ReviewAngiogenesis

Positive and Negative Modulation of Angiogenesis by VEGFR1 Ligands

See allHide authors and affiliations

Science Signaling  24 Feb 2009:
Vol. 2, Issue 59, pp. re1
DOI: 10.1126/scisignal.259re1

Abstract

Vascular endothelial growth factor–A (VEGF-A) is a key target for new antiangiogenic drugs for the treatment of both malignant and nonmalignant human diseases. Vascular effects of VEGF family members are mainly mediated by VEGF receptor 2 (VEGFR2). Conversely, the function and signaling of VEGFR1, which is present on endothelial and nonendothelial cells, are poorly understood. Intriguingly, two of five members in the VEGF family—VEGF-B and placental growth factor (PlGF)—are exclusive ligands for VEGFR1 and do not interact with the other VEGFRs, VEGFR2 and VEGFR3. These VEGFR1-specific ligands may be important therapeutic targets for the treatment of cancer. This review discusses the distinctive roles of VEGFR1 and its ligands PlGF and VEGF-B in the mediation of angiogenic signaling and considers the therapeutic potential of targeting these particular vascular factors.

Introduction

Members of the vascular endothelial growth factor (VEGF) family of proteins exhibit multiple functions, including modulation of angiogenesis, vasculogenesis, vascular leakage, neurogenesis, inflammation, hematopoiesis, and lymphangiogenesis (14). This diversity of biological function makes VEGF biology a dynamic research field (3). Furthermore, the impairment of VEGF-regulated vascular functions may contribute to the onset, development, and progression of several common and lethal human diseases, including cancer, cardiovascular disorders, diabetic complications, retinal degeneration, and chronic inflammation (5). Thus, the therapeutic targeting of the VEGF-mediated signaling system has the potential to be an effective approach to the treatment of a range of malignant and nonmalignant diseases. Successful examples of this strategy include the demonstrated statistically significant benefit associated with the addition of bevacizumab (Avastin), a neutralizing humanized monoclonal antibody specific for VEGF-A, to conventional regimens used in the treatment of colorectal cancer, non–small cell lung carcinoma, renal cancer, and breast cancer (69). Additionally, VEGF-A inhibitors, including pegaptanib (Macugen) and ranibizumab (Lucentis), show remarkable therapeutic benefits in the treatment of age-related macular degeneration (AMD) (10).

With the exception of renal cancer, the clinical benefit conferred by drugs that interfere with VEGF-A to patients with other cancers remains modest (11, 12). Additionally, unconventional drug resistance—which involves compensatory switching to non–VEGF-A angiogenic factors, ligand-independent autophosphorylation of VEGF receptors, recruitment of myeloid cells, tissue hypoxia, and abnormality of tumor endothelial cells—tends to develop after relatively long periods of therapy (13, 14). Although the mechanisms underlying the low therapeutic efficacy and drug resistance remain poorly understood, compensatory expression of genes encoding other VEGF-related and nonrelated angiogenic factors may, in part, explain the relative clinical ineffectiveness of agents that block VEGF-A function. For example, placenta growth factor (PlGF), fibroblast growth factor (FGF), and platelet-derived growth factor (PDGF) are often abundant in various tumor types and can substantially modulate angiogenesis or the effects of VEGF-A on vascular functions (13, 15, 16). Thus, therapeutically blocking the function of PlGF and these other factors may be an important concurrent approach that will improve the efficacy of existing antiangiogenic cancer therapies.

This review is focused on discussion of emerging evidence relating to the functional biology of VEGF receptor 1 (VEGFR1, also known as FLT1). In particular, the therapeutic potential of targeting the ligand components of this receptor signaling system is discussed in detail.

Diversity in the VEGF Ligand and Receptor Families

Five structurally related glycoproteins in the VEGF family—VEGF-A, PlGF, VEGF-B, VEGF-C, and VEGF-D—bind to the extracellular immunoglobulin (Ig)–like domains of three transmembrane receptor tyrosine kinases (VEGFR1, 2, and 3), which are mainly present on endothelial cells (Fig. 1) (1721). In addition to binding to receptor tyrosine kinases, certain VEGF family isoforms also interact with neurophilin (NP) 1 and 2, which serve as co-receptors of the VEGFR1, VEGFR2, and VEGFR3 receptor tyrosine kinases and modulate the vascular functions mediated by VEGFR2 and VEGFR3 (2227), including angiogenesis and lymphangiogenesis, processes critical for tumor growth and metastasis.

Fig. 1

The VEGF family comprises a series of structurally related secreted glycoproteins that bind to homo- or heterodimeric transmembrane receptor tyrosine kinases to modulate a range of cellular processes, including angiogenesis and lymphangiogenesis. Certain isoforms also bind to neurophilin (NP1 and NP2) co-receptors, which appear to contribute to the cellular responses to these ligands. In the VEGFRs, filled circles in the extracellular regions represent the Ig-like domains and black rectangles represent the intracellular tyrosine kinase domains. In the NP co-receptors, the extracellular domain contains two tandem complement binding factor (CUB) domains (membrane-distal brown ovals), two tandem Factor V/VIII homology domains (brown squares), and one MAM (meprin, A5 antigen, receptor tyrosine phosphatase m) domain (membrane-proximal brown oval).

According to their functions and receptor binding patterns, the VEGF family can be further divided into three subgroups. The type I ligands, PlGF and VEGF-B, bind only to VEGFR1, and their physiological functions are the least well characterized of the VEGF family members. The type II ligand, VEGF-A—which is the prototype of the ligand family and is commonly called VEGF—binds to VEGFR1 and VEGFR2 (also known as KDR in humans and Flk-1 in mice) and displays potent angiogenic and vascular permeability activity. The type III ligands, VEGF-C and VEGF-D, bind to both VEGFR2 and VEGFR3 and exert angiogenic and lymphangiogenic activity. Accumulating evidence shows that VEGFR2 is the crucial receptor mediating angiogenic and vascular permeability activity, whereas VEGFR3 mainly mediates the lymphangiogenic function (28). In contrast to VEGFR2 and VEGFR3, the functions mediated by VEGFR1 are less well understood, although several vascular-related and nonvascular-related functions have been suggested (29, 30).

Not only are there three main types of VEGF ligands, but in humans, several of the ligands are further diversified through alternative splicing. This results in the generation of different protein isoforms with variable affinity for heparin (Fig. 2). There are at least four different isoforms of PlGF; of these, only PlGF-2 and PlGF-4 bind heparin (21, 31, 32). VEGF-A and VEGF-B transcripts are also commonly alternatively spliced. In particular, VEGF-B encodes two isoforms, VEGF-B167 and VEGF-B186; the shorter VEGF-B167 form has a strongly basic cysteine-rich, heparin-binding domain, whereas VEGF-B186 lacks such a domain at the C terminus, does not bind to heparin, and thus is more soluble when secreted (33). Indeed, the heparin-binding affinity of the different protein isoforms might establish a gradient effect from the location where these factors are produced, with heparin-binding forms of PlGF, VEGF-B, and VEGF-A interacting tightly with heparan sulfate proteoglycans distributed on the cell surface and extracellular matrix, thereby preventing their further diffusion. The complexity of encoded products from these genes raises the possibility of a diversified signaling capability.

Fig. 2

Through alternative splicing of their primary gene transcripts, the human PlGF, VEGF, and VEGFB genes can encode multiple different protein isoforms. When produced within the same cell, the monomers can, in principle, combine to form different heterodimers that may have different functional properties. The balance of different types of heterodimers produced is likely to be modulated at least in part by the relative expression rate of each gene and the degree of alternative splicing of each gene transcript. Not all heterodimeric combinations shown here have been reported.

Additional functional diversity may occur through the formation of dimers between VEGF family members. For example, VEGF-A may form heterodimers with either PlGF (34) or VEGF-B (33). In humans, because of the alternative splice variants, this has the potential to create enormous diversity and presents a challenge to assessing the functional consequences of heterodimerization (35). For example, the six human VEGF-A and four human PlGF isoforms could theoretically form 24 VEGF-PlGF heterodimers, and VEGF-A could theoretically form 12 different combinations with the two human VEGF-B variants. Different heterodimers may well have different matrix binding affinities, by virtue of the strength of their interaction with heparan sulfate proteoglycans, and could create a gradient from their producing sites associated with the level of that affinity (Fig. 2)

Although VEGF-A appears to share overlapping functions with VEGF-C and -D, including stimulation of hemangiogenesis, vascular permeability, and lymphangiogenesis, PlGF and VEGF-B do not seem to show such obvious vascular functions under normal physiological conditions. Indeed, homozygous deletion of the genes encoding either VEGF-B or PlGF did not result in any severe phenotypic changes in adult mice (36, 37). However, both VEGF-B and PlGF may modulate the vascular functions of VEGF-A and may have important roles in triggering pathological angiogenesis (19, 34, 3841). Thus, these VEGFR1-exclusive binding ligands are potentially important therapeutic targets.

Exploring the Biological Actions of VEGFR1

The gene encoding VEGFR1, unlike that encoding VEGFR2, is widely expressed in many nonendothelial cell types, including bone marrow progenitor cells, monocytes, macrophages, vascular smooth muscle cells, and various tumor cells (42); hence, VEGFR1 might mediate a broad spectrum of biological functions (Fig. 3A). Despite its high affinity for VEGF-A, VEGFR1 when overexpressed in endothelial cells, pericytes, vascular smooth muscle cells, or fibroblasts does not result in substantial proliferative activity when those cells are stimulated with VEGF-A, VEGF-B, or PlGF (43, 44).

Fig. 3

(A) The presence of VEGFR1 on endothelial cells is essential during embryological vasculogenesis and persists in adult animals, suggesting a continued role in endothelial maintenance. VEGFR1 is also present many other cell types, including hepatocytes, neural cells (NCs), vascular smooth muscle cells (VSMCs), and monocytes, which is consistent with the involvement of VEGFR1-mediated signaling in a spectrum of cellular processes. TCs, tumor cells; ICs, inflammatory cells. (B) The administration of drugs that inhibit angiogenesis drugs (anti-VEGF) leads to elevation in circulating VEGFR1 ligands including PlGF and VEGF-B, which may be an adaptive response from the tumor linked to drug resistance.

Intriguingly, VEGFR1 can be generated as both a full-length transmembrane form and a short soluble form (sVEGFR1) consisting of only the extracellular ligand-binding domain (45). The soluble form of VEGFR1 has the potential to act as a decoy receptor for VEGF-A, and thus it may indirectly inhibit VEGF-A’s proangiogenic activities. For example, a native soluble VEGFR1 produced by tumors inhibits tumor growth, angiogenesis, and metastasis (46). Similarly, a soluble VEGFR1 completely inhibited corpus luteum angiogenesis, which is essential for ovulation and fertility (47). Thus, VEGFR1 may function both directly (through stimulation of its tyrosine kinase domain) and indirectly, effectively reducing the availability of its ligands for other receptors.

Although VEGFR1 contains a classical tyrosine kinase domain, VEGFR1-mediated downstream signaling remains poorly understood primarily because of its low activity in cultured cells, even in VEGFR1-overexpressing endothelial cells (43, 44). Although several tyrosine residues of VEGFR1 become autophosphorylated upon VEGF-A stimulation, this generally leads to only weak responses of endothelial cell proliferation, migration, and tube formation (48, 49). However, these in vitro assay systems may not necessarily recapitulate the in vivo angiogenesis process; for example, ablation of the Vegfr1 gene in mice leads to embryonic lethality due to disorganization of the vascular architecture (50). The formation of functional vascular networks and vascular remodeling is almost impossible to recapitulate with in vitro assays. A number of studies in different cell types, including endothelial cells and fibroblasts, have indicated that PlGF-stimulated VEGFR1 signaling can modulate the phosphatidylinositol 3-kinase–Akt (PI3K/Akt) pathway, the ERK mitogen-activated protein kinase (MAPK) pathway, and the Janus kinase–signal transducer and activator of transcription 3 (JAK/STAT3) pathway (5155).

Homozygous deletion of VEGFR1 in mice results in embryonic lethality due to overgrowth of endothelial cells, leading to disorganization and dysfunction of the vasculature (50). Surprisingly, loss of the tyrosine kinase domain of VEGFR1 alone produces a nearly healthy phenotype with normal vasculatures (56). These findings show that a membrane-anchored and a soluble VEGFR1 might coordinately regulate angiogenic activity triggered by VEGFR1-specific ligands and thereby ensure the development of healthy vasculature during embryonic growth. However, little is known about the regulatory mechanisms controlling the differential expression of these VEGFR1 isoforms. Perhaps a net balance in the production of these VEGFR1 isoforms determines vessel growth and regression.

These studies in mice suggest the possibility that the primary function of VEGFR1 is as a negative regulator of vascular development. One possible mechanism may be that VEGFR1 acts as a decoy receptor, competing with VEGFR2 for binding to proangiogenic VEGF ligands. However, it is also possible that ligand binding to the receptor and activation of its tyrosine kinase activity are important for modulation of angiogenesis. Indeed, VEGFR1 has been shown to play positive roles in mediating inflammatory responses, tumor growth, and atherosclerosis (42). The underlying mechanism might involve transphosphorylation of VEGFR2 by the phosphorylated VEGFR1 (41).

Exploring the Functions of PlGF and VEGF-B

Clues from expression patterns

Insight into the biological functions of these VEGFR1-specific ligands can be obtained by examining their expression patterns. The PlGF gene (mapped to human chromosome 14q24.3) was originally cloned from a human placental complementary DNA (cDNA) library (18). In addition to the placenta, the gene is abundantly expressed in the thyroid under normal physiological conditions (57). VEGF-B, encoded by the VEGFB gene at human chromosome 11q13 (58), is abundant in tissues and organs with high metabolic rates, such as the myocardium, brown fat, and striated muscles (59, 60). During development, the expression of VEGFB in embryos and newborns is elevated in the myocardium relative to the adult myocardium (61).

Under pathological conditions, PlGF abundance is elevated in various cell types, including vascular endothelial cells and smooth muscle cells, keratinocytes, hematopoietic cells, retinal pigment epithelial cells, and many different tumor cells (34, 39, 6270). In addition, in cultured immortalized or transformed mouse embryonic fibroblasts and NIH 3T3 cells, PlGF is also abundant (40). As seen for the gene encoding VEGF, in tumors the expression of the gene encoding PlGF may increase after administration of cytotoxic anticancer treatments (71). The circulating levels of PlGF protein can also markedly increase during anti-VEGF therapy, which suggests that PlGF might contribute to the compensatory mechanism of drug resistance (72) (Fig. 3B). These findings imply that tissue and organ stress could be important factors that stimulate PlGF expression.

Consistent with its discovery in a placental tissue cDNA library, PlGF is abundant in the trophoblasts of the placenta at the middle to late stages of pregnancy (32, 73). The amount of PlGF persistently increases toward the terminal development of the placenta, and its biosynthesis seems to be limited to trophoblasts and stromal cells (32, 73). This temporal pattern of PlGF production may reflect a role in preventing excessive neovascularization and overgrowth of the placenta tissue by down-regulating angiogenesis. PlGF and VEGF-A are not always cosynthesized in the same cell population, and thus the role of PlGF may not be limited to countering or limiting VEGF-A–mediated angiogenesis (40, 74). Interestingly, high concentrations of placenta-derived sVEGFR1, which binds to both PlGF and VEGF-A, may contribute to the development of preeclampsia (75) (a dangerous condition characterized by high blood pressure, protein in the urine, tissue swelling, sudden weight gain, and often vision impairment that occurs during pregnancy and the postpartum period), perhaps indicating the crucial function of one or both of these ligands in maintaining normal functioning of the maternal vasculature.

Nonvascular functions of PlGF and VEGF-B

The biological function of VEGF-B remains enigmatic. VEGF-B might have important nonvascular roles, such as promoting the survival of smooth muscle cells through the inhibition of apoptosis and inducing inflammatory responses (37, 59, 61, 76, 77). Although the source of VEGF-B in different tissues might vary, skeletal muscle tissues and myocardium have been shown to have high levels of VEGF-B mRNA expression (19). PlGF also has nonvascular functions. For example, the high-molecular-weight form PlGF2 also binds to NP1, which may mediate neuronal outgrowth and guidance (78). The nonvascular functions of VEGF-B, and possibly those of PlGF, are consistent with the broad tissue distribution of their receptors, VEGFR1 and NP1.

Positive and negative modulation of VEGF function by PlGF

Several studies in various experimental models show that PlGF directly induces angiogenesis; examples include PlGF-induced neovascularization in the rabbit cornea and in the chick choriallantoic membrane (79). Although in the cornea and choriallantoic membrane PlGF appears to be proangiogenic, studies with cultured endothelial cells from the human umbilical vein show that PlGF does not appear to directly stimulate cell migration and proliferation (34, 74). Again, differential effects of PlGF in vitro and in vivo might reflect different aspects of the angiogenic process. Furthermore, PlGF homodimers seem to elicit less potent angiogenic responses than do VEGF-A–PlGF heterodimers (34, 80). Similar to VEGF-A, PlGF may also act as a survival factor for endothelial and nonendothelial cells (81). The conflicting data regarding the endothelial cell response to PlGF may be partly due to abundant endogenous PlGF production by the cultured endothelial cells, which may mask the effects of exogenously added PlGF. Indeed, PlGF-deficient endothelial cells show enhanced responses to exogenous PlGF (40).

Although PlGF might directly regulate endothelial and angiogenic activity in vitro and in vivo, it also modulates VEGF-A function in both a positive and negative manner, depending on its relationship to VEGF-A production (34, 38, 82). If PlGF and VEGF-A are produced in the same cell population, these two factors preferentially form heterodimers rather than homodimers, and VEGF-A–PlGF heterodimers display weaker angiogenic activity than do VEGF-A homodimers (38). Thus, PlGF might inhibit VEGF-induced angiogenesis by the formation of functionally less active (with respect to VEGFR2 signaling) heterodimers, resulting indirectly in the generation of fewer VEGF homodimers. The inhibition of angiogenesis and growth by overexpression of PlGF in either xenografted or spontaneous tumor models may be explained by the production of VEGF-A–PlGF heterodimers (83, 84). Note that these PlGF-mediated antagonistic effects on tumor angiogenesis have only been observed in tumors that have been genetically engineered to overexpress PlGF. The negative role of PlGF in relation to the regulation of physiological and pathological angiogenesis consequently warrants further investigation.

In contrast to the negative regulation of VEGF-A function when PlGF and VEGF-A are produced in the same cells, PlGF may also potentiate VEGF-induced angiogenic activity when both factors are produced in different cells (38, 74). PlGF homodimers may potentially compete with VEGF-A homodimers for VEGFR1 binding, indirectly rendering more VEGF molecules available for binding to VEGFR2, which would thereby increase proangiogenic and vasculogenic signals (Fig. 4). Another mechanism by which PlGF may enhance VEGF-A–induced vascular function is through activation of VEGFR1 by PlGF, which in turn may lead to phosphorylation and transactivation of VEGFR2 (41). Intriguingly, VEGFR1 and VEGFR2 form heterodimers in blood vessel endothelial cells and, in theory, VEGFR heterodimers could bind to VEGF-A–PlGF heterodimers (85). Currently, the biological functions mediated by VEGFR1-VEGFR2 heterodimers are not understood, because it is impossible to separate the responses of receptor heterodimers from those of homodimers within cells and in vivo angiogenesis models.

Fig. 4

(A) The functional consequences of PlGF production are likely to be radically different according to whether VEGF is also produced in the same cell. Where cells produce both ligands, a proportion of heterodimers will form, with the balance between particular homodimers and heterodimers influenced by the relative expression rates of each gene. (A) Left panel: When the balance is such that PlGF is more abundant than VEGF-A, then the proportion of secreted VEGF homodimers will likely be small, and the stimulatory signal for angiogenesis will be relatively weak. In this situation, up-regulation of PlGF has a negative effect on VEGF-associated signaling. Right panel: If PlGF and VEGF are produced in similar amounts from different cells, then homodimers of each will predominate. In this case, the exclusive binding of PlGF homodimers to VEGFR1 is likely to result in an increased proportion of VEGF homodimers available to bind the proangiogenic VEGFR2 homodimers. In this situation, high levels of PlGF have a positive effect on VEGF-associated signaling. (B) Active PlGF signaling through VEGFR1 may also modify the behavior of a range of nonepithelial cell types, which may influence angiogenesis indirectly or regulate the behavior of these cells in processes that are independent of angiogenesis.

PlGF-2, but not PlGF-1, binds to NP1, and this interaction might be mediated by amino acid sequences encoded by exons 6 and 7 (25). However, this interaction does not seem to influence PlGF-induced endothelial cell migration, and neither isoform stimulates cell proliferation (25). Experiments in mouse xenograft tumor models show that administration of neutralizing monoclonal antibodies to either NP1 or NP2 inhibits tumor growth and metastasis by impairing VEGFR-induced angiogenesis (22, 86). Thus, the interaction between PlGF and NP receptors could potentially lead to further positive regulation of VEGF-A function by enhancing VEGFR2-trigerred angiogenic signaling (27). This possibility warrants further investigation.

Modulation of Tumorigenesis and Progression

The effect of hypoxia

The net effect of PlGF and VEGF-A on angiogenesis is likely also influenced by tissue hypoxia, which is especially important in the context of solid tumors. In solid tumors, tissue hypoxia may shift the ratio of expression of various angiogenic factors. For example, hypoxia increases the expression of VEGFA through transcriptional regulation mediated by hypoxia-inducible factor (HIF)–1α or HIF-2α. The effect of tissue hypoxia on PlGF expression may be more complicated, with some studies showing enhanced expression (40, 63, 8791) and others showing no change as a consequence of tissue hypoxia (82). Nevertheless, differential regulation of PlGF and VEGFA expression by hypoxia has the capacity to affect angiogenesis. For example, in tumor cells that produce both VEGF-A and PlGF, hypoxia may affect the ratio of their production and thereby the degree of heterodimerization. Thus, tissue hypoxia may not only elevate the abundance of VEGF-A, but may also alter the abundance of other angiogenic regulatory factors, thus leading to altered angiogenic activity.

PlGF in primary human tumors

In contrast to the restricted pattern of expression in normal cells that are not under hypoxic stress, PlGF protein is abundant in various types of human tumors (34, 39). Some studies have reported that, relative to matched normal tissue from the same organ, PlGF expression is elevated in breast, lung, colorectal, and gastric cancers (69, 9294). Another example of elevated PlGF expression in a tumor is rare vascular neoplasia that arises from a reciprocal t(10;14)(p13;q24) translocation. The breakpoint on chromosome 14 is in the vicinity of the PlGF gene, and immunohistochemical analysis of the tumor tissue from the patient revealed strong expression of the messenger RNA (mRNA), implying that the translocation had resulted in up-regulation of PlGF, which may have provided a selective advantage to the tumor cell (95). Thus, in some cases, the levels of PlGF protein or PlGF mRNA in tumors correlate with the clinical characteristics or outcome of disease (69, 9294, 96, 97), such that tumors with higher amounts of PlGF protein or PlGF mRNA are associated with more aggressive disease or reduced survival.

Not all gene expression data are consistent with this hypothesis of elevated PlGF as causally related to disease prognosis. Conflicting data for lung and colorectal cancers have also been reported in which PlGF expression appeared to be lower than that in matched normal tissue (84). This same study also reported that the PlGF promoter was heavily methylated, and thus likely to be less active, in tumor cell lines. PlGF expression is apparently lower in prostatic tumors than in normal prostatic tissue (97); a similar situation occurs in the thyroid, where strong PlGF expression has been noted in normal tissue and its production is decreased in thyroid tumors (57).

Because the gene expression data from tumors show variability, it may be that PlGF is an attractive target in a subset of tumors. However, several examples have been reported where the data are consistent with the hypothesis that high-level expression of PlGF in common human tumor types may be linked to a more aggressive cancer phenotype.

Although correlative, overexpression of genes in tumors relative to expression in normal tissue, or the association of high levels of expression with poor clinical outcome, is not necessarily proof of a causal link between deregulation of the gene and malignancy, nor of the suitability of a gene product as a therapeutic target. However, gene expression analysis can help to focus efforts.

PlGF as a positive regulator of tumorigenesis and metastasis

Model systems can provide useful insight into the therapeutic usefulness of a particular gene identified through expression analysis as potentially important for tumorigenesis. In a transgenic mouse model, constitutive overexpression of PlGF under the control of the keratin 14 promoter, which limits expression to the skin, was associated with an increase in dermal vascularization that was reflected by an increase in the number, branching, and size of blood vessels and an increase in vascular permeability (98). In the same transgenic PlGF-expressing mice, overexpression of PlGF in keratinocytes was also linked to an increase in tumor growth, invasiveness, and the numbers and sizes of metastases in animals inoculated intradermally with B16-BL6 melanoma cells (99). When injected into mice, B16F10 melanoma cells transfected to overexpress PlGF also exhibited increased tumor vasculogenesis relative to tumors that arose from mice injected with B16F10 melanoma control cells (100).

Experiments with cultured cancer cell lines suggest that PlGF has a role in controlling cell motility and invasiveness. In breast cancer cell lines, exogenous addition of PlGF-2, but not VEGF-A, to culture medium stimulated two key characteristics associated with metastatic potential: motility (detected by cell migration assay) and invasiveness (detected by Matrigel spheroid assay) (101). Although this may represent a direct effect of PlGF on tumor cells, it may also be the result of PlGF competing with VEGF to display indirect effects on tumor cells. The PlGF-2–associated stimulation of motility and invasion was suppressed in this in vitro system by the addition of a peptide that blocked the heparin-binding site of VEGFR1 or by an antibody to PlGF (101). Treatment of a human non–small cell lung cancer cell line with a ribozyme transgene construct targeting PlGF inhibited cellular motility (102).

Studies with xenografted tumors also support a role for PlGF in metastasis or growth. For example, administration of the blocking peptide to mice bearing subcutaneous or orthotopic breast cancer xenografts resulted in decreased rates of spontaneous metastasis (101). Furthermore, a neutralizing monoclonal antibody to mouse PlGF2 (αPlGF) inhibited growth, metastases, or both in various model systems, including human tumor xenografts. The antibody was effective against tumors that were either sensitive or resistant to VEGFR inhibitors. The αPlGF inhibited tumor angiogenesis, lymphangiogenesis, and intratumoral macrophage recruitment (39).

PlGF as a negative regulator of tumor growth angiogenesis

Several investigations have produced data that, at least superficially, appear to contradict the clinical and experimental observations suggesting that the abundance of PlGF positively contributes to tumor growth and tumor-associated angiogenesis. For example, Xu and colleagues reported that overexpression of PlGF impaired tumor growth in xenograft models (84). In this study, a full-length PlGF-2 plasmid expression construct was stably transfected into three human cell lines (lung carcinoma, colon carcinoma, and glioblastoma) that produced high amounts of VEGF-A, and clones with the highest amounts of PlGF-2 were selected for implantation (84). Although there was no effect on the growth rate of the cell lines in culture, subcutaneous or orthotopic implantation of the PlGF-overexpressing cell lines into mice revealed an inhibition of tumor growth and a reduction in tumor-associated angiogenesis. One possible explanation for these results may be that in the presence of elevated PlGF, there was a reduction in the production of VEGF-A homodimers, with the majority of the VEGF forming the less active (in angiogenesis) heterodimers with PlGF (103).

The antagonism of VEGF-induced angiogenesis by production of PlGF within the same population of cells has also been demonstrated in a murine model of fibrosarcoma (38). VEGF-A–PlGF-1 heterodimers failed to activate VEGFR2-mediated signaling and failed to induce angiogenesis in vitro and in vivo. Although overexpression of human PlGF-1 in murine fibrosarcoma cells did not alter the growth rate of the cells in culture, forced production of PlGF-1 markedly reduced the rate of growth of tumors arising from subcutaneously implanted inoculates. In a spontaneous mouse tumor model for pancreatic cancer, introduction of PlGF-1 under the control of the insulin promoter resulted in slower-growing tumors with reduced tumor angiogenesis relative to the parent transgenic mouse. The inhibition of tumor growth and angiogenesis appeared to be due to the formation of VEGF-A–PlGF-1 heterodimers in the tumor cell population at the expense of the more angiogenic VEGF-A–VEGF-A homodimers. Thus, the net effect of PlGF on tumorigenesis, tumor progression, and tumor angiogenesis appears to be dependent on the tumor’s VEGF production, such that PlGF can have different effects according to its spatiotemporal relationship to VEGF production (Fig. 4).

PlGF as a therapeutic target for cancer

The interplay between PlGF and VEGF-A makes it difficult to propose a single clinically relevant strategy for targeting PlGF. Nevertheless, if information about the VEGF-A status of the tumor is available, it should be easier to decide whether increasing PlGF abundance or inhibiting PlGF activity is the most appropriate approach.

One agent of particular interest that targets tumor angiogenesis is the soluble decoy receptor aflibercept (VEGF Trap). This agent, which is currently under clinical evaluation, is a fusion protein incorporating the second Ig-like extracellular domain of VEGFR1 joined to the third Ig-like extracellular domain of VEGFR2, which is linked to the Fc portion of human IgG1 (104). The fused receptor-derived regions of aflibercept bind with high affinity to VEGF-A, PlGF, and VEGF-B proteins (105). The capability to simultaneously target these three proteins therapeutically is attractive and may prove to be a clinically effective strategy. Phase I dose escalation and phase II studies have shown that aflibercept is generally well tolerated and have provided some early evidence of efficacy. They have also suggested that this agent may be safely combined with standard chemotherapy regimens (106116). Aflibercept is currently undergoing clinical validation in four phase III trials, which include patients with prostate, non–small cell lung, colorectal, and pancreatic cancers.

Paradoxically, it may also be possible in the future to use PlGF itself to inhibit VEGF-A–induced angiogenesis and tumor growth (103). A genetically modified PlGF possessing a retention signal of the endoplasmic reticulum (ER) could be used to force the formation of heterodimers with VEGF-A in tumor cells. The ER-retained PlGF could consequently hold VEGF intracellularly and thereby block its function. For therapeutic development, viral or nonviral vectors carrying this version of PlGF would have to be delivered into tumor cells. Such an approach warrants further investigation.

Insight from Knockout Mice

Ligands

Knockout of only one allele of the Vegfa gene, encoding VEGF-A, is lethal in the mouse embryo (heterozygous lethality) with impaired angiogenesis and blood-island formation leading to developmental abnormalities (117, 118). Among all gene deletion studies, a haploinsufficiency phenotype resulting in obvious vascular abnormalities and heterozygous lethality in embryos has been observed only in Vegf and Delta-like 4 (Dll4, a Notch ligand) knockout mice (119). Like Vegfa, Dll4 is involved in vascular development (119, 120).

In contrast, several studies suggest that genetic deletion of either Vegfb or Plgf genes does not result in obvious impairment of the vascular system (36, 37, 61), with knockout mice essentially developing normally and living healthily during their entire adulthood. However, one study reported that Vegfb knockout mice have coronary artery abnormalities and smaller hearts, although this observation was not confirmed in a second study (37, 61).

Although Plgf knockout mice under normal physiological conditions do not exhibit obvious phenotypic changes, these animals show vascular defects under several pathological settings (36, 39, 40, 121). The response to VEGF-A was also impaired, including reduced angiogenesis and vascular permeability under ischemic insult in a hindlimb model and recruitment of monocytes and macrophages in a skin wound assay. Transplantation of wild-type bone marrow restored angiogenesis and collateral growth in this knockout model under ischemic conditions, which suggests that PlGF might contribute to vessel growth in adult animals through the mobilization of bone marrow–derived cells (40). Alternatively, it is possible that lack of PlGF leads to severe defects of bone marrow cells themselves. Deletion of Plgf in mice was also associated with a reduced rate of vascular leakage induced by skin wounding, allergens, and neurogenic inflammation (36). Although the underlying mechanisms of the PlGF-related pathological functions are not fully understood, they are probably associated with defective recruitment of inflammatory cells including monocytes and macrophages expressing VEGFR1.

Receptors

Consistent with the phenotypes of ligand deletion, knockout of the murine VEGFR2 (Flk1) gene leads to early embryonic lethality due to lack of vasculogenesis and angiogenesis, supporting the positive role of this receptor in the mediation of vessel growth (122, 123). In contrast, deletion of the murine VEGFR1 gene (Flt1) also results in early embryonic death, but this is due to uncontrollable growth of blood vessel endothelial cells (50), which suggests that VEGFR1 mediates negative regulation of embryonic angiogenesis. The tyrosine kinase activity of the receptor appears to be dispensable for this embryonic function of VEGFR1, because mice with a deletion of only the kinase domain of Flt1 (flt-1TK−/−) developed nearly normal vasculature, whereas VEGF-A–induced macrophage migration was impaired in adult animals (56). However, there may be a role for VEGFR1-mediated tyrosine kinase signaling in pathological angiogenesis, which was revealed in studies with murine flt-1TK−/− models of human disease processes, including choroidal neovascularization, rheumatoid arthritis, and metastasis (124127). In each case, the flt-1TK−/− mice showed decreased disease progression or symptoms. In addition, PlGF-expressing tumors showed retarded growth in flt-1TK−/− mice (124, 128).

Conclusions and Prospects

Production of PlGF is restricted to a few tissues in humans. In contrast, it is commonly expressed in tumor cells, and high-level expression may play a causal role in some malignancies or may modify the clinical characteristics of disease. The role of VEGF-B in promoting tumor angiogenesis remains poorly understood and warrants further investigation.

The spatiotemporal expression pattern of PlGF (and perhaps VEGF-B) in relation to VEGF-A appears to be one of the key determinants of the effect of PlGF on angiogenesis. When PlGF and VEGF-A are produced within a tumor cell population, PlGF may antagonize VEGF-A function through the formation of VEGF-A–PlGF heterodimers, leading to a concomitant reduction in the amount of the more angiogenically active VEGF-A–VEGF-A homodimers. Conversely, when PlGF (and possibly VEGF-B) and VEGF-A are produced by separate cell populations, PlGF may promote angiogenesis by enhancing VEGF-A signaling through VEGFR2.

The abundance of PlGF, VEGF-B, VEGF-A, VEGFR1, and VEGFR2, as well as the relative proportions of these molecules in tumors, might serve as predictive biomarkers of tumor response to agents that target PlGF or VEGF-A. During therapy that inhibits the activity of VEGF-A, the amount of PlGF in the circulation increases (72, 129, 130), which suggests that tumors might use PlGF together with other angiogenic factors to promote tumor angiogenesis. This compensatory mechanism might contribute to the development of resistance therapies that inhibit VEGF-A (Fig. 2). Thus, blocking VEGFR1-specific ligands in combination with other broad-spectrum inhibitors may offer a more effective approach for cancer therapy, as well as offering an alternative approach to targeting the receptor tyrosine kinases. An organizing principle for antiangiogenic cancer therapy, which consists of using several antiangiogenic agents targeting different pathways of angiogenic signaling pathways or different steps of the angiogenic process, has already been validated in experimental tumor models (131, 132). This is an exciting time of opportunities to improve the therapeutic efficacy of existing treatments by developing more agents that inhibit tumor angiogenesis.

Acknowledgments

I thank S. Lim for the artistic work. I regret that, owing to space limitations, I am unable to cover all aspects of the large amount of basic research on this topic or to refer to all of the primary literature; in many instances, I have cited only reviews. The author’s laboratory is supported by research grants from the Swedish Research Council, the Swedish Heart and Lung Foundation, the Swedish Cancer Foundation, the Karolinska Institute fund, the Karolinska Gender foundation, the Söderberg Foundation, European Union Integrated Projects of Angiotargeting contract 504743 (to Y.C.), and European VascuPlug Contract STRP 013811 (to Y.C.). The author is also a Chang Jiang scholar of the Chinese Ministry of Education.

References and Notes

View Abstract

Stay Connected to Science Signaling

Navigate This Article