Placental growth factor and its receptor, vascular endothelial growth factor receptor-1: novel targets for stimulation of ischemic tissue revascularization and inhibition of angiogenic and inflammatory disorders
The Center for Transgene Technology and Gene Therapy, Flanders Interuniversity Institute for Biotechnology; University of Leuven, B-3000 Leuven, Belgium
P. Carmeliet MD PhD, Center for Transgene Technology and Gene Therapy, Flanders Interuniversity Institute for Biotechnology, KU Leuven, Campus Gasthuisberg, Herestraat 49, B-3000, Leuven, Belgium. Tel.: +32 1634 5772; fax: +32 1634 5990; e-mail: email@example.com
Summary. In contrast to VEGF and its receptor VEGFR-2, PlGF and its receptor VEGFR-1 have been largely neglected and therefore their potential for therapy has not been previously explored. In this review, we describe the molecular properties of PlGF and VEGFR-1 and how this translates into an important role for PlGF in the angiogenic switch in pathological angiogenesis, by interacting with VEGFR-1 and synergizing with VEGF. PlGF was effective in the growth of new and stable vessels in cardiac and limb ischemia, through its action on different cell types (i.e. endothelial, smooth muscle and inflammatory cells and their precursors) that play a cardinal role in blood vessel formation. Accordingly, blocking its receptor VEGFR-1 with monoclonal antibodies (anti-VEGFR-1 mAb), expressed on al these cell types, successfully attenuated blood vessel formation during cancer, ischemic retinopathy and rheumatoid arthritis. In addition, while blocking this receptor was effective in reducing inflammatory disorders like atherosclerosis and rheumatoid arthritis, blocking the anti-angiogenic receptor VEGFR-2 was without effect. This indicates that in the latter diseases the beneficial effects of anti-VEGFR1 mAb were mainly due to its effect on inflammatory cells. Importantly, VEGFR-1 was also present on hematopoietic stem/progenitor cells, the precursors of inflammatory cells. Thus, these preclinical studies show proof-of-principle that PlGF and VEGFR-1 are promising therapeutic targets to treat angiogenesis and inflammation related disorders. Clinical trials will reveal whether this is also true for patients.
Discovering the molecules determining the angiogenic switch is of great value, not only to better understand how angiogenesis (i.e. the birth of new blood vessels) contributes to more than 70 diseases but also ultimately to develop more efficient and safer medicaments for these angiogenic disorders. After a decade of intense research efforts, there is now general consensus that vascular endothelial growth factor (VEGF) is a major player of angiogenesis (reviewed in ). VEGF, also referred to as VEGF-A, regulates angiogenesis by interacting with mainly two tyrosine kinase receptors, VEGF receptor-1 (VEGFR-1, also known as Fms-like tyrosine kinase, Flt1) and VEGF receptor-2 (VEGFR-2, also known as fetal liver kinase, Flk1, and, in humans, as kinase insert domain-containing receptor, KDR) . While VEGFR-2 has been generally considered to be the main transducer of the VEGF-A-dependent angiogenic signals (reviewed in [3,4]), the role of VEGFR-1 has remained largely elusive. Since the discovery of VEGF-A, several other homologs have been identified, including VEGF-B, VEGF-C, VEGF-D, VEGF-E (also called orf virus VEGF) and placental growth factor (PlGF) (reviewed in ). However, much less is known about the specific functions of all these factors. The purpose of this review is to focus on the recently unveiled role and therapeutic potential of two largely neglected molecules of the VEGF family, PlGF and VEGFR-1.
PlGF: molecular properties
PlGF was originally discovered in human placenta by Persico in 1991 , only 2 years after the discovery of VEGF-A . Although primarily expressed in the placenta, PlGF transcripts have also been detected in the heart, lung, thyroid gland and skeletal muscle . Many cell types produce PlGF: quiescent endothelial cells release minimal amounts of PlGF but, when activated, angiogenic endothelial cells produce abundant amounts of PlGF and thereby regulate the VEGF-A-dependent angiogenic switch (see below). In addition, other cell types, including vascular smooth muscle cells, inflammatory cells, bone marrow cells, neurons and many tumor cells also produce PlGF, especially when activated or stressed [9–11]. Alternative splicing of human PlGF primary transcript generates three isoforms, which differ in size and binding properties [12,13]: (i) PlGF-1 (PlGF131), (ii) PlGF-2 (PlGF152), which, due to the insertion of 21 basic amino acids at the carboxy-terminus, is able to bind polyanionic substances (e.g. heparin) and the coreceptors neuropilin-1 and neuropilin-2 [14–16], and (iii) the longest isoform PlGF-3 (PlGF203), which, like PlGF-1, lacks the amino acid insert that confers the ability to bind heparin. PlGF-2 is the only isoform present in the mouse . Like VEGF-A and the other VEGF homologs, PlGF belongs to the so-called cystein-knot superfamily of growth factors, whose members are all characterized by a common motif of eight spatially conserved cysteines, which are involved in intra- and inter-molecular disulfide bonds, the latter allowing the formation of dimers . A recent study on the three-dimensional structure of human PlGF-1 has shown that, overall, this protein is strikingly similar to VEGF-A, although the two proteins only share 42% amino acid sequence identity . The two monomers of PlGF are held together by one interchain disulfide bond, and the dimeric structure is also stabilized by a hydrophobic core region (one for each monomer). The side-by-side, head-to-tail orientation of the monomers places the receptor-binding interface at each pole of the dimer (Fig. 1). It is interesting to note that half of the residues located on the receptor-binding interface are different between PlGF and VEGF-A, an important finding which is the explanation of the different functional properties of the two proteins.
VEGFR-1: molecular properties
PlGF specifically binds VEGFR-1, but not VEGFR-2 , and induces VEGFR-1 auto-phosphorylation . However, compared with the strong tyrosine phosphorylation of VEGFR-2, VEGFR-1 is only weakly phosphorylated in cultured endothelial cells [22,23]. This has cast doubts on whether VEGFR-1 transmits any significant angiogenic signals at all. Some in vitro studies have shown that PlGF stimulates endothelial cell growth and migration [8,24], while others have documented minimal effects on proliferation, migration and permeability of endothelial cells [20,25,26]. When comparing wild-type and PlGF-deficient endothelial cells, we observed, however, that the abundant endogenous production of PlGF by endothelial cells in culture reduces their sensitivity to PlGF, presumably because endogenous PlGF saturates VEGFR-1 and renders the cells insensitive to additional exogenous PlGF . In a sense, cultured endothelial cells are therefore not a good model to study the effects of PlGF. Since PlGF is able to enhance the angiogenic activity of VEGF-A in vitro, it has been proposed that PlGF competes with VEGF-A for binding to VEGFR-1, so that more VEGF-A is available to activate VEGFR-2. PlGF can also form heterodimers with VEGF-A both in vitro and in vivo[28,29], which induce angiogenesis almost to the same extent as VEGF-A/VEGF-A homodimers [29,30]. Based on their receptor binding selectivity, they might transmit angiogenic signals via binding to VEGFR-1 homodimers or VEGFR-1/VEGFR-2 heterodimers (see below).
VEGFR-1, the first VEGF receptor to be identified , was originally found to be expressed only on vascular endothelial cells . Recently, several reports have shown the expression of this receptor on non-endothelial cells, including smooth muscle cells , monocytes , trophoblasts , mesangial cells , and osteoblasts . Importantly, expression of VEGFR-1 is highly upregulated in disease [11,38,39], thereby contributing to the angiogenic switch (see below). Both mutagenesis and crystallographic studies have demonstrated that the second immunoglobulin domain of VEGFR-1 is the major binding site for VEGF-A and PlGF [22,40–43]. Interestingly, not all the amino acid residues located on the ligand interface of VEGFR-1 are identical for PlGF and VEGF-A [19,40]. The affinity of VEGFR-1 for VEGF-A (Kd, 1–20 pmol L−1) is higher than that for PlGF (Kd, about 170 pmol L−1) [21,25,44,45].
VEGFR-1, like other tyrosine kinase receptors including VEGFR-2, mediates signal transduction following well-defined molecular phases: (i) receptor dimerization upon ligand stimulation; (ii) activation of the tyrosine kinase with consequent receptor autophosphorylation; and (iii) binding and activation of adaptors to the autophosphorylation sites . However, since tyrosine phosphorylation of VEGFR-1 is much weaker than that of VEGFR-2, it has remained unclear whether and to what extent VEGFR-1 is indeed involved in controlling angiogenesis. In vitro, VEGFR-1 has been implicated in endothelial cell migration and growth in some but not in other studies [15,20,21,25,47–50]. It should be noticed, however, that expression of VEGFR-1 in cultured endothelial cells is rapidly downregulated in vitro. Upon VEGF-A stimulation, VEGFR-1 has been reported either to stimulate, to be ineffective against, or even to suppress VEGFR-2-dependent activity in vitro[47,51–56]. The inhibitory effect may be attributable, in part, to a short stretch of residues in the juxtamembrane cytosolic domain of VEGFR-1, which mediates inhibitory signals . Furthermore, gene-targeting studies in mice have suggested that VEGFR-1-dependent signaling has a minimal role in vascular development. Mice deficient in VEGFR-1 succumbed around embryonic day 8.5 due to vascular overgrowth and disorganization , but mice lacking the tyrosine kinase domain were viable . Based on these findings, it has been proposed that VEGFR-1 mainly functions as an inert reservoir (sink) for VEGF-A and prevents VEGF-A from binding its signaling receptor VEGFR-2 [1,58].
Soluble VEGFR-1: a natural VEGF inhibitor in disease
The VEGFR-1 primary transcript gives rise to two polypeptides by alternative splicing: (i) the full-length receptor, consisting of seven extracellular immunoglobulin-like domains, a transmembrane region, and an intracellular consensus tyrosine kinase domain interrupted by a kinase-insert domain; and (ii) a soluble form of VEGFR-1, which lacks the seventh immunoglobulin-like domain, the transmembrane and the cytoplasmic regions, but retains the ability to bind its ligands  (Fig. 1). It is remarkable that very few naturally occurring mechanisms have thus far been identified to inhibit VEGF-A or its receptors directly. By capturing VEGF-A, soluble VEGFR-1 functions as an inhibitor of VEGF-A-dependent angiogenic signaling and has even been successfully used to block VEGF-A-dependent angiogenic disorders in preclinical animal models [60,61], but its endogenous role in health and disease remains largely enigmatic. Plasma levels of soluble VEGFR-1 are elevated in individuals with cancer, ischemia and pre-eclampsia [62–64], but it remains unknown whether it causally contributes to the pathogenesis of the disease or is simply a surrogate marker of the disease progress. A recent study indicated, however, that soluble VEGFR-1 may play a more important role in pre-eclampsia than anticipated. Indeed, in pregnant women with pre-eclampsia, maternal spiral artery remodeling in the uterus is defective, leading to hypoxia in the placenta. The hypoxic placenta releases elevated levels of soluble VEGFR-1 in the circulation, which capture free VEGF-A and PlGF and thereby deprive the endothelium of established vessels from critical survival and maintenance signals. As a result, edema develops in the brain and liver, the blood pressure rises (VEGF-A has hypotensive activity), the glomerular capillaries disintegrate, and widespread thrombosis occurs; all signs of pre-eclampsia .
PlGF: a key molecule in the angiogenic switch
A gene targeting study in mice has provided clues for understanding the functional role of PlGF in angiogenesis . PlGF, while affecting vascular development, is not essential for physiological angiogenesis during development and reproduction. For instance, there are almost 50% fewer vessels in the ovarian corpus luteum in PlGF-deficient mice, but this defect does not cause any functional insufficiency and reproduction is grossly normal. Thus, while we cannot exclude minor defects in vascular development, loss of PlGF consistently impairs angiogenesis in pathological conditions, including ischemia, inflammation and cancer (Fig. 2). PlGF contributes to this angiogenic switch via several mechanisms: (i) by affecting endothelial cells, in particular by amplifying the angiogenic activity of VEGF-A (Fig. 3); (ii) by affecting smooth muscle cells, and thereby stimulating vessel maturation and stabilization (Fig. 4); (iii) by recruiting inflammatory cells, which play a critical role in growth of collateral vessels (Fig. 5); and (iv) by mobilizing vascular and hematopoietic stem cells/progenitors from the bone marrow. These genetic insights have prompted us to evaluate the therapeutic potential of PlGF which, indeed, has been shown to be effective in relieving ischemia and improve function of the ischemic heart and limbs  (Fig. 2). Conversely, blocking VEGFR-1 reduced angiogenesis and/or inflammation during angiogenic and inflammatory disorders  (Fig. 2). These findings, together with the evidence that deletion of the tyrosine kinase domain of VEGFR-1 or inhibition of VEGFR-1 by antagonists impairs pathological angiogenesis [65,66] and that PlGF stimulates angiogenesis in a variety of conditions in vivo[11,24,27,67–69] (see below), indicate that PlGF and VEGFR-1 are key molecules in regulating the angiogenic switch during pathological conditions.
So, why is PlGF/VEGFR-1 only angiogenic in pathological but not in physiological conditions? One possible explanation may be that both PlGF and VEGFR-1, while minimally expressed in the adult quiescent vasculature, are strongly upregulated in pathological conditions [11,27,70]. Furthermore, since VEGFR-1 is predominantly expressed as a soluble receptor during vascular development , the minimal levels of transmembrane VEGFR-1 might be insufficient to transmit PlGF-dependent angiogenic signals in embryogenesis [1,58]. Thus, as a reconciling model, we propose that during embryonic development, the abundant soluble VEGFR-1 primarily functions as an inactive ‘sink’; this is evidenced by the lack of angiogenic defects in mice with a truncated VEGFR-1 receptor (VEGFR-1/TK−) or in mice genetically deficient in the VEGFR-1-specific ligands PlGF or VEGF-B. In contrast, during pathological angiogenesis, membrane-bound VEGFR-1 is upregulated and functions as a positive signaling receptor; this is evidenced by the impaired angiogenesis after VEGFR-1 blocking or by the pathology-specific angiogenic defects in PlGF-deficient mice.
The amplification, by PlGF, of VEGF-A-driven angiogenesis also contributes to the angiogenic switch (other mechanisms will be described below). We recently unveiled several mechanisms underlying this synergism, which are likely to cooperate in vivo. First, by activating VEGFR-1, PlGF induces an intermolecular crosstalk between VEGFR-1 and VEGFR-2, which determines the activation of VEGFR-2, with consequent enhancement of the response to VEGF-A. Receptor transphosphorylation, involving an intermolecular transphosphorylation from G-coupled receptors [72,73] or shear-stress-induced integrins [74,75] to receptor tyrosine kinases, or between EGF receptors , has been previously documented, but a crosstalk between VEGF receptors has not yet been demonstrated. Second, PlGF, heterodimerized with VEGF-A, stimulates angiogenesis by inducing the formation of VEGFR-1/VEGFR-2 receptor heterodimers, which trans-phosphorylate each other in an intramolecular reaction, analogous to the activation of heterodimer receptor subunits of the PDGF and EGF family [76,77]. PlGF does, however, not only enhance angiogenesis via amplifying VEGF-A-driven signals. Recent gene-profiling experiments indicate indeed that PlGF is capable of triggering its own signaling via VEGFR-1. Remarkably, this biological response is distinct from that reported for VEGF-A . Taken together, these novel findings indicate that PlGF, by activating VEGFR-1, is able to induce: (i) an intermolecular crosstalk between VEGFR-1 and VEGFR-2 with consequent enhancement of the response to VEGF-A; (ii) an intramolecular crosstalk between the two receptors when in complex with VEGF-A; and (iii) its own signaling cascade.
Efficient blockers of angiogenic disorders
In the past, most strategies designed to block VEGF-A-driven angiogenesis were developed to inhibit the activity of VEGF-A or VEGFR-2 (for an overview, see: http//cancertrials.nci.nih.gov and reference ). PlGF and VEGFR-1 have not been considered attractive therapeutic targets, primarily because insufficient insight in their role in disease was available. In the following chapters, we will discuss the therapeutic potential of VEGFR-1 inhibitors to block angiogenesis during ischemic retinopathy, cancer, atherosclerosis and rheumatoid arthritis (RA), and the therapeutic potential of PlGF to restore blood flow in ischemic diseases including myocardial infarction and peripheral arterial occlusive disease.
The rationale for evaluating VEGFR-1 inhibitors was deduced, in part, by the growing evidence that this receptor played a significant role in the pathological angiogenic switch. Another reason was that VEGFR-1 and VEGFR-2 differ significantly in their expression pattern. While the expression of VEGFR-2 is largely restricted to vascular endothelial cells, VEGFR-1 is also expressed on myeloid cells such as monocytes/macrophages [34,79,80] and neutrophils , and on osteoclasts and preosteoclasts [78,81,82]. We therefore hypothesized that blocking VEGFR-2 on vessels would only block angiogenesis (as amply documented previously [83–86]), while blocking VEGFR-1 would affect both inflammation and angiogenesis (Fig. 6). We therefore compared the effect of an anti-VEGFR-1 monoclonal antibody (mAb), which blocked the binding of VEGF-A and PlGF to VEGFR-1 and inhibited VEGF-A- or PlGF-driven endothelial cell growth , with that of an anti-VEGFR-2 mAb for their potential to suppress angiogenesis in different models. Anti-VEGFR-1 mAb efficiently suppressed VEGF-A-driven neovascularization in the cornea and matrigel implants, and blocked neovascularization in the ischemic retina to a degree comparable with genetic deficiency of PlGF  or inhibition of VEGFR-2 . Compared with a control IgG-treated group, anti-VEGFR-1 mAb dose-dependently blocked angiogenesis and growth of human epidermoid A431 tumors in nude mice, resulting in pale, poorly vascularized and necrotic tumors, in contrast to the large vascularized control tumors. Remarkably, anti-VEGFR-1 mAb was only slightly less active than anti-VEGFR-2 mAb. VEGFR-1 was expressed in tumor-associated vessels, but not in malignant tumor cells, indicating that the effect on tumor growth was mediated through the effect on the tumor vasculature. Anti-VEGFR-1 mAb also attenuated the growth and vascularization of PlGF- or VEGF-A-transduced rat C6 gliomas implanted in nude mice, consistent with other findings .
The observation that anti-VEGFR-1 mAb suppressed pathological angiogenesis compared with anti-VEGFR-2 mAb indicates that VEGFR-1 is a more important therapeutic target for inhibition of angiogenesis than previously presumed. Others have reported that anti-VEGFR-1 antibody does not affect tumor growth/vascularization . However, insufficient amounts of antibody may have been used, since a more complete inhibition of VEGFR-1 reduced tumor angiogenesis in our study . In addition, our data extend previous findings that ribozyme-mediated downregulation of VEGFR-1 suppressed VEGF-A-driven tumor angiogenesis  and that a bifunctional antibody (diabody) against VEGFR-1 and VEGFR-2 was a stronger inhibitor of endothelial cell proliferation than its monofunctional parent antibodies . Growth of Lewis lung carcinomas, endogenously expressing VEGF-A or overexpressing PlGF, was also impaired in mice expressing VEGFR-1 without tyrosine kinase domain (VEGFR-1/TK−) .
Direct effects on endothelial cells
The antiangiogenic effect of the anti-VEGFR-1 antibody can be explained by a direct inhibition of endothelial cell growth. VEGFR-1 has been proposed to regulate VEGF-A/VEGFR-2-driven angiogenesis by functioning as a non-signaling reservoir for VEGF-A and PlGF . However, if this is the only role of VEGFR-1, then the anti-VEGFR-1 antibody would enhance, not reduce, endothelial cell growth in vitro and angiogenesis in vivo, as more VEGF-A would become available to bind VEGFR-2. However, when cultured endothelial cells were treated with the antibody, endothelial cell growth was blocked, strongly suggesting that binding of PlGF to VEGFR-1 activates this receptor and induces an angiogenic signaling cascade. It is also possible that the antiangiogenic effect of anti-VEGFR-1 mAb may have blocked endothelial precursor cells (EPCs) or VEGFR-1+ multipotent adult progenitor cells (MAPCs), which have been documented as contributing to pathological angiogenesis [91–93]. Moreover, the anti-VEGFR-1 mAb may have blocked VEGF-A-driven angiogenesis by lowering the expression of VEGF-A, since activation by PlGF of VEGFR-1 on peri-endothelial fibroblasts, smooth muscle or inflammatory cells, all present in wound or tumor stroma, has been reported to upregulate expression of VEGF-A .
Efficient blockers of inflammatory disorders
Atherosclerosis and rheumatoid arthritis (RA) are chronic inflammatory disorders, characterized by angiogenesis. Although VEGF-A is expressed in atherosclerotic and restenotic vessels ([95,96] and references therein), its role in atherogenesis has remained controversial, since VEGF-A can have both pro- and anti-atherogenic effects [97,98]. On the one hand, VEGF-A may favor plaque growth and destabilization by: (i) stimulating ingrowth of vessels from the adventitial vasa vasorum into the plaque; (ii) increasing the vascular permeability, which facilitates extravasation of plasma proteins and fibrin formation in the plaque; (iii) activating endothelial cells through upregulation of adhesion molecules and monocyte chemoattractant factors, which recruit monocytes to the plaque; and (iv) enhancing smooth muscle proliferation and migration [99–102]. On the other hand, VEGF-A may be atheroprotective by protecting endothelial cells against the toxicity of oxidized low density lipoprotein and by securing endothelial integrity and survival [103–107]. However, it remains unclear whether VEGFR-2 and VEGFR-1 have distinct roles in atherosclerosis. In arthritic joints, expression of VEGF-A, PlGF and VEGF-C and their receptors is upregulated in the synovium, synovial fluid and plasma of RA patients [94,108–113]. While anti-VEGF-A antibodies [108,114] or soluble VEGF receptors [109,115] protect against RA, it remains undetermined which VEGF receptor is involved in the pathogenesis of RA.
We therefore used anti-VEGFR-1 and anti-VEGFR-2 mAb to study the effect of both receptors on the growth and stability of initial (avascular) fatty streak lesions, and of intermediate and advanced (vascularized) plaques in atherosclerosis-prone apolipoprotein-E-deficient mice. Treatment with anti-VEGFR-1 mAb reduced the size of early and intermediate lesions at the aortic root by 50% and the growth of advanced atherosclerotic lesions by ∼ 25% compared with control IgG-treated mice. Remarkably, the anti-VEGFR-2 mAb failed to affect atherosclerotic plaque development at all stages. The atheroprotective effect of anti-VEGFR-1 mAb was attributable to a reduced macrophage infiltration in early as well as in advanced lesions. Based on the antiangiogenic effect of the anti-VEGFR-1 antibody in tumors and ischemic retina (see above), we had anticipated that the reduced plaque growth might, at least in part, be attributable to inhibition of plaque neovascularization. Surprisingly, however, neither the anti-VEGFR-1 nor the anti-VEGFR-2 mAb blocked angiogenesis in atherosclerotic lesions or the surrounding adventitia. Thus, the anti-VEGFR-1 mAb suppressed plaque growth and vulnerability via inhibition of inflammatory cell infiltration, independently of angiogenesis, while the anti-VEGFR-2 mAb, which normally blocks angiogenesis, was ineffective.
The effect of selectively blocking VEGFR-1 or VEGFR-2 was also evaluated in collagen-induced arthritis, which in many aspects reproduces rheumatoid arthritis in humans . Immunostaining of affected joints revealed that VEGF-A was present in inflammatory cells, chondrocytes and cells at the pannus/bone interface, and on endothelial cells in synovial neovessels. VEGFR-2 was only present on synovial neovessels, whereas VEGFR-1 and PlGF were expressed by inflammatory and endothelial cells in the inflamed synovium. Treatment with anti-VEGFR-1 mAb reduced the incidence of joint disease by 60%, while all IgG-treated mice developed signs of arthritis in the paws and ankles. Remarkably, anti-VEGFR-1 mAb treatment suppressed the development of paw swelling, erythema and ankylosis by 85%. In addition, anti-VEGFR-1 mAb significantly protected against bone destruction (Fig. 2). The effect observed with treatment by anti-VEGFR-1 mAb was specific, since anti-VEGFR-2 mAb was ineffective. Synovial angiogenesis and infiltration by inflammatory cells was reduced by anti-VEGFR-1 mAb treatment, which also suppressed the activation of leukocytes and their production of tumor necrosis factor α and monocyte/macrophage chemoattractant protein-1, cytokines implicated in arthritis  (Fig. 7). Previous in vitro studies have documented VEGFR-1-mediated signaling in monocytes and macrophages [34,79,80]. Thus, inhibition of VEGFR-1, but not of VEGFR-2, protects mice against arthritic joint destruction by suppressing synovial inflammation and neovascularization. Taken together, the failure of the anti-VEGFR-2 mAb, but not of the anti-VEGFR-1 mAb, to block arthritis and atherosclerosis and the angiogenesis-independent atheroprotective effect of the anti-VEGFR-1 mAb indicates that suppression of inflammation, not angiogenesis, was primarily responsible for the observed effects (Figs 6 and 8). In addition, these data argue in favor of a pro-atherogenic effect of the VEGFR-1 ligands VEGF-A and PlGF, by stimulating plaque growth and destabilization (Fig. 8). Which ligand is primarily responsible for plaque growth and destabilization awaits further studies in mice with a specific deficiency of these ligands.
VEGFR-1: novel role in hematopoietic stem cells
Transplantation of bone marrow, transduced with a retroviral GFP-expressing vector, revealed that the anti-VEGFR-1 mAb blocked the accumulation of GFP-labeled bone marrow-derived cells in atherosclerotic lesions and arthritic joints . Reduced leukocyte accumulation could result from an effect of anti-VEGFR-1 mAb on the infiltration of circulating myeloid cells in inflamed lesions, and/or from an effect on the differentiation or mobilization of these cells or their progenitors from the bone marrow into the peripheral blood. In support of the latter mechanism, the anti-VEGFR-1 mAb partially abrogated the disease-associated increase in the numbers of circulating monocytes and granulocytes (Fig. 7). In addition, mobilization of hematopoietic progenitors into the peripheral blood was suppressed by the anti-VEGFR-1 mAb by 75% (Fig. 7). In another publication, a molecular mechanism for the role of PlGF/VEGFR-1 in hematopoietic stem or progenitor cell mobilization was proposed in a model of hematopoietic recovery after myelosuppression [118,119]. While anti-VEGFR-1 impaired recovery, exogenous PlGF mobilized VEGFR-1+ hematopoietic stem and progenitor cells. In the early phase, PlGF recruited VEGFR-1+ hematopoietic precursors through chemotaxis, while in the late phase, PlGF acted through MMP-9-mediated release of soluble Kit ligand . Whether the same molecular mechanisms also apply to stem cell recruitment during ischemia/inflammation is currently under investigation. Furthermore, VEGFR-1 appeared to be crucial for hematopoietic stem cell survival in vitro.
PlGF: a novel attractive candidate for stimulation of ischemic tissue revascularization
Growth of blood vessels in the adult occurs via several mechanisms: (i) angiogenesis and arteriogenesis; new vessels, which initially only consist of endothelial cells, are subsequently stabilized via coverage with smooth muscle cells ; (ii) postnatal vasculogenesis, or the recruitment of bone marrow-derived circulating endothelial progenitors (CEP) [27,122–125]; and (iii) collateral growth, where pre-existing collaterals are remodeled and grow in size and number. The recruitment of inflammatory cells (monocytes/macrophages) to the activated endothelium of collateral vessels and their production of proteinases, cytokines and growth factors is critical in the initial phase of this process, whereas adventitial fibroblasts may fine-tune the collateral remodeling in the late phase . Revascularization of ischemic tissues, by stimulation of all these mechanisms, is an attractive therapeutic goal, but a formidable challenge. While administration of VEGF-A markedly improved perfusion and rescued ischemic tissues in preclinical studies [107,127–139], clinical trials with VEGF-A have thus far not yielded the expected results . This has raised questions whether additional molecules may be required to form mature, durable and functional vessels. However, novel strategies should avoid hemangiomagenesis, the occurence of edema, or hypotension. In addition, new vessels should not regress, but should be durable and persist after the angiogenic treatment is stopped. Delivery of VEGF-A suffers from some of these side-effects, even after local administration [141,142]. We recently documented that treatment of mice with PlGF improved revascularization of ischemic heart and limbs. PlGF is an attractive therapeutic candidate, as it affects all three mechanisms of vascular growth. We will discuss the role of PlGF on these processes in the following sections.
The rationale for evaluating PlGF in therapeutic revascularization was deduced from gene-targeting studies (Fig. 2). Indeed, collateral growth and revascularization of ischemic myocardium were significantly impaired in PlGF-deficient mice . Moreover, PlGF and VEGFR-1 expression were upregulated in postischemic collateral vessels and myocardium [11,27]. Administration of PlGF not only rescued the impaired revascularization in the ischemic heart and limbs of PlGF-deficient mice , but also stimulated revascularization in these tissues in wild-type mice  and rabbits , indicating that PlGF has therapeutic effects. Both PlGF isoforms were equally potent, and the therapeutic potential of PlGF to stimulate myocardial angiogenesis or collateral growth in the ischemic limbs was comparable or even superior to that of VEGF-A, respectively . PlGF increased postischemic limb perfusion up to 3-fold more than VEGF-A or saline (Fig. 2). This effect was attributable to an increase in the number of collaterals, the enlargement of their luminal diameter, and of the total collateral vascular perfusion area . Importantly, PlGF treatment also enhanced the functional motor recovery of the ischemic limb to normal levels, when analyzed in baseline conditions and after endurance-exercise test . VEGF-A treatment only minimally affected postischemic collateral growth and functional recovery , consistent with previous reports .
An important observation was that PlGF administration stimulated the formation of durable, mature, stable vessels . Upon transient expression of PlGF using adenoviral gene transfer in the skin, vessels significantly grew and enlarged. These vessels became covered by perivascular smooth muscle cells and, remarkably, these mature vessels persisted for more than a year (Fig. 9). This finding is important, as it demonstrates that transient expression of PlGF during a short period (<10 days) initiates the formation of durable, mature and functional vessels, even long after PlGF transgene expression was arrested. Another beneficial characteristic of PlGF therapy is that it did not cause any undesired hemangiomagenesis, nor did it induce edema or bleeding; characteristic side-effects of VEGF-A treatment  (Fig. 4). In addition, systemic delivery of PlGF, even at high doses, was well tolerated, while comparable doses of VEGF-A killed the animals due to hypotension, edema and circulatory shock.
Many patients with ischemic heart and limb disease, eligible for therapeutic angiogenesis, are old and have diabetes and/or atherosclerosis. These conditions are known to impair the angiogenic response to VEGF-A treatment [128,130]. We therefore examined whether adjunctive VEGF-A plus PlGF treatment was superior than monotherapy with each factor alone, using the urokinase-deficient mouse model, as it is resistant to therapeutic angiogenesis with VEGF-A or PlGF alone . Noticeably, revascularization of the ischemic myocardium was only stimulated when VEGF-A and PlGF were coadministered, indicating that PlGF, in this setting, is also an attractive therapeutic candidate.
PlGF stimulates ischemic tissue revascularization via effects on multiple cell types
The therapeutic efficacy of PlGF may be explained by its intrinsic capacity to affect, directly and indirectly, multiple cell types, involved in all stages of vascular growth in the adult [11,27] (Fig. 10). First, PlGF has direct effects on endothelial cell growth, migration and survival. It also synergizes with VEGF-A and amplifies the angiogenic activity of the latter via an intermolecular crosstalk between VEGFR-1 and VEGFR-2 (see above). PlGF may also upregulate the expression of VEGF-A, thereby further amplifying this synergism . These effects may explain why PlGF stimulates the sprouting of new capillaries (angiogenesis). Second, PlGF directly affects smooth muscle cells and fibroblasts, which express VEGFR-1 [11,33,147], but may also indirectly influence smooth muscle cell proliferation and migration through cytokine release from activated endothelial cells . Through these effects, PlGF recruits smooth muscle cells around nascent vessels, which are initially only naked endothelial tubes, thereby stabilizing them into mature, durable and non-leaky vessels. Moreover, PlGF may further fine-tune remodeled collateral vessels by affecting adventitial fibroblasts. Third, PlGF stimulates the mobilization of VEGFR-1 positive hematopoietic progenitors from the bone marrow [27,89] and, indirectly via upregulation of VEGF-A expression, recruitment of VEGFR-2-positive endothelial progenitors to the ischemic tissue [11,148] (Fig. 10). These cells may stimulate new vessel growth by direct incorporation into the vessel wall and/or creating a proangiogenic microenvironment through the release of angiogenic molecules. Fourth, PlGF is chemoattractive for monocytes and macrophages, which express VEGFR-1 [11,81]; activated macrophages produce PlGF, thereby providing a positive feedback  (Fig. 7). Tissue infiltration of macrophages is impaired in PlGF-deficient mice, and the impaired postischemic collateral growth in PlGF-deficient mice can be rescued by transplantation with PlGF positive bone marrow . Interestingly, PlGF treatment amplifies the chemotactic recruitment of monocytes and macrophages [80,118], and increases monocytic tissue infiltration two-fold during postischemic collateral growth , through direct VEGFR-1 activation [11,58,66]. Additionally, PlGF deficiency impairs macrophage activation, whereas exogenous PlGF stimulates macrophage activation , thereby increasing the release of cytokines . PlGF may also prolong macrophage survival . Moreover, the inflammation-mediated remodeling of the extracellular matrix, a prerequisite for smooth muscle cell migration during collateral growth , is decreased in PlGF-deficient mice , whereas exogenous PlGF stimulates proteolytic activity through direct VEGFR-1 activation [118,151]. All these effects of PlGF on inflammatory cells may contribute to the growth of collateral vessels. Fifth, PlGF stimulates the mobilization of bone marrow-derived VEGFR-1-positive hematopoietic stem and progenitor cells, which contribute to postischemic revascularization . Finally, PlGF, by increasing VEGF-A secretion, may affect postischemic muscle regeneration .
PlGF and VEGFR-1 have been neglected as therapeutic candidates for over a decade. Recent gene targeting studies have revealed, however, that PlGF and VEGFR-1 are key regulators of the angiogenic switch in ischemia, inflammation and cancer. Proof of principle has now been provided in preclinical animal models that inhibition of VEGFR-1 efficiently blocks cancer, arthritis and inflammation, while delivery of PlGF stimulates revascularization of ischemic heart and limbs. One of the reasons for this unsuspected efficacy of PlGF is that this growth factor affects multiple cell types, including endothelial cells, smooth muscle cells, inflammatory cells and hematopoietic progenitor/stem cells. Ongoing studies in larger animal models and clinical trials will reveal the therapeutic potential of PlGF and VEGFR-1 for humans.