Endocytosis regulates VEGF signalling during angiogenesis

Endocytosis has proved to be a versatile mechanism regulating diverse cellular processes, ranging from nutrient uptake to intracellular signal transduction. New work reinforces the importance of endocytosis for VEGF receptor signalling and angiogenesis in the developing eye, and describes a mechanism for its differential regulation in angiogenic versus quiescent endothelial cells.

Members of the vascular endothelial growth factor (VEGF) family, primarily VEGF-A and VEGF-C, and their receptors, VEGFR-1–3, are key regulators of normal and pathological angiogenesis. VEGF signalling is tightly regulated in multiple ways, including VEGF receptor endocytosis. In this issue, Nakayama et al.1 shed light on how VEGFR signalling is regulated through spatial control of VEGFR endocytosis, mediated by the activities of cell polarity factors (principally aPKC).

The activity of many ligand-activated transmembrane receptors depends on their internalization and subsequent intracellular trafficking. The amplitude, specificity and duration of signalling are, for example, dependent on whether a receptor is targeted for degradation or recycling to the plasma membrane. Furthermore, many receptors signal from endosomal compartments, and specific signalling events were suggested to occur at distinct types of endosomes (reviewed in ref. 2). Endosomes may also serve as scaffolds that facilitate the assembly of signalling complexes and their transport to relevant subcellular locations2. Several recent publications have demonstrated the importance of VEGFR-2 endocytosis for its biological function and signalling capability3. Following VEGF-A binding, VEGFR-2 is rapidly internalized in a clathrin- and dynamin-dependent manner4, ultimately leading to its proteolytic degradation5,6. VEGF also triggers trafficking of intracellular VEGFR-2 from the Golgi apparatus to the plasma membrane, which requires syntaxin 6, a Golgi-localized t-SNARE (target SNAP receptor)7.

VEGFR-2 endocytic trafficking has also been shown to require synectin (a protein involved in intracellular trafficking) and myosin VI (a retrograde actin-based motor linked to endosomal transport) in the context of arterial morphogenesis8. Arterial endothelial cells lacking either of these proteins had reduced VEGFR-2 phosphorylation (on Tyr 1175) and compromised downstream signalling through the MAP kinase ERK1/2 and phospholipase Cγ1. Interestingly, VEGF-A-mediated VEGFR-2 internalization occurred normally in cells lacking synectin or myosin VI, but VEGFR-2 entry into the early endosome antigen 1 (EEA1)-positive endosomes was delayed. As a consequence, VEGFR-2 remained in the vicinity of the plasma membrane, where its dephosphorylation at Tyr 1175 by the phosphatase PTP1b resulted in reduced downstream signalling8. Ephrin-B2, one of the transmembrane ligands for the EphB receptor tyrosine kinases, also promotes VEGFR-2 and VEGFR-3 endocytosis, and its absence compromises downstream signalling through RAC1, AKT and ERK1/2. Ablation of ephrin-B2 in mice causes defects in angiogenesis and lymphangiogenesis, yet how ephrin-B2 regulates VEGFR endocytosis is mechanistically unclear9,10.

Although a role for endocytosis in VEGFR signalling has already been described, Nakayama et al. provide important insight into its spatial control1. They make extensive use of high-resolution in vivo imaging of the neonatal mouse retina, including intracellular tracking of ligand-induced receptor endocytosis to reach a resolution normally only achieved in cultured cells. They verify the proposed functions of the genes under study using tissue-specific gene ablation in the relevant tissue, and achieve double-mutant phenotypic rescue when proteins with opposing functions were ablated.

The neonatal mouse retina has become a popular model for developmental angiogenesis because it lends itself to high-resolution imaging and in vivo manipulation. Retinal vascularization starts at the centre and expands radially by centrifugal sprouting. At the vascular front, endothelial cells show high migratory and mitogenic activity, and they become progressively stabilized at central locations. Since retinal angiogenic sprouting is VEGF-A- and VEGFR-2-dependent11, it would be expected that VEGFR-2 levels are high at the angiogenic sprouting front of the retina. However, Nakayama et al.1 found that immunostaining for VEGFR-2 was lower in vessels at the sprouting front than in the more quiescent proximal vessels. They hypothesized that this reflected differences in turnover rate of VEGFR-2, which was supported by the ability of proteasome, lysosome or endocytosis inhibitors to increase VEGFR-2 (and VEGFR-3) protein levels at the angiogenic front. To visualize the spatial pattern of ligand-induced receptor trafficking, the authors carried out intra-ocular injections of fluorescently labelled VEGF-A and found that it was internalized in endothelial cells, where it often co-localized with EE1A (indicating that VEGF-A localizes to the early endosome). Impairment of endocytosis through inhibition of dynamin (which acts in clathrin-mediated endocytosis) or inhibition of the clathrin-associated sorting protein Disabled 2 (Dab2) reduced the endothelial uptake of labelled VEGF, and endothelial-specific deletion of VEGFR-2 (Flk-1iΔEC) had similar effects. These results show that fluorescently labelled VEGF bound to VEGFR-2 is internalized through clathrin-mediated endocytosis.

They also establish that VEGFR-2 endocytosis occurs preferentially at the angiogenic front in the neonatal retina. The authors further demonstrate that VEGFR endocytosis is important for normal retinal angiogenesis and that it requires the cell polarity factor PAR-3. Dab2 and PAR-3 were shown to bind each other as well as to VEGFR-2 and VEGFR-3, suggesting that they form an endocytic trafficking complex. Endothelial-specific ablation of Dab2 or Par3 reduced VEGF uptake and increased VEGFR-2 and -3 immunoreactivity at the angiogenic front. The authors also showed that Dab2 binds to ephrin-B2, and that the angiogenic phenotypes associated with loss of Dab2 or PAR-3 (reduced sprouting and proliferation) resemble those of ephrin-B2 mutants. Providing further corroborating biochemical data on retinal VEGFR signalling is challenging owing to the small proportion of endothelial cells in the minute retina, and also because of possible interference from VEGFR-2-mediated signalling events in retinal neurons. To avoid this pitfall, the authors moved to using cultured cells and showed that knockdown of Dab2 or PAR-3 led to decreased uptake of VEGF and reduced ERK1/2 and RAC1 activation, whereas AKT seemed unaffected.

Mechanistically, the authors identified atypical protein kinase λ (aPKC-λ) and aPKC-ζ as negative regulators of VEGFR endocytosis. The authors demonstrate that these aPKCs phosphorylate the PTB domain in Dab2 to reduce its interaction with VEGFR-2 and VEGFR-3. Conversely, endothelial-specific deletion of aPKC-λ in mice resulted in increased Dab2–VEGFR-3 complex formation, and inhibition of aPKC in cultured cells led to accelerated VEGFR-2 and VEGFR-3 internalization. In line with these observations, aPKC inhibition enhanced the VEGF-A- and VEGF-C-induced ERK1/2 activation. Together, these findings suggest a role for aPKC-mediated Dab2 phosphorylation in negatively regulating VEGF receptor endocytosis and signalling (Fig. 1). In the neonatal retina, immunostaining for activated aPKC was weak at the sprouting front, but stronger in the quiescent proximal vasculature, implying that a graded aPKC activity may cause the differences in VEGFR-2 endocytosis between angiogenic and quiescent endothelial cells. In support of this interpretation, endothelial-specific ablation of aPKC-λ increased the accumulation of labelled VEGF and reduced VEGFR-2 and VEGFR-3 immunostaining at central locations, where the vessels also acquired characteristics of a sprouting phenotype. Strikingly, and consistent with the proposed mechanism, this phenotype was corrected by simultaneously deleting aPKC-λ together with Dab2 or PAR-3.

Figure 1: Schematic representation of VEGFR endocytosis and signalling.

(a) VEGF-A (pink) binding to VEGFR-2 at the plasma membrane leads to receptor phosphorylation and activation of downstream effectors ERK1/2, RAC1 and AKT. The presence of protein tyrosine phosphatase 1b (PTP1b) in proximity to VEGFR-2 triggers its dephosphorylation and inhibits signalling. (b) A protein complex including ephrin-B2, Dab2 and PAR-3 associates with VEGFR-2 in the clathrin-coated vesicle and mediates trafficking towards the early endosome. Atypical PKC (aPKC) is a negative regulator of this process, as it mediates phosphorylation of Dab2, thereby reducing VEGFR-2 endocytosis. (c) VEGFR-2 in the early endosome signals mainly through ERK1/2 and RAC1. The VEGF-A–VEGFR-2 complex can then be recycled to the plasma membrane via the recycling endosome or undergo lysosomal degradation. Rab proteins implicated in the different trafficking steps are indicated.

In summary, these results identify aPKC-λ as an inhibitor of Dab2/PAR-3-dependent VEGFR endocytosis and signalling in retinal angiogenesis. It remains to be elucidated how aPKC adopts its differential activity along the retinal vasculature, and how VEGFR signalling from the endosomal compartment is qualitatively and quantitatively different from its signalling at other subcellular locations. Although biochemical evidence suggests that VEGFR-2 endocytosis is required for full ERK1/2 and RAC1 activation, it would be valuable to see the many possible signalling pathways downstream of VEGFR-2 being investigated by endothelial-specific gene knockout in the retina, using the same rigorous standards as Nakayama et al. This would help us understand which pathways are physiologically relevant for angiogenesis in vivo, and whether these coincide with the pathways delineated in in vitro studies. RAC1 has been implicated in cell migration downstream of VEGFRs in vitro, and embryonic endothelial-specific deletion of RAC1 leads to early lethality associated with vascular defects12. However, the analysis done so far has not revealed whether defective cell migration, altered cell death or even other mechanisms are causing this phenotype. Surprisingly, RAC1 deletion in the neonatal retina failed to provoke an obvious angiogenic phenotype13.

What should we expect from future research in angiogenic signalling? Although genetic toolbox and imaging techniques develop quickly, high-resolution live imaging of angiogenesis in mammals is still technically challenging and seldom reaches a satisfactory resolution. This is in marked contrast to zebrafish, where this approach has pushed the field significantly forward. Currently, we study the vast majority of developmental and pathological processes through 'snapshots'. However, as most biological processes during development and disease are dynamic, we would certainly benefit from live imaging at good resolution. Efforts to develop fluorescent markers and mouse lines to address this issue will definitely pay off.


  1. 1

    Nakayama. et al. Nat. Cell Biol. 15, 249–260 (2013).

    CAS  Article  Google Scholar 

  2. 2

    Sadowski, L., Pilecka, I. & Miaczynska, M. Exp. Cell Res. 315, 1601–1609 (2009).

    CAS  Article  Google Scholar 

  3. 3

    Simons, M. Physiology 27, 213–222 (2012).

    CAS  Article  Google Scholar 

  4. 4

    Lampugnani, M. G., Orsenigo, F., Gagliani, M. C., Tacchetti, C. & Dejana, E. J. Cell Biol. 174, 593–604 (2006).

    CAS  Article  Google Scholar 

  5. 5

    Ewan, L. C. et al. Traffic 7, 1270–1282 (2006).

    CAS  Article  Google Scholar 

  6. 6

    Bruns, A. F. et al. Traffic 11, 161–174 (2010).

    CAS  Article  Google Scholar 

  7. 7

    Manickam, V. et al. Blood 117, 1425–1435 (2011).

    CAS  Article  Google Scholar 

  8. 8

    Lanahan, A. A. et al. Dev. Cell 18, 713–724 (2010).

    CAS  Article  Google Scholar 

  9. 9

    Wang, Y. et al. Nature 465, 483–486 (2010).

    CAS  Article  Google Scholar 

  10. 10

    Sawamiphak, S. et al. Nature 465, 487–491 (2010).

    CAS  Article  Google Scholar 

  11. 11

    Gerhardt, H. et al. J. Cell Biol. 161, 1163–1177 (2003).

    CAS  Article  Google Scholar 

  12. 12

    Tan, W. et al. FASEB J. 22, 1829–1838 (2008).

    CAS  Article  Google Scholar 

  13. 13

    D'Amico, G. et al. PLoS ONE 5, e9766 (2010).

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Christer Betsholtz.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Gaengel, K., Betsholtz, C. Endocytosis regulates VEGF signalling during angiogenesis. Nat Cell Biol 15, 233–235 (2013). https://doi.org/10.1038/ncb2705

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing