NeuN stained CA3 region were acquired and the number of neuronal nuclei counted from three different brain sections and four animals per genotype. Based on the L1 staining, the dimensions of the hippocampal layers were determined by measuring the thickness of each layer and normalizing it to the total thickness of all layers as previously described (Terauchi et al., 2010).

Analysis of axon branching in vivo

Request a detailed protocol

Control, Nes-cre;Kdrlox/- and Nes-cre;Kdrlox/+;EfnbB2lox/+ pups sacrifized, the brains removed and embedded in 5% low-melting temperature agarose. 250 μm transverse sections were obtained using a vibratome (Leica) and collected in ice-cold PBS. The PBS was removed and sections fixed for 2 hr in 4% PFA at RT. Following three washes with PBS, single sections were transferred to a 12 well plate. four sections per brain and experiment were used. Under a stereomicroscope, all liquid was carefully removed with a P1000 pipette and a tiny DiI crystal (Molecular Probes, D-3911) was inserted into the CA3 pyramidal cell layer with the help of a tungsten needle attached to a glass pasteur pipette. The slices were covered with PBS and incubated at 37°C for 16–20 hr. Subsequently, the sections were stained with TO-PRO3 (1:1000, Life technologies) and mounted. Image stacks of Schaffer collaterals in the stratum radiatum of the CA1 region were acquired using a confocal microscope (Zeiss LSM510, 20x objective). Single axons were traced using the FIJI plugin Simple Neurite Tracer and the number and length of axon branches were determined.

Transmission electron microscopy

Request a detailed protocol

For electron microscopy experiments, mice were perfused with 2.5% glutaraldehyde and 2% PFA. First, 400 μm thick vibratome section were cut, and those containing the hippocampus were collected. The vibratome sections were rinsed in cacodylate buffer (3–10 min) and then incubated in 0.5% OsO4 in cacodylate buffer for 30 min. Afterwards the tissue was dehydrated in a series of increasing concentrations of acetone (30%, 50%, 70%, 85% and 2x 100% for 10 min each). The tissue was pre-incubated in a 1:1 mixture of acetone and Epon for 30 min, followed by incubation in pure Epon overnight at RT. On the next day the tissue was transferred into fresh Epon for polymerization for at least 48 hr in a heating cabinet at 60C. Vibratome sections were flat-embedded in Epon between two plastic sheets, cut out after polymerization and glued onto Epon blocks. Ultrathin sections were collected on slot grids, which were coated with 2% polyvinyl butyral in chloroform. An ultramicrotome (Ultracut UCT, Leica) with a diamond knife (Ultra 45, Diatome, Hatfield) was used to cut ultrathin sections at a thickness of 50-60 nm. The ultrathin sections were transferred with an eyelash onto the grids. Afterwards, the grids were counterstained with uranyl acetate and lead citrate using the automatic staining machine EM Stain (Leica). The transmission electron microscope (TEM) Leo 912 AB Omega (Zeiss) was used to acquire images from ultrathin sections. Images were taken by using a CCD camera (Troendle) in combination with the Image SP software (Troendle). Electron microscopy images of 20–50 μm from the middle region in stratum radiatum of the hippocampal CA1 region were acquired at a magnification of 5000x. At least eight images from different ultra-thin sections were acquired per mouse, which corresponds to over 700 counted synapses per mouse. Synapses were identified and counted based on the presence of presynaptic vesicles, the synaptic cleft and postsynaptic density.

Electrophysiology

Hippocampal slice preparation

Request a detailed protocol

Transverse 400-um-thick hippocampal slices were made from pups (P8-P10) using a vibratome (Leica VT1200). For dissection and slicing, sucrose-based ACSF containing (in mM) 87 NaCl, 25 NaHCO3, 2.5 KCl, 1.25 NaH2PO4, 10 glucose, 0.5 CaCl2, 7 MgCl2, and 75 sucrose was used. For storage and superfusion of slices, ACSF solution containing (in mM) 125 NaCl, 25 NaHCO3, 2.5 KCl, 1.25 NaH2PO4, 25 glucose, 2 CaCl2, 1 MgCl2, 2 TTX and 20 bicuculline was used. Slices were incubated at 34°C for 30 min after slicing and subsequently stored at room temperature. Both solutions were equilibrated with 95% O2 and 5% CO2 throughout the procedure.

Electrophysiological recordings

Request a detailed protocol

Whole-cell patch-clamp recordings were made from CA1 pyramidal neurons with pipettes with a resistance of 4–6 MΩ using an Axopatch 200B amplifier. The internal solution of pipettes contained 120 mM potassium gluconate, 20 mM KCl, 0.1 mM EGTA, 2 mM MgCl2, 4 mM Na2ATP, 0.5 mM GTP, 10 mM HEPES, 7 mM disodium phosphocreatine, (pH 7.2, 300 mOsm). Neurons were voltage clamped at –70 mV while the series resistance was left uncompensated during the recordings. mEPSCs were analyzed offline using Stimfit software (Guzman et al., 2014) employing a template matching algorithm. Selected mEPSC events were individually screened with an amplitude threshold of >5 pA and an exponential decay. Recordings were started 5 min after patching and the minis are usually recorded for 10 min. Statistical differences between experimental conditions were determined by the unpaired students t-test. The experimenter was blind to the identity of the mice during acquisition.

Preparation of neuronal cultures

Request a detailed protocol

Hippocampal neuron cultures were prepared from E16.5 mouse embryos as described previously (Segura et al., 2007) and cultured on coverslips or tissue culture plates coated with poly-L-lysine (1 mg/ml, Sigma) and laminin (20 µg/ml, Sigma). Neurons were maintained in neurobasal medium (Life technologies) supplemented with B27 (Invitrogen) and 0.5 mM Glutamine (GIBCO). For morphological analysis at 3 DIV neurons were cultured at a density of 20.000 neurons/coverslip.

Neuronal transfection

Request a detailed protocol

For subcellular localization of the VEGFR2-GFP and tracking of actin dynamics, neurons were plated on four well Labtek chambers (Thermo Scientific, 155383) at a density of 100.000 neurons/well. At 1 DIV neurons were co-transfected with an hVEGFR2-GFP fusion plasmid and UtrCH-mCherry plasmid using Lipofectamine 2000 (Invitrogen) following manufacture instructions. mCherry-UtrCH plasmid was described in Burkel et al. (2007) and provided by Dr. Ulrike Engel (University of Heidelberg Heidelberg, Germany).

Morphological analysis of neuronal cultures

Request a detailed protocol

1 DIV hippocampal neurons were stimulated for 48 hr with or without 100 ng/ml VEGF, fixed for 20 min in 4% PFA/4% sucrose at RT and stained for beta-III-tubulin. Where indicated, neurons were pre-treated with 2 µg/ml α-VEGFR2 blocking antibody (αVEGFR2 (DC101), gift from Dr. T Schmidt), anti-Nrp1 blocking antibody (αNrp1), respective IgG control, 1 µM PP2 (Calbiochem) or 1 µM PP3 (Calbiochem) for 2 hr prior to VEGF stimulation. The neurons were imaged using a Zeiss Axiovert 200 epifuorescence microscope. The longest neuronal process was defined as the axon and the number and length of axonal branches were quantified using the ImageJ plugin NeuronJ.

Time-lapse video microscopy

Request a detailed protocol

For time-lapse video microscopy 50.000 neurons/well were plated in a 12 well plate. At 1 DIV neurons were pre-treated for 2 hr with 2 µg/ml α-VEGFR2 blocking antibody, respective IgG control, 1 µM PP2, 1 µM PP3. To block internalization, neurons were pre-treated with 80 µM dynasore or the equal volume of DMSO vehicle control for 30 min. Subsequently, 50 ng/ml VEGF was added and at least 10 neurons per condition were imaged simultaneously at the Nikon Imaging Center, Heidelberg, over the course of 4 hr at 4 min intervals using an inverted Nikon Ti microscope with Nikon perfect focus system and encased by an environmental box to keep temperature, CO2 and humidity constant. Phase contrast images were acquired using a Nikon Plan Fluor 10x NA 0.3 phase contrast objective. Newly formed and growing axon branches were defined as extension, while loss and shrinkage of branches were defined as retraction. The number and length of extending and retracting axon branches were quantified using the ImageJ plugin NeuronJ and the growth rates calculated. The number of newly formed filopodia was quantified within the first 80 min of the time-lapse movies.

Dynamic VEGFR2-GFP movement and actin tracking

Request a detailed protocol

1 DIV hippocampal neurons were double transfected with VEGFR2-GFP and mCherry-UtrCH. At 3 DIV neurons were imaged using an inverted Nikon Ti microscope with objective TIRF illumination (Nikon Apo TIRF 60x NA 1.49) and equipped with an on-stage incubation chamber from TokiaHit for temperature, CO2- and humidity control. GFP and mCherry were imaged sequentially with 488 nm and 561 nm excitation and emission filters 520/40 nm and 630/60 nm onto an EM-CCD (Andor Xion). To follow actin nucleation and patch formation, mCherry-UtrCH was imaged for 2 min at a frame rate of 1.6 fps before and 5 min after stimulation with 100 ng/ml VEGF. To investigate protrusion dynamics, mCherry-UtrCH images were acquired every 5 s over the course of 10 min before and after stimulation with 100 ng/ml VEGF. VEGFR2-GFP trafficking was observed for 1 min at a frame rate of 1.6 fps before and 5 min after stimulation with 100 ng/ml VEGF.

Western Blotting

Request a detailed protocol

Neurons were harvested in ice-cold lysis buffer (150 mM NaCl, 20 mM Tris, 5 mM EDTA, 10% glycerol, 1% Igepal, protease and phosphatase inhibitor tablets (Roche)), lysed for 30 min on ice and cleared by centrifugation at 15.000 rpm for 15 min at 4C. Protein concentrations were determined using Roti-Quant Universal kit (Roth). Protein samples were loaded on 7.5% or 10% polyacrylamide gels for SDS-PAGE. Following transfer to 0.2 µm pore-size PVDF membranes, blots were blocked in 5% BSA in TBST for 1 hr at RT. Blots were then probed with antibodies against VEGFR2, (phospho-) VEGFR2 (Y1175), GFAP and VE-Cadherin. After incubation with appropriate HRP-conjugated secondary antibodies, bands were detected with ECL Western Blotting detection reagent (Millipore) and visualized using ImageQuant LAS 4000 (GE Healthcare).

Reverse transcription PCR

Request a detailed protocol

Total RNA was prepared from neuronal and primary endothelial culture, as well as from postnatal brains using Superscript II (Invitrogen). Reverse transcription-PCR was carried out using Q5 High Fidelity DNA polymerase (New England Biolabs) according to manufacturer instruction. Forward (F) and reverse (R) primers used were as follows: CD31: F, 5’-AGCCTAGTGTGGAAGCCAAC-3’; R, 5’- AGCCTTCCGTTCTCTTGGTG-3’; GAPDH: F, 5’- GGTCCTCAGTGTAGCCCAAG-3’; R, 5’-AATGTGTCCGTCGTGGATCT-3’; GFAP: F, 5’-AAATCCGTGTCAGAAGGCCA-3’; R, 5’-TAATGACCTCACCATCCCGC-3’; Neuropilin 1: F, 5’-AACGTGTGCTTCTGTCCAAC-3’; R, 5’-AAGGGAGAGGGGAAAGCAAT-3’; VEGF: F, 5’-CGTTCACTGTGAGCCTTGTT-3’; R, 5’-CTTGGCTTGTCACATCTGCA-3’; Vegfr1: F, 5’-ACCCAGGAGTGCAAATGGAT-3’; R, 5’-TGTTGGACGTTGGCTTGAAG-3’; Vegfr2 (Kdr): F, 5’-TTCACAGTCGGGTTACAGGC-3’; R, 5’-CTGCCGACGTTCCTCTCTTT-3’.

Quantitative RT-PCR

Request a detailed protocol

Total RNA was prepared from neuronal cultures and reverse transcription was performed using Superscript II (Invitrogen). Expression levels of VEGFR2 mRNA were quantified by real-time RT-PCR using SYBR-Mix (Applied Biosystems). GAPDH was served as an endogenous control. Following forward (F) and reverse primers (R) were used: Vegfr2 (Kdr): F, 5’- TTCACAGTCGGGTTACAGGC-3’; R, 5’- CTGCCGACGTTCCTCTCTTT-3’; GAPDH: F, 5’- GGTCCTCAGTGTAGCCCAAG-3’; R, 5’- AATGTGTCCGTCGTGGATCT-3’.

Quantification of VEGF protein levels by ELISA

Request a detailed protocol

Conditioned medium of hippocampal neuronal cultures was collected at various times in vitro and processed for further measurements of VEGF levels using the Quantikine mouse VEGF ELISA kit (R and D, MMV00).

Statistical analysis

Request a detailed protocol

Results are expressed as the mean ± SEM, unless stated otherwise. To calculate statistical significance the unpaired Student’s t-test, one-way ANOVA followed by Tukey’s multiple comparisons test and two-way ANOVA followed by Dunnett multiple comparisons test were calculated using Prism software. An excel file including all detailed statistical information of the data presented in the study is provided as a source data file.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
    Early specification and autonomous development of cortical fields in the mouse Hippocampus
    1. S Tole
    2. C Christian
    3. EA Grove
    (1997)
    Development 124:4959–4970.
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43

Decision letter

  1. Didier Y Stainier
    Senior Editor; Max Planck Institute for Heart and Lung Research, Germany
  2. Gou Young Koh
    Reviewing Editor; Institute of Basic Science and Korea Advanced Institute of Science and Technology (KAIST), Republic of Korea
  3. Injune Kim
    Reviewer; Korea Advanced Institute of Science and Technology, Republic of Korea
  4. Bassem A Hassan
    Reviewer; ICM, France

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

It is known that VEGF-A signaling plays a certain role neuronal development. In this study, the authors uncover molecular and mechanical insights into how neuronal VEGF-A-VEGFR2 regulates axonal branching and synapse formation in the different subsets of hippocampal neurons. Overall the study is intriguing and the results are rigorous and solid.

Decision letter after peer review:

Thank you for submitting your article "VEGFR2 controls axon branching of hippocampal neurons during development" for consideration by eLife. Your article has been reviewed by Didier Stainier as the Senior Editor, a Reviewing Editor, and three reviewers. The following individuals involved in review of your submission have agreed to reveal their identity: Injune Kim (Reviewer #1); Bassem A Hassan (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

The comments are relatively favorable, complimentary and constructive. However, the authors are required to address following.

1) It is essential to further clarify the complexities of VEGFR2/VEGFR2 ligands – VEGF-A, C and D in the axon branching using a primary neuron culture.

2) Please address the points of all three reviewers point-by-point in a data-driven manner or with further analyses.

I believe the authors are capable of addressing most of the comments, but please provide the reasons for not implementing the suggested changes where necessary.

Reviewer #1:

This work elucidated the role of the VEGF signaling in CA3 hippocampal axon branching. Although the experiments were designed and performed well, the complex behavior of axon responding to VEGF activation and the loss of VEGFR2 remains to be clarified. This issue needs to be further discussed for improving the manuscript.

1) Together with the accompanying manuscript by Harde et al., this work shows the opposite effect of the loss of VEGFR2 on the branching of axons and dendrites in hippocampal neurons. How does the axon sense the VEGF cue in the absence of VEGFR2? What about other VEGF receptors?

2) Does the VEGFR2 interact with EphrinB2 for its internalization in the axon?

3) Hippocampal neurons appear to be in proximity of blood vessels. Does the loss of neuronal VEGFR2 gene affect angiogenesis in neighboring vessels due to the potential increase of free VEGF level?

Reviewer #2:

Urban et al. present convincing and clear evidence that neuronal VEGF signaling via the VEGFR2 receptor regulates axonal branching. In an interesting twist both gain and loss of function of VEGFR2 increase axonal branching, but with what appear to be opposite consequences on synapse formation.

Overall, the manuscript is very well written, the evidence clear and convincing, and interpretation consistent with the data. In principle, the manuscript can be published in eLife in its current form. I do, however have some suggestion for the authors, which I think would add the "completeness" of the story.

1) The authors show that loss of VEGFR2 reduces the number of synapses. Perhaps I missed this, but I couldn't find data as what happens to synapses upon stimulation with VEGF. Do they increase in number, as would be predicted from the increase in what appear to be mature branches?

2) In reference to the accompanying manuscript, is VEGFR2 internalization also required for its role in axonal branching?

3) Very recent work from the Hiesinger lab (Ozel et al., Dev Cell 2019) shows how filopodial dynamics regulate synapse formation in Drosophila. Given the role of VEGFR2 shown here in actin regulation and branching, discussing potential regulation of filopodial dynamics by VEGFR2 in this context might be interesting.

4) In my personal opinion terms like "X controls Y process" are outdated and reflect linear deterministic models that are no longer relevant in modern biology and should in my view no longer be used. Perhaps a term like "regulates" or "contributes to" are more appropriate in the title, but I leave that to the discretion of the authors.

Reviewer #3:

The manuscript by Urban et al. investigates the role of VEGF-A signaling in CA3 neurons in the hippocampus with regard to axon branching. Expression of both ligand and receptor VEGFR2 is investigated, then primary culture of neurons is used to show that VEGF stimulation leads to excess branching. However, in vivo the neuron-specific conditional genetic loss of the receptor also leads to excess branches, although these are thought to be less mature. A blocking antibody to VEGFR2 recapitulates the excess branching seen in vivo in primary cultures. The signaling is linked to Src-like kinases using inhibitors, and some very elegant cell biology indicates that VEGFR2 relocalizes to actin patches upon stimulation.

Overall the data is rigorous and uses both in vitro and in vivo approaches. It is known that VEGF-A signaling is important in neuronal development. The major novelty here is that the neuronal loss of VEGFR2 leads to excess branching of a subset of hippocampal neurons that does not appear to lead to productive synapse formation, while VEGF-A stimulation in primary neurons also leads to excess branching. The authors speculate that the lack of synapses may lead to a feedback loop with more short axon branches. It would strengthen the work to test some other ideas – for example perhaps VEGF-A is not the relevant ligand in vivo since VEGF-C and VEGF-D also bind to VEGFR2. What are the effects of alternate ligands on the parameters in primary neuron culture?

The other aspect of the novelty is the idea that VEGFR2 relocalizes upon VEGF-A stimulation to actin patches, where it supposedly promotes neurite outgrowth. It would strengthen the work to move the kymographs to the main figures and label better to highlight these novel findings, and if possible, to include the movies from which the kymographs were derived (the co-labeling of actin and VEGFR2). Also, would strengthen to show spatial changes by a different approach – is it possible to use Proximity Ligation Assay or pull down to investigate VEGFR2 association with something localized to the actin patches and/or filopodia?

Also, as described in review of co-submitted paper, it would be useful to better integrate the findings from the co-submitted paper in the Discussion section – it appears that endocytosis is important for dendrite biology, is it also important for the axon branching effects of VEGF/VEGFR2? Are SPKs important for dendrites as well as axons? It is not clear whether axons and dendrites use different mechanisms or whether they were tested differently.

https://doi.org/10.7554/eLife.49818.sa1

Author response

The comments are relatively favorable, complimentary and constructive. However, the authors are required to address following.

1) It is essential to further clarify the complexities of VEGFR2/VEGFR2 ligands – VEGF-A, C and D in the axon branching using a primary neuron culture.

It is indeed plausible that VEGF-C and VEGF-D might contribute to and affect axonal branching in the hippocampus during development. To address this concern we checked their expression pattern in the developing hippocampus as well as their effect in primary hippocampal neuron cultures. For detailed description and interpretation of the new results please see our response to reviewer #3, comment #1.

2) Please address the points of all three reviewers point-by-point in a data-driven manner or with further analyses.

In our response letter we address point-by-point all reviewers’ comments.

I believe the authors are capable of addressing most of the comments, but please provide the reasons for not implementing the suggested changes where necessary.

Reviewer #1:

This work elucidated the role of the VEGF signaling in CA3 hippocampal axon branching. Although the experiments were designed and performed well, the complex behavior of axon responding to VEGF activation and the loss of VEGFR2 remains to be clarified. This issue needs to be further discussed for improving the manuscript.

We thank the reviewer for his comments. Below we have addressed point by point these general concerns. In addition, as indicated by this reviewer, in the Discussion section of our manuscript, we have further discussed the complex behavior of axons upon activation as wells as loss of VEGFR2.

1) Together with the accompanying manuscript by Harde et al., this work shows the opposite effect of the loss of VEGFR2 on the branching of axons and dendrites in hippocampal neurons. How does the axon sense the VEGF cue in the absence of VEGFR2? What about other VEGF receptors?

We thank the reviewer for his comment. In the submitted manuscript, we show that stimulation of hippocampal neurons with VEGF increased axon branching via activation of its cognate receptor VEGFR2. At the same time, blockage, as well as knockout of VEGFR2 showed the same phenotype of increased axonal branching, both in vitro and in vivo (in Nes-cre;Kdrlox/- mice). In our manuscript we speculate that the increased axon branching observed upon loss of VEGFR2 is a compensatory mechanism. In this regard, as this reviewer suggested, it might be possible that interference with VEGFR2 function shifts the signaling of VEGF to other VEGF receptors, thus in turn being the reason for the increase in axon branching. Equally likely, the compensatory mechanism that is activated could be totally independent of VEGF receptor signaling.

To address whether VEGF could be binding to other VEGF receptors in the absence of VEGFR2 we focused on VEGFR1 (Flt1) and Neuropilin 1 (Nrp1), the other two receptors via which VEGF can induce a signal transduction cascade independent of VEGFR2 (de Vries et al., 1991; Soker et al., 1998; Roth et al., 2016).

VEGFR1

We have analyzed the expression of Vegfr1 in the hippocampus at similar time points as we did for VEGFR2 (late embryonic (E18.5) and early postnatal stages P4 and P8) using ISH. At these time points Vegfr1 mRNA is primarily detected in blood vessel structures (Author response image 1A and 1B). Similar to our results, the data provided by the Allen Brain Atlas for the developing mouse brain (http://developingmouse.brain-map.org) shows Vegfr1 mRNA expression primarily in blood vessels in the hippocampus (Author response image 1C).

Author response image 1
Flt1 is expressed by blood vessels during hippocampal development.

(A, B) Flt1-ISH was performed in the mouse hippocampus at E18.5, P4 and P8 developmental stages. Antisense (AS, in A) and sense (SE, in B) signals are shown. CA1 and CA3 hippocampal areas, as well as the DG are indicated. Lower panel in A) shows higher magnification inserts. Scale bars 500 µm (A – higher panel and B) and 250 µm (A – lower panel). C) Flt1-ISH from the Allen Brain Atlas, showing AS signal in the hippocampus at E18.5, P4 and P14. Scale bars 500 µm.

These data also correlates with our analysis shown in Figure 1—figure supplement 1D of our revised manuscript where no expression of Vegfr1 mRNA is detected in isolated hippocampal neurons at 1 DIV and 3 DIV. Based on this expression pattern we believe that VEGFR1 is a rather improbable receptor for VEGF in hippocampal neurons during development.

Nrp1

Nrp1 is a known mediator of the repulsive effect exerted by Semaphorins on axon development, which is also specifically described for the hippocampus (He and Tessier-Lavigne, 1997; Chedotal et al., 1998). Nrp1 is also described to function as co-receptor for VEGFR2 as well as to transduce VEGF signals independent of VEGFR2 (Soker et al., 1998; Roth et al., 2016; Erskine et al., 2011). To investigate whether Nrp1 could act as a receptor for VEGF and regulate axonal branching in hippocampal neurons, we first analyzed its expression pattern during hippocampal development using ISH. We detected Nrp1 mRNA in the CA1, CA2 and CA3 hippocampal regions, as well as in the dentate gyrus at late embryonic (E18.5) and early postnatal (P4 and P8) stages (Figure 5—figure supplement 4A of our revised manuscript). This is in line with our expression analysis in isolated hippocampal neurons (1 DIV and 3 DIV; Figure 1—figure supplement 1D of our revised manuscript) and with previous publications showing the expression of Nrp1 mRNA in the hippocampus and neocortex during development (Chedotal et al., 1998; Kawakami et al., 1996). The spatiotemporal expression of Nrp1 is similar to the one of Vegfr2 and Vegf, raising the possibility that Nrp1 can act either alone or together with VEGFR2 to mediate VEGF signaling in the developing hippocampus.

First, to elucidate whether VEGF could also signal via Nrp1 to modulate axon branching, we blocked Nrp1 using a previously described α-Nrp1-neutralizing antibody (Ruiz de Almodovar et al., 2011) and analyzed the VEGF-mediated effect on axonal branching as before. While axon branch number and length was significantly increased in control neurons upon VEGF stimulation, this effect was not blocked in the presence of α-Nrp1-neutralizing antibody (Figure 5—figure supplement 4B and C of our revised manuscript). This result suggests that Nrp1 is dispensable for VEGF-induced axon branching.

Second, we also considered the possibility that the increase in axon branching upon α-VEGFR2 treatment could be due to a swift towards VEGF-Nrp1 signaling. To test this hypothesis, we co-administered α-Nrp1-neutralizing antibody together with αVEGFR2 in hippocampal neurons and analyzed axon branching in the different conditions. Consistent with our data, α-VEGFR2 led to an increase in axon branch number and length (Figure 5—figure supplement 4D and C of revised manuscript). This effect was however not blocked with the addition of the α-Nrp1-neutralizing antibody (Figure 5—figure supplement 4D and C of revised manuscript), indicating that Nrp1 is not the receptor that leads to increased axon branching upon inactivation of VEGFR2.

Collectively, these data show that Nrp1 does not mediate the effect of VEGF in axon branching by itself, nor does it represent a strong compensatory mechanism that takes over when VEGFR2 is blocked. Our data also suggests that this compensatory mechanism might be independent of VEGF signaling. In the revised version of our manuscript we have now integrated the data for Nrp1 in Figure 5—figure supplement 4 of the revised manuscript.

2) Does the VEGFR2 interact with EphrinB2 for its internalization in the axon?

To address the reviewer’s question of whether VEGFR2 interacts with EphrinB2 for its internalization in the hippocampal axon, similar as it occurs in dendrites (shown in the accompanying manuscript by Harde et al., we isolated hippocampal neurons from EphrinB2 knockout mice (from here on EphrinB2-/-) and analyzed the number and length of axonal branches in the presence and absence of VEGF. Stimulation of hippocampal neurons from EphrinB2-/- with VEGF resulted in a similar increase of axon branch number and length as in neurons from control littermates (Figure 6—figure supplement 1C and D of revised manuscript), indicating that EphrinB2 is not necessary for the VEGF-induced effect on axonal branching.

To confirm this in vitro data, we further analyzed CA3 hippocampal axon branching in vivo in VEGFR2 and EphrinB2 compound mice (from here on Nes-cre;Kdr+/-;ephrinB2+/-), which showed a phenotype in dendritic branching in the accompanying paper (Harde et al.,) Using the same DiI-labeling and quantification approach we used for analyzing CA3 axon branching in Nes-cre;Kdrlox/-, we labeled single axons of CA3 pyramidal neurons in wildtype and Nes-cre;Kdr+/-;ephrinB2+/- littermates and analyzed their branching. This analysis revealed no significant differences between the genotypes (Figure 6—figure supplement 1E of revised manuscript), indicating that the interaction between VEGFR2 and EprhinB2 is not necessary for the VEGF-mediated regulation of hippocampal axon branching.

Altogether, these new results show that VEGF utilizes different receptor complexes and molecular mechanisms to regulate axon and dendritic branching in hippocampal neurons. The new data generated with EphrinB2 are now integrated in the updated version of our manuscript (Figure 6—figure supplement 1 of revised manuscript).

3) Hippocampal neurons appear to be in proximity of blood vessels. Does the loss of neuronal VEGFR2 gene affect angiogenesis in neighboring vessels due to the potential increase of free VEGF level?

We understand the reviewer’s reasoning that upon loss of neuronal VEGFR2 expression there might be potential increase of free VEGF level, which subsequently might affect angiogenesis in neighboring vessels. A similar mechanism has been described before in the developing mouse retina (Okabe et al., 2014). To address this question, we analyzed the area covered by blood vessels in the hippocampus of Nes-cre;Kdrlox/- mice at P10 using ImageJ. We did not observe a significant difference in blood vessel density in the hippocampus of Nes-cre;Kdrlox/- mice when compared to their control littermates (Author response image 2A).

Author response image 2
Neuronal loss of VEGFR2 does not affect blood vessel density in the hippocampus.

(A) Blood vessels were stained using Isolectin-B4 at P10 in control and Nes-cre;Kdrlox/-animals. The area of Isolectin-B4-positive signal was quantified using ImageJ and normalized to the hippocampal area. Data are represented as mean ± SEM, >15 neurons from n=4 animals. n.s. not significant; unpaired Student’s t-test.

Reviewer #2:

Urban et al. present convincing and clear evidence that neuronal VEGF signaling via the VEGFR2 receptor regulates axonal branching. In an interesting twist both gain and loss of function of VEGFR2 increase axonal branching, but with what appear to be opposite consequences on synapse formation.

Overall, the manuscript is very well written, the evidence clear and convincing, and interpretation consistent with the data. In principle, the manuscript can be published in eLife in its current form.

We thank the reviewer for their positive comments.

1) The authors show that loss of VEGFR2 reduces the number of synapses. Perhaps I missed this, but I couldn't find data as what happens to synapses upon stimulation with VEGF. Do they increase in number, as would be predicted from the increase in what appear to be mature branches?

Indeed, in Figure 5 of our manuscript we show that despite the increase in axon branching upon loss of VEGFR2 in neurons, the number of functional synapses in the CA1 is reduced, suggesting that those branches are rather immature and/or nonfunctional. In our manuscript we did not analyze synapse formation in the axon upon stimulation with VEGF.

In the accompanying manuscript of Harde et al., the authors investigated the effect of VEGF on the dendritic spines of hippocampal neurons in vitro (which would be the postsynaptic side for the axon). They transfected primary hippocampal neurons from wildtype mice with a plasmid expressing EGFP at 11 DIV, stimulated the neurons with 50 ng/ml VEGF for 24h or 48h and analyzed spine morphology and number at 14 DIV. Quantification revealed a significant increase in the number of mature spines upon VEGF stimulation when compared to unstimulated control neurons. The authors further show that the effect of VEGF is mediated through activation of its receptor VEGFR2.

Previous studies have also attributed a role for VEGF in regulating synapse formation and synaptic plasticity in the adult hippocampus (De Rossi et al., 2016; Kim et al., 2008; Latzer et al., 2019; Licht et al., 2011). De Rossi et al. for example demonstrated that VEGF signaling via VEGFR2 increases the postsynaptic responses mediated by a specific type of glutamate receptors (GluNRs) in hippocampal neurons, thus highlighting the potential of VEGF signaling in neurons for regulating proper synaptic function (De Rossi et al., 2016). Using conditional transgenic systems to reversibly overexpress VEGF or block endogenous VEGF in the hippocampus of adult mice, Licht et al. demonstrated a crucial role for VEGF in modulating plasticity of mature neurons and subsequently regulating/affecting memory formation and LTP responses (Licht et al., 2011).

Altogether, although we did not specifically analyzed synapses in the axon upon VEGF stimulation, the above-mentioned studies point to a role for VEGF in synapse formation and plasticity.

2) In reference to the accompanying manuscript, is VEGFR2 internalization also required for its role in axonal branching?

To investigate whether VEGFR2 internalization is required for its role in axonal branching, we followed a similar approach as in the accompanying manuscript (Harde et al.). We inhibited VEGFR2 internalization with dynasore, a dynamin inhibitor that 6 blocks both clathrin-dependent and clathrin-independent endocytic pathways, and analyzed the VEGF-mediated effect on axonal branching. Hippocampal neurons from wildtype mice were pretreated with dynasore or DMSO (as control) at 1 DIV for 30 min, followed by stimulation with 50 ng/ml VEGF. Time-lapse videos were recorded over the course of 4h and the dynamics of axon branching was analyzed. Treatment with dynasore blocked the increase in branch number and length induced by VEGF when compared to control conditions (DMSO treated neurons) (Figure 6—figure supplement 1A-B of revised manuscript).

These results indicate that internalization of VEGFR2 is also required for the VEGF-mediated effect on axon branching. However, in contrast to what occurs for dendritic branching, the internalization of VEGFR2 required for regulating axon branching is not dependent of EphrinB2 as we did not observed any significant inhibition of VEGF-induced axon branching in primary hippocampal neurons isolated from EphrinB2 knockout mice (see response to reviewer #1, comment 2 and Figure 6—figure supplement 1C-E of revised manuscript).

3) Very recent work from the Hiesinger lab (Ozel et al., Dev Cell 2019) shows how filopodial dynamics regulate synapse formation in Drosophila. Given the role of VEGFR2 shown here in actin regulation and branching, discussing potential regulation of filopodial dynamics by VEGFR2 in this context might be interesting.

We thank the reviewer for pointing out this interesting study to us. Indeed, we show that loss of VEGFR2 changes filopodia dynamics resulting in an increased number of filopodia and increased axon branching. With this in mind and considering that we also observe a reduced number of functional synapses on CA1 pyramidal neurons, it could very well be the case that the presence of VEGFR2 is somehow required for the competitive distribution of synaptic building material between synaptogenic filopodia. If this is the case, based on our data, we could speculate that in the absence of VEGFR2 the synaptic building material is reduced or distributed aberrantly, allowing only the formation and maturation of a reduced number of synapses. As in its current stage this is all speculation, and perhaps a bit far from the focus of our study, we prefer not to discuss it in our manuscript, but will consider following up on this possibility in further studies.

4) In my personal opinion terms like "X controls Y process" are outdated and reflect linear deterministic models that are no longer relevant in modern biology and should in my view no longer be used. Perhaps a term like "regulates" or "contributes to" are more appropriate in the title, but I leave that to the discretion of the authors.

We thank the reviewer’s suggestion. We have modified the title accordingly and it now reads: “VEGF/VEGFR2 signaling regulates hippocampal axon branching during development.”

Reviewer #3:

The manuscript by Urban et al. investigates the role of VEGF-A signaling in CA3 neurons in the hippocampus with regard to axon branching. Expression of both ligand and receptor VEGFR2 is investigated, then primary culture of neurons is used to show that VEGF stimulation leads to excess branching. However, in vivo the neuron-specific conditional genetic loss of the receptor also leads to excess branches, although these are thought to be less mature. A blocking antibody to VEGFR2 recapitulates the excess branching seen in vivo in primary cultures. The signaling is linked to Src-like kinases using inhibitors, and some very elegant cell biology indicates that VEGFR2 relocalizes to actin patches upon stimulation.

Overall the data is rigorous and uses both in vitro and in vivo approaches.

We thank the reviewer for their positive comments. Below we have addressed point by point their general concerns.

It is known that VEGF-A signaling is important in neuronal development. The major novelty here is that the neuronal loss of VEGFR2 leads to excess branching of a subset of hippocampal neurons that does not appear to lead to productive synapse formation, while VEGF-A stimulation in primary neurons also leads to excess branching. The authors speculate that the lack of synapses may lead to a feedback loop with more short axon branches. It would strengthen the work to test some other ideas – for example perhaps VEGF-A is not the relevant ligand in vivo since VEGF-C and VEGF-D also bind to VEGFR2. What are the effects of alternate ligands on the parameters in primary neuron culture?

To address this question, we performed a thorough literature research on the expression and role of VEGF-C and VEGF-D ligands in the developing hippocampus. We also analyzed expression of VEGF-C and VEGF-D in the developing hippocampus ourselves and tested their effect on axon branching of isolated hippocampal neurons.

VEGF-C

VEGF-C binds primarily to VEGFR3 but it can also bind VEGFR2 after proper proteolytic cleavage (Joukov et al., 1996), leading to the formation and activation of VEGFR2/VEGFR3 heterodimers (Tammela and Alitalo, 2010; Koch and Claesson-Welsh, 2012; Thomas et al., 2013). Ward et al. described VEGF-C expression during embryonic and postnatal development in various brain regions (Ward and Cunningham, 2015). In the hippocampus, VEGF-C expression appears to be higher in embryonic stage E16 and declines in the maturing hippocampus, where it becomes restricted to the hilus and the granule cell layer (GCL) of the dentate gyrus (DG) (Ward and Cunningham, 2015).

To confirm that VEGF-C is expressed in the developing hippocampus, as indicated by Ward and Cunningham (2015), we micro-dissected the hippocampus from the developing mouse brain at P4 and P8 and performed an ELISA for VEGF-C. As positive controls we used the adult mouse heart and lung as VEGF-C is expressed in those tissues (Kukk et al., 1996). In contrast to Ward et al., we did not detect VEGF-C in the developing hippocampus at any of the analyzed stages, while its expression was positive in heart and lung (Author response image 3A). The difference between our results and Ward et al. could be potentially explained by the sensitivity of the different approaches used to detect VEGF-C protein.

Nevertheless, to not exclude the possibility that VEGF-C is expressed in low (undetected by ELISA) levels and thus could exert an effect in developing hippocampal neurons, we cultured primary hippocampal neurons and stimulated them with 100 ng/ml VEGF-C at 1 DIV over the course of 48 hours. Analysis of axonal morphology in response to VEGF-C stimulation from three independent experiments showed no significant increase of axon branch number or length when compared to control or VEGF (VEGF-A) stimulated neurons (Author response image 3C and 3D).

Author response image 3
VEGF-C and VEGF-D do not show the same growth promoting effect on hippocampal axon branches as VEGF-A.

(A, B) Hippocampal tissue was collected from mice at P4 and P8 stages. Heart and lung tissue was collected from adult mice. The concentration of VEGF-C (A) and VEGF-D (B) was measured in whole tissue lysates using ELISA. Data are represented as mean ± SEM, n>2 independent animals. nd not detectable. (C, D) 1 DIV hippocampal neurons were stimulated with 100 ng/ml VEGF-C, VEGF-D, VEGF-A or vehicle control for 48h. The number of axonal branches (C) and the average branch length (D) were analyzed. Data are represented as% of non-stimulated control. Mean ± SEM, >79 neurons from n=3 independent experiments. *p<0.05; **p<0.001; ***p<0.0001; one-way ANOVA.

VEGF-D

VEGF-D can also bind and activate VEGFR2 and VEGFR3 receptor (Achen et al., 1998; Davydova et al., 2016). VEGF-D expression has been described in the hippocampus in early postnatal and adult stages where its expression is regulated by neuronal activity and nuclear calcium signaling (Mauceri et al., 2011). Mauceri et al. showed that VEGF-D promotes dendritic maintenance, both in cultured hippocampal neurons as well as in the adult mouse hippocampus (Mauceri et al., 2011). Indeed, our own analysis of VEGF-D expression via ELISA also confirmed those previous results. As VEGF-D and its receptors VEGFR2 and VEGFR3 are expressed in the hippocampus, we set out to investigate a possible effect of VEGF-D on axonal development. We cultured primary hippocampal neurons and stimulated them with 100 ng/ml VEGF-D at 1 DIV over the course of 48hours (neurons from three independent experiments) and analyzed the axonal morphology. VEGF-D stimulation of primary hippocampal neurons did not lead to an increase in the branch number when compared to control and VEGF (VEGF-A) stimulated neurons. VEGF-D however induced a significant increase in the length of axon branches.

Altogether, our new results point to a potential effect of VEGF-D in regulating the length of hippocampal axon branches. Still, VEGF-A seems to be the main factor that, acting via VEGFR2, leads to an increase number of axon branches. We believe that describing the entire mechanism of how VEGF-D regulates axon branch growth, and whether it does so via VEGFR2 activation, would require substantially more detailed experiments, which could be performed in a follow up study. For this reason, we prefer to answer this part of the reviewer’s comment just in the response letter and not include the data in the current manuscript.

The other aspect of the novelty is the idea that VEGFR2 relocalizes upon VEGF-A stimulation to actin patches, where it supposedly promotes neurite outgrowth. It would strengthen the work to move the kymographs to the main figures and label better to highlight these novel findings, and if possible, to include the movies from which the kymographs were derived (the co-labeling of actin and VEGFR2).

We thank the reviewer for their suggestion. We have now moved the kymographs to the main figure (Figure 4 of revised manuscript) and included also the movies from which the kymographs were derived (Video 5, Video 6, Video 7 and Video 8).

Also, would strengthen to show spatial changes by a different approach – is it possible to use Proximity Ligation Assay or pull down to investigate VEGFR2 association with something localized to the actin patches and/or filopodia?

In the submitted manuscript we indeed investigated the specific localization and dynamics of VEGFR2 along the axon, as well as its potential association with actin patches and filopodia (Figure 4—figure supplement 1B-D of revised manuscript). For this aim, the approach we chose to follow was co-transfection of primary hippocampal neurons with a VEGFR2-GFP and a mCherry-UtrCH plasmid. This decision was based on the fact that (i) Utrophin is frequently used as a marker for stable and dynamic F-actin, both in fixed and living cells (Burkel, von Dassow and Bement, 2007; Melak, Plessner and Grosse, 2007); and that (ii) the calponin homology domains of utrophin interacts with adjacent actin subunits, thus further confirming that utrophin is a reliable marker for actin filaments/patches (Lin at al., 2007).

Our analysis confirmed the co-localization of VEGFR2 with utrophin in primary hippocampal neurons in vitro, suggesting an association between VEGFR2 and the actin cytoskeleton and a role of VEGFR2 in actin remodeling (Figure 4A). We therefore believe that the approach chosen is appropriate to show VEGFR2 localization in actin patches/filopodia.

In support of such an association, the direct effect of VEGFR2 on the actin cytoskeleton has been previously described in the context of angiogenesis, where VEGFR2 signaling leads to actin remodeling by phosphorylation of profilin-1 (Pfn-1), a small actin-binding protein, either directly or through Src tyrosine kinases (Fan et al., 2012). Actin remodeling mediated via VEGF/VEGFR2 signaling has also been described in the neuronal growth cones of chicken dorsal root ganglia (Olbrich et al., 2007).

Taken together, these data support a role of VEGFR2 in regulating actin dynamics and strengthen our results on a direct implication of VEGF/VEGFR2 signaling on actin remodeling in the axons hippocampal neurons.

Also, as described in review of co-submitted paper, it would be useful to better integrate the findings from the co-submitted paper in the Discussion section – it appears that endocytosis is important for dendrite biology, is it also important for the axon branching effects of VEGF/VEGFR2?

In the revised version of the manuscript, we have better integrated the findings from the co-submitted paper in the Discussion section.

We also checked whether internationalization of VEGFR2 is required for axon branching. For this, we followed a similar approach as in the accompanying manuscript (Harde et al.). We inhibited VEGFR2 internalization with dynasore, a dynamin inhibitor that blocks both clathrin-dependent and clathrin-independent endocytic pathways, and analyzed the VEGF-mediated effect on axonal branching. Hippocampal neurons from wildtype mice were pretreated with dynasore or DMSO (as control) at 1 DIV for 30 min, followed by stimulation with 50 ng/ml VEGF. Time-lapse videos were recorded over the course of 4h and the dynamics of axon branching was analyzed. Treatment with dynasore blocked the increase in branch number and length induced by VEGF when compared to control conditions (DMSO treated neurons) (Figure 6—figure supplement 1A and B of revised manuscript).

These results indicate that internalization of VEGFR2 is required for the VEGF-mediated effect in axon branching. However, in contrast to what occurs for dendritic bragf-aching, the internalization of VEGFR2 required for regulating axon branching is not dependent of EphrinB2 as we did not observed any significant inhibition of VEGF-induced axon branching in primary hippocampal neurons isolated from EphrinB2 knockout mice (see response reviewer 1; comment 2; and Figure 6—figure supplement 1C-E of revised manuscript).

Are SPKs important for dendrites as well as axons? It is not clear whether axons and dendrites use different mechanisms or whether they were tested differently.

We assume that the reviewer meant to ask whether SFKs are important for dendrites as well as for axons. In the accompanying paper, Harde et al. show that VEGF stimulation of 14 DIV hippocampal neurons induces SFKs activation (Figure 3C, accompanying manuscript). As SFKs activity is required for EphrinB2 mediated spine morphogenesis (Segura et al., 2007), and as EphrinB2 is also required for VEGF/VEGFR2-mediated regulation of dendritogenesis and spine morphogenesis in hippocampal neurons, we assume that indeed SFK activity is also required in the context of VEGF/VEGFR2induced dendritogenesis and spinogenesis.

https://doi.org/10.7554/eLife.49818.sa2

Article and author information

Author details

  1. Robert Luck

    1. Biochemistry Center (BZH), University of Heidelberg, Heidelberg, Germany
    2. European Center for Angioscience, Medicine Faculty Mannheim, Heidelberg University, Heidelberg, Germany
    3. Institute for Transfusion Medicine and Immunology, Medicine Faculty Mannheim, Heidelberg University, Heidelberg, Germany
    Contribution
    Conceptualization, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing
    Contributed equally with
    Severino Urban
    Competing interests
    No competing interests declared
    Additional information
    These authors are listed alphabetically
  2. Severino Urban

    Biochemistry Center (BZH), University of Heidelberg, Heidelberg, Germany
    Contribution
    Conceptualization, Investigation, Methodology, Writing—original draft
    Contributed equally with
    Robert Luck
    Competing interests
    No competing interests declared
    Additional information
    These authors are listed alphabetically
  3. Andromachi Karakatsani

    1. Biochemistry Center (BZH), University of Heidelberg, Heidelberg, Germany
    2. European Center for Angioscience, Medicine Faculty Mannheim, Heidelberg University, Heidelberg, Germany
    3. Institute for Transfusion Medicine and Immunology, Medicine Faculty Mannheim, Heidelberg University, Heidelberg, Germany
    Contribution
    Investigation, Writing—review and editing
    Competing interests
    No competing interests declared
  4. Eva Harde

    1. Institute of Cell Biology and Neuroscience, University of Frankfurt, Frankfurt am Main, Germany
    2. Neurovascular Interface group, Max Planck Institute for Brain Research, Frankfurt am Main, Germany
    3. Buchmann Institute for Molecular Life Sciences (BMLS), University of Frankfurt, Frankfurt am Main, Germany
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  5. Sivakumar Sambandan

    Department of Synaptic Plasticity, Max Planck Institute for Brain Research, Frankfurt am Main, Germany
    Present address
    Max Planck Institute for Biophysical Chemistry, Goettingen, Germany
    Contribution
    Investigation, Visualization
    Competing interests
    No competing interests declared
  6. LaShae Nicholson

    1. Institute of Cell Biology and Neuroscience, University of Frankfurt, Frankfurt am Main, Germany
    2. Neurovascular Interface group, Max Planck Institute for Brain Research, Frankfurt am Main, Germany
    3. Buchmann Institute for Molecular Life Sciences (BMLS), University of Frankfurt, Frankfurt am Main, Germany
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  7. Silke Haverkamp

    Imaging Facility, Max Planck Institute for Brain Research, Frankfurt am Main, Germany
    Present address
    Department of Computational Neuroethology, Center of Advanced European Studies and Research (Caesar), Bonn, Germany
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  8. Rebecca Mann

    Biochemistry Center (BZH), University of Heidelberg, Heidelberg, Germany
    Contribution
    Investigation
    Competing interests
    No competing interests declared
  9. Ana Martin-Villalba

    Department of Molecular Neurobiology, German Cancer Research Center (DKFZ), Heidelberg, Germany
    Contribution
    Resources, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-9405-8910
  10. Erin Margaret Schuman

    Department of Synaptic Plasticity, Max Planck Institute for Brain Research, Frankfurt am Main, Germany
    Contribution
    Resources, Supervision
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7053-1005
  11. Amparo Acker-Palmer

    1. Institute of Cell Biology and Neuroscience, University of Frankfurt, Frankfurt am Main, Germany
    2. Neurovascular Interface group, Max Planck Institute for Brain Research, Frankfurt am Main, Germany
    3. Buchmann Institute for Molecular Life Sciences (BMLS), University of Frankfurt, Frankfurt am Main, Germany
    Contribution
    Resources, Funding acquisition, Writing—review and editing
    Competing interests
    No competing interests declared
  12. Carmen Ruiz de Almodóvar

    1. Biochemistry Center (BZH), University of Heidelberg, Heidelberg, Germany
    2. European Center for Angioscience, Medicine Faculty Mannheim, Heidelberg University, Heidelberg, Germany
    3. Institute for Transfusion Medicine and Immunology, Medicine Faculty Mannheim, Heidelberg University, Heidelberg, Germany
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    carmen.ruizdealmodovar@medma.uni-heidelberg.de
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5975-7815

Funding

Deutsche Forschungsgemeinschaft (FOR2325)

  • Robert Luck
  • Carmen Ruiz de Almodóvar
  • Amparo Acker-Palmer

Schram Foundation (Individual grant)

  • Carmen Ruiz de Almodóvar
  • Robert Luck
  • Andromachi Karakatsani

European Research Council (ERC-StG-311367)

  • Carmen Ruiz de Almodóvar

FP7 People: Marie-Curie Actions (FP7-PEOPLE-2011-CIG-304054)

  • Carmen Ruiz de Almodóvar

Deutsche Forschungsgemeinschaft (SFB1366)

  • Carmen Ruiz de Almodóvar

Deutsche Forschungsgemeinschaft (SFB873)

  • Carmen Ruiz de Almodóvar
  • Andromachi Karakatsani

Deutsche Forschungsgemeinschaft (SFB1158)

  • Carmen Ruiz de Almodóvar

Deutsche Forschungsgemeinschaft (GRK2099)

  • Carmen Ruiz de Almodóvar

European Research Council (ERC_AdG_Neurovessel_ project 669742)

  • Amparo Acker-Palmer

Deutsche Forschungsgemeinschaft (SFB834)

  • Amparo Acker-Palmer

Deutsche Forschungsgemeinschaft (SFB1080)

  • Amparo Acker-Palmer

Deutsche Forschungsgemeinschaft (SFB1193)

  • Amparo Acker-Palmer

Deutsche Forschungsgemeinschaft (EXC 2026)

  • Amparo Acker-Palmer

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank all members of CRA and AAP labs for their help, critical inputs and fruitful discussions. We also acknowledge the Nikon Imaging Center of Heidelberg University for all their help and support, and Gong Sun Nam for excellent technical assistance. Research in CRA is supported by the Marie Curie Career integration grant (FP7-PEOPLE-2011-CIG-304050), by the European Research Council (ERC-StG-311367 NeuroVascular Link), by Deutsche Forschungsgemeinschaft (FOR2325; SFB1366 (Project number 394046768-SFB 1366); SFB873; SFB1158 and GRK2099). and by the Schram Foundation. Research in AAP is supported by the European Research Council (ERC_AdG_Neurovessel_project 669742), by the Deutsche Forschungsgemeinschaft (SFB 834, SFB1080, SFB1193, FOR2325, EXC 2026) and the Max Planck Fellow Program.

Ethics

Animal experimentation: All animal experiments were conducted in accordance with the German institutional guidelines and ethical committees (references refs: FU1090, T60/12, T79/13, T49/15, T69/14, T46/16, T36/17, T48/18, and I19/13).

Senior Editor

  1. Didier Y Stainier, Max Planck Institute for Heart and Lung Research, Germany

Reviewing Editor

  1. Gou Young Koh, Institute of Basic Science and Korea Advanced Institute of Science and Technology (KAIST), Republic of Korea

Reviewers

  1. Injune Kim, Korea Advanced Institute of Science and Technology, Republic of Korea
  2. Bassem A Hassan, ICM, France

Publication history

  1. Received: July 1, 2019
  2. Accepted: November 14, 2019
  3. Version of Record published: December 23, 2019 (version 1)
  4. Version of Record updated: February 7, 2020 (version 2)

Copyright

© 2019, Luck et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 476
    Page views
  • 94
    Downloads
  • 1
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Computational and Systems Biology
    2. Developmental Biology
    Zhe Yang et al.
    Research Article Updated
    1. Developmental Biology
    2. Immunology and Inflammation
    Samuel Liegeois, Dominique Ferrandon
    Insight