Neuropilin-1 (NRP1), together with neuropilin-2, belongs to the neuropilin family. Neuropilins are transmembrane proteins essential for vascular and neural development and act as co-receptors for secreted signalling molecules of the class 3 semaphorin and vascular endothelial growth factor A (VEGF-A) families. NRP1 promotes VEGF-A signal in blood vascular endothelium and semaphorin signal in lymphatic endothelium, by forming complexes with its co-receptors. Mouse mutant studies established that NRP1 expression is essential during development because mice lacking NRP1 expression die embryonically and show severe neuronal and cardiovascular defects. Even though the contribution of NRP1 to vascular development has been mainly ascribed to its function as a VEGF-A receptor, recent evidence suggests that NRP1 contributes to angiogenesis through VEGF-independent mechanisms. In the present paper, we provide an overview of NRP1 functions in the vasculature and discuss current knowledge of NRP1-dependent signalling in the endothelium.
Neuropilin-1 (NRP1) plays a central role in endothelial cells (ECs). Loss of NRP1 in mouse embryos reduces angiogenesis in the brain and spinal cord and leads to embryonic lethality [1,2]. Binding of vascular endothelial growth factor A (VEGF-A) isoform to NRP1 promotes the formation of a trimeric complex with VEGF receptor 2 (VEGFR2) in ECs, stimulating signalling downstream of VEGF-A . Although NRP1 is required for VEGF-A-dependent arteriogenesis , new evidence shows that VEGF-A binding to NRP1 is dispensable for developmental angiogenesis in vivo because knockin mice expressing NRP1 with a mutation in the VEGF-binding pocket domain are viable without the severe and lethal cardiovascular phenotype observed in full or endothelium-specific NRP1-knockout mice . In accordance with the VEGF-A-independent function of NRP1, we have recently identified a novel mechanism by which NRP1 promotes developmental and pathological angiogenesis by mediating extracellular matrix (ECM) signalling via ABL1 in a VEGFR2-independent fashion . These observations imply that NRP1 does not exclusively function as a VEGF-A receptor as generally accepted, but it also controls VEGF-A-independent signalling to promote angiogenesis.
NRP1 structure and biological function
NRP1 was originally identified in the nervous system as an adhesion molecule , but it has since been studied mostly as a receptor of the semaphorin 3A in axon patterning and of VEGF-A in ECs . NRP1 is a single-pass transmembrane protein with a large N-terminal extracellular domain consisting of two complement-binding homology domains, named a1 and a2, essential for semaphorin binding; two coagulation factor V/VIII homology domains, named b1 and b2 involved in binding of VEGF165 isoform of VEGF-A [9,10] and placenta growth factor-2 (Plgf-2) , and a C domain that mediates NRP1 interactions with other receptors . NRP1 also presents a transmembrane domain and an intracellular C-terminus that contains a PSD-95/DIg/ZO-1 (PDZ)-binding motif that binds synectin, also known as Gα-interacting protein (GAIP)-interacting protein C-terminus 1 (GIPC1) .
NRP1-dependent VEGF-A signalling
During development, all vertebrates form a cardiovascular system to overcome limited oxygen diffusion in tissues. The first embryonic blood vessels arise from coalescing angioblasts, but thereafter blood vessels sprout from pre-existing vessels in a process termed angiogenesis. This is also the main mechanism of new vessel growth in adults, where it is usually referred to as neo-angiogenesis. Several different types of molecules stimulate the proliferation or migration of angiogenic ECs. Among these, VEGF-A is essential for all stages of cardiovascular development  and for pathological neo-angiogenesis in adults, which occurs, for example, during wound healing, diabetic retinopathy and solid tumour growth [13–15].
VEGF-A exists in three major forms produced by alternative splicing from a single VEGFA gene composed of eight exons and are termed VEGF121, VEGF165 and VEGF189 in humans or VEGF120, VEGF164 and VEGF188 in mice. The number refers to the total of amino acids in the mature protein . Exons 6 and 7 determine the ability of the different VEGF-A isoforms to bind heparin in vitro and are associated with the sequestration of VEGF-A to the ECM, thus determining the differential distribution in the environment of the different VEGF-A isoforms [16–18].
VEGF189 presents the domains encoded by exons 6 and 7, VEGF165 has the domain encoded by exon 7, and VEGF120 lacks both domains encoded by exons 6 and 7. All isoforms bind VEGFR1 and VEGFR2 tyrosine kinase receptors. Although VEGFR2 is the main signalling receptor in ECs that promotes VEGF signalling, VEGFR1 transduces VEGF-A signals in macrophages and serves as a VEGF decoy during angiogenesis . VEGF165 is the main VEGF-A isoform to bind NRP1, although VEGF189 was also recently shown to bind NRP1 . The formation of a trimeric complex between NRP1, VEGF and VEGFR2 is mediated by the cysteine knot motif located in exon 4 of VEGF165, which contacts VEGFR2, and the domain presents in exon 7/8-encoded region that interacts with the b1 domain of NRP1 (Figure 1). The cytoplasmic NRP1 tail is also essential for complex formation. The VEGF120 isoform lacks exons 6 and 7 and has a lower affinity for NRP1 than does VEGF165 .
NRP1 is composed of three extracellular domains, a transmembrane domain and an intracellular C-terminal domain containing a PDZ-binding domain
The importance of the NRP1–VEGFR2–VEGF165 complex in promoting VEGF signalling has been established mainly by overexpression studies in ECs, which showed that co-expressing NRP1 and VEGFR2 in porcine aortic endothelial (PAE) cells, which lack endogenous expression of either protein, increases the VEGF-induced phosphorylation of extracellular-signal-regulated kinase (ERK) and p38 mitogen-activated protein kinase (MAPK) more than expressing VEGFR2 alone, although expressing only NRP1 did not promote these signals . Consistent with a role for NRP1 in promoting optimal VEGF-A signalling through VEGFR2, inhibiting VEGF165 binding to NRP1, using a specific peptide that does not affect VEGF165 binding to VEGFR2 in PAE cells co-expressing NRP1 and VEGFR2 proteins, reduced VEGF-A-induced VEGFR2 tyrosine phosphorylation and the activation of ERK. However, inhibition of VEGF165 binding to NRP1 had little effect on the proliferation/pro-survival signalling promoted by Akt phosphorylation. Accordingly, treatment of human umbilical vein endothelial cells (HUVECs) with anti-VEGF antibody completely blocked VEGF-A-induced VEGFR2 phosphorylation, but treatment with anti-NRP1 blocking antibodies modestly reduced VEGFR2, ERK and Akt phosphorylation levels . More importantly, treatment with anti-NRP1 antibodies has an additive effect when combined with anti-VEGF agents in reducing tumour growth and tumour vascular density in tumour xenograft models. This observation suggests that, although NRP1 enhances VEGF-A signalling by interacting with VEGFR2, it also has VEGF-independent roles and that NRP1 and VEGF-A control different but synergistic pathways in the vasculature .
VEGF-independent NRP1 signalling
The role of NRP1 as a VEGF-A receptor has been mainly investigated by biochemical studies in vitro. A recent study examined the relevance of VEGF-A binding to NRP1 in vivo by using knockin mice that express NRP1 with a point mutation of Tyr297 located in the VEGF-binding pocket , previously shown to disrupt VEGF165 binding to NRP1 in vitro . Although the Nrp1Y297 allele caused a severe decrease in NRP1 expression, these mice were viable and showed a mild vascular phenotype in hindbrain vasculature, with a similar number of radial vessels compared with wild-type littermates, but reduced vessel branching in the hindbrain subventricular plexus. Even though the severe reduction in NRP1 expression did not permit definitive discrimination of VEGF binding from other NRP1 functions, these data demonstrate that VEGF binding to NRP1 is not essential for embryonic angiogenesis. Interestingly, adult Nrp1Y297/Y297 mice suffered reduced body size and increased postnatal mortality due to congestive heart failure. Hence Nrp1Y297/Y297 mice present a mild embryonic vascular phenotype and extended viability compared with full or endothelial-specific NRP1 knockout. Therefore NRP1 functions in embryonic ECs predominantly in a VEGF-independent manner.
NRP1 modulates signalling of receptors other than VEGFRs
It has recently been reported that NRP1 can interact with or modulate signalling of several receptors other than VEGFRs. NRP1 regulates phosphorylation of the actin cytoskeleton-associated protein p130Cas in response to hepatocyte growth factor (HGF) in the U87 malignant glioma cell line or upon stimulation of vascular smooth muscle cells with platelet-derived growth factor (PDGF) [25–27]. In both cases, p130Cas phosphorylation requires the NRP1 cytoplasmic tail [27,28]. Also, it has been reported that NRP1 can modulate the signalling of transforming growth factor β1 (TGFβ1) by binding the latent and active forms of TGFβ1 and by simultaneously binding TGFβ receptors 1 and 2 [29–31].
NRP1 functions as an adhesion molecule
NRP1 was originally identified as an adhesion molecule in the nervous system , and further studies identified two sequences of 18 amino acid residues located within the NRP1 b1 and b2 domains that promote cell–cell adhesion by interacting with heterologous proteins on adjacent cells yet to be identified . It has been reported that NRP1 interacts with integrins in vitro [33,34], and promotes α5β1 integrin-mediated adhesion of HUVECs to ECM fibronectin in vitro, in a mechanism that depends on the NRP1 cytoplasmic tail, which increases endocytosis of active α5β1 . However, genetic mouse studies showed that the cytoplasmic tail of NRP1 is dispensable for developmental and pathological angiogenesis [4,36] and therefore signal transduction downstream of the NRP1 cytoplasmic domain is unlikely to play major roles in angiogenesis. Accordingly, NRP1 contributes to angiogenesis exclusively via its extracellular domain . In agreement with a role of NRP1 in promoting both cell–cell and cell–matrix adhesion, NRP1 promotes cell attachment of HUVECs to fibronectin, laminin and gelatin in vitro, independently of VEGFR2, and it affects filamentous actin (F-actin) polymerization [36,37]. Yet, the relative significance of NRP1 for VEGF-A/VEGFR2 signalling compared with integrin ligand-stimulated processes for angiogenesis is not completely understood.
NRP1 regulates ABL1 signalling in angiogenic ECs
It has recently been reported that NRP1 drives angiogenesis independently of VEGFR2 by promoting ECM signalling and actin remodelling  (Figure 2). By combining inhibitor studies with genetic studies in mice, we have identified a VEGFR2-independent pathway involving NRP1-dependent activation of the non-receptor tyrosine kinase ABL1. This, in turn, promotes ABL1-mediated phosphorylation of the focal adhesion-related protein paxillin (PXN) downstream of integrin activation in ECs. The formation of a protein complex between NRP1 and ABL1 is required to recruit PXN to the complex and to stimulate PXN phosphorylation by ABL1 in vitro. NRP1 and ABL1 promote PXN phosphorylation also in vivo in the perinatal mouse retina, where endothelial sprouts headed by filopodia-studded tip cells migrate on a template of astrocyte-derived fibronectin towards the retinal periphery following a VEGF gradient [18,38,39]. Accordingly, inducible EC-specific Nrp1-knockout mouse mutants and mice treated with the ABL1 inhibitor imatinib, a U.S. Food and Drug Administration (FDA)-approved drug used to treat acute myeloid leukaemia, showed a similar phenotype in the perinatal angiogenic retina, with fewer tip cells, impaired lateral sprout extension and reduced vascular branch points . Importantly, imatinib-treated mice showed a small decrease in the vascular extension across the retina, due to the inhibition of the signalling promoted by astrocytic fibronectin at the vascular front , whereas the Nrp1 mutant presented a more severe reduction in agreement with the consensus role of NRP1 in promoting VEGF signalling. Thus NRP1 plays a dual role in angiogenesis and vascular morphogenesis by independently promoting ECM-stimulated and growth-factor-induced signals in ECs (Figure 2).
Schematic representation of the NRP1/ABL1/PXN pathway and its synergism with known VEGF signalling pathways transduced by VEGFR2
NRP1–ABL1 pathway is a potential therapeutic target for vascular eye diseases
Age-related macular degeneration (AMD) and diabetic retinopathy are the leading causes of blindness in people over 50 years of age. AMD exists in two different clinical forms called dry and wet AMD. Although the dry form is caused by accumulation of cellular debris called drusen that may lead to detachment of the retina, the wet form of AMD is caused by abnormal growth of the choroid blood vessel behind the retina . In proliferative diabetic retinopathy (PDR), abnormal retinal angiogenesis occurs to compensate for the loss of functionality of damaged existing blood vessels in the retina. In both diseases, the newly formed vessels leak blood and fluid causing oedema, resulting in impaired vision and ultimately blindness . Anti-VEGF therapy (i.e. using Lucentis®, Macugen® or Avastin®) is the approved treatment for AMD and PDR. These drugs are administered by monthly injection into the eye. In the case of wet AMD, anti-VEGF therapy stabilizes the sight in over 90% of cases, but only 30% of people show improved vision . Also, recent evidence suggests that anti-VEGF therapy is not curative because oedema returns as soon as the treatment is discontinued . Furthermore, a 7-year multicentre cohort clinical study showed that after 7 years of treatment with anti-VEGF therapies, only one-third of patients showed good visual outcome, whereas another one-third had poor outcome . Because long-term anti-VEGF monotherapy has limited efficacy, there is the need to develop alternative treatments. We have recently reported that inhibition of the NRP1/ABL1 pathway by treatment with imatinib reduces neovascularization in a mouse model of oxygen-induced retinopathy (OIR) . In this model, exposing pups to hyperoxia, followed by exposure to normoxia, induces first vaso-obliteration of central retinal capillaries due to the high concentration of oxygen, and leads to the formation of neovascular lesions in normoxic conditions, similar to those observed in patients with PDR . Specifically, in the OIR model, inducible EC-specific Nrp1-knockout mouse mutant and mice treated with imatinib showed a similar reduction in revascularization of vaso-obliterated areas and reduced formation of neovascular lesions . Interestingly, depletion of NRP1 in ECs caused a slightly stronger effect compared with treatment with imatinib, in accordance with a dual role for NRP1 in promoting ECM-stimulated and growth factor-induced signals in EC during angiogenesis (Figure 2). Thus targeting NRP1-mediated ABL1 signalling prevents pathological angiogenesis in mice and may represent a novel therapeutic target to enhance the efficacy of current anti-angiogenic therapies to treat vascular eye diseases.
NRP1 is a multifunctional transmembrane receptor that signals downstream of several ligands such as growth factors and ECM components. Even though it is known that NRP1 has several functions, the consensus model depicts NRP1 mainly as a receptor for semaphorin and VEGFR. Nevertheless, new evidence suggests that the molecular mechanism by which NRP1 controls EC behaviour also involves VEGF-independent signalling pathways. A better understanding of the role of NRP1 in physiological and pathological blood vessel growth represents an important progress to develop potential treatments to cure diseases that rely on angiogenesis or are caused by abnormal blood vessel growth. Therefore NRP1 can serve as a molecular target for therapeutic intervention to prevent pathological angiogenesis in cancer or vascular eye diseases or to promote new vessel growth in ischaemic tissues.
Membrane Morphology and Function: A Biochemical Society Focused Meeting held at Hotel del Camerlengo, Fara San Martino, Abruzzo, Italy, 5–8 May 2014. Organized and Edited by Banafshé Larijani [IKERBASQUE, Basque Foundation for Science and Unidad de Biofísica (CSIC-UPV/EHU), University of the Basque Country, Spain] and Marco Falasca (Barts and The London School of Medicine and Dentistry, U.K.)
age-related macular degeneration
human umbilical vein endothelial cell
porcine aortic endothelial
proliferative diabetic retinopathy
transforming growth factor β1
vascular endothelial growth factor
vascular endothelial growth factor receptor
I am grateful to C. Ruhrberg, A. Fantin and A. Lampropoulou for helpful discussions, and to A. Chikh, T. Maffucci and M. Falasca for a critical reading of the article.
I am funded by a research fellowship from the British Heart Foundation [grant number FS/13/35/30148].