The vascular endothelial growth factor receptor Flt-1 mediates biological activities. Implications for a functional role of placenta growth factor in monocyte activation and chemotaxis.

Two distinct receptors for vascular endothelial growth factor (VEGF), the tyrosine kinase receptors Flt-1 and Flk-1/KDR, have been described. In this study we show that monocytes, in contrast to endothelium, express only the VEGF receptor Flt-1, and that this receptor specifically binds also the VEGF homolog placenta growth factor (PlGF). Both VEGF and PlGF stimulate tissue factor production and chemotaxis in monocytes at equivalent doses. In contrast, endothelial cells expressing both the Flt-1 and the Flk-1/KDR receptors produce more tissue factor upon stimulation with VEGF than after stimulation with PlGF. Neutralizing antibodies to the KDR receptor reduce the VEGF-stimulated tissue factor induction in endothelial cells to levels obtained by stimulation with PlGF alone, but do not affect PlGF-induced tissue factor induction in endothelial cells nor the VEGF-dependent tissue factor production in monocytes. These findings strongly suggest Flt-1 as a functional receptor for VEGF and PlGF in monocytes and endothelial cells and identify this receptor as a mediator of monocyte recruitment and procoagulant activity.

Two distinct receptors for vascular endothelial growth factor (VEGF), the tyrosine kinase receptors Flt-1 and Flk-1/KDR, have been described. In this study we show that monocytes, in contrast to endothelium, express only the VEGF receptor Flt-1, and that this receptor specifically binds also the VEGF homolog placenta growth factor (PlGF). Both VEGF and PlGF stimulate tissue factor production and chemotaxis in monocytes at equivalent doses. In contrast, endothelial cells expressing both the Flt-1 and the Flk-1/KDR receptors produce more tissue factor upon stimulation with VEGF than after stimulation with PlGF. Neutralizing antibodies to the KDR receptor reduce the VEGF-stimulated tissue factor induction in endothelial cells to levels obtained by stimulation with PlGF alone, but do not affect PlGF-induced tissue factor induction in endothelial cells nor the VEGF-dependent tissue factor production in monocytes. These findings strongly suggest Flt-1 as a functional receptor for VEGF and PlGF in monocytes and endothelial cells and identify this receptor as a mediator of monocyte recruitment and procoagulant activity.
Vascular endothelial growth factor (VEGF), 1 also known as vascular permeability factor, was described as an inducer of angiogenesis in a variety of physiological and pathological processes including embryogenesis (1), corpus luteum formation (2), tumor growth (3), wound healing (4), and compensatory angiogenesis in the heart (5). In addition, VEGF induces vascular permeability in vivo (6) and exerts procoagulant activity via its ability to stimulate the production of the potent initiator of coagulation tissue factor (7) in endothelial cells and monocytes (8).
There are two known phosphotyrosine kinase receptors for VEGF: the fms-like tyrosine kinase Flt-1, and the fetal liver kinase, Flk-1 (derived from the mouse), or its human homolog KDR (kinase insert domain-containing receptor), with apparent K d values of 16 -114 pM for Flt-1 (9 -12) and 400-1000 pM for Flk-1/KDR (11)(12)(13)(14). The K d values derived from studies using transfected cell lines correspond closely to the K d values of the two binding sites which have been observed in endothelial cell types tested in vitro (11,15,16). So far, biological activities observed in both in vitro and in vivo studies appear to be mediated exclusively by the Flk-1/KDR receptor. Transfection of either the human Flt-1 or the KDR receptor cDNA in porcine aortic endothelial (PAE) cells, which do not express either VEGF receptor, demonstrated that, at least in PAE cells, the ability of VEGF to stimulate chemotaxis and proliferation occurs in KDR-transfected (KDR-PAE) but not in Flt-transfected (Flt-PAE) PAE cells (11). In addition, using a dominant negative approach, the Flk-1 receptor was shown to be essential for tumor angiogenesis in a glioblastoma model in vivo. (17) In addition to its endothelial cell-specific activities, VEGF also attracts peripheral blood monocytes, raises their intracellular Ca 2ϩ levels (18), and induces tissue factor production (8). In contrast to the expression of both VEGF receptors in endothelial cells, only one specific binding site has been found on cells of the monocyte/macrophage lineage. Upon activation with VEGF, a high molecular weight band corresponding to a tyrosine-phosphorylated protein can be observed following immune precipitation with an anti-phosphotyrosine antibody (18).
Placenta growth factor (PlGF) has been described recently as a secreted growth factor with strong homology to VEGF based on amino acid and cDNA sequences. PlGF is expressed in human umbilical vein endothelial (HUVE) cells and placenta (19,20). PlGF is a very weak stimulator of endothelial chemotaxis and proliferation (20), a finding which can be explained by the hypothesis that PlGF displays lower affinity to one or both of the two VEGF receptors. Indeed, when binding competition studies are performed with extracellular domains from either Flk-1/KDR or Flt-1, PlGF appears to be able to bind to Flt-1 but not to KDR/Flk-1 (12,21).
In this study we compared the human heparin binding isoform of PlGF, also known as PlGF-2 (20), with the corresponding human VEGF 165 isoform. Using PAE cells, stably transfected with either KDR or human Flt-1, we identified Flt-1 as a specific PlGF receptor. In addition, we demonstrate by assessing tissue factor production and chemotaxis that Flt-1-expressing human monocytes react similarly to PlGF as to VEGF, while Flt-1 and KDR coexpressing human endothelial cells are primarily responsive to VEGF. Furthermore, neutralizing polyclonal antibodies to the KDR receptor inhibit VEGF-but not PlGF-mediated induction of endothelial tissue factor, whereas in monocytes these antibodies do not affect VEGF-mediated tissue factor induction. These data identify Flt-1 as a functional receptor for both VEGF and PlGF in cells of the endothelial and monocyte/macrophage lineage.

EXPERIMENTAL PROCEDURES
Materials-Laboratory reagents not listed otherwise were purchased from Sigma (Mü nchen, FRG). Media and other cell culture reagents were obtained from Life Technologies, Inc. (Eggenstein, FRG). Heparin-, concanavalin A-and phenyl-Sepharose and Superdex S200 were obtained from Pharmacia (Freiburg, FRG). Human umbilical cords were kindly donated from hospitals in the "Wetterau." Citrated pooled plasma was obtained from volunteers.
Cell Culture and Assays-HUVE cells were prepared by the method of Jaffe et al. (22) as modified by Thornton et al. (23). HUVE cells were cultured in medium 199 supplemented with 10 mM Hepes (pH ϭ 7.4), 10% fetal calf serum, 10% human serum (Sigma, Mü nchen, FRG), 100 g/ml endothelial cell growth factor (ccpro, Neustadt, FRG), heparin (20 g/ml), 100 units/ml penicillin, 100 g/ml streptomycin, 2 mM glutamine, 2.5 g/ml fungizone. Experiments were carried out within 48 h of the cells achieving confluency. Expression of tissue factor in endothelial cells was assessed by incubating cultures with purified recombinant VEGF or PlGF in medium 199 containing 10 mM Hepes (pH ϭ 7.4), 10% fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, in the presence/absence of antibodies. For studies with neutralizing antibodies against the VEGF receptor KDR (r-212), cells were submitted to a 15-min incubation period before addition of the cytokines. Cells were further incubated for 6 h at 37°C. Assays were carried out with whole cells obtained in suspension following scraping, and tissue factor activity equivalents were determined as described previously (8). Briefly, endothelial cells were removed from the dish by scraping with a rubber policeman. 100 l of these resuspended cells were mixed with 100 l of citrated plasma, and clotting times were measured after recalcification with 100 l of a 25 mM CaCl 2 solution. Tissue factor equivalents were determined by using a standard curve of purified human tissue factor. In certain experiments, cells were incubated for 15 min at 37°C with a monoclonal tissue factor neutralizing antibody (Loxo, Dossenheim, FRG) prior to the clotting assay.
Peripheral human blood monocytes were purified from buffy coats employing gradient centrifugation over Ficoll (histopaque 1077) and subsequent adhesion to tissue culture plastic as described previously (8). Chemotaxis of monocytes was investigated using the method of Quinn et al. (24) as modified recently (25). Briefly, cells were placed in the upper chamber, and a test substance was placed in the lower chamber. Chemotactic assays were performed after 3 h of incubation, and cells in at least 4 power fields were counted for each condition assessed.
RT-PCR analysis was performed essentially as described (27). In short, synthesis of cDNA templates was carried out using 2 g of total RNA. The cDNA contained in 1-l aliquots of the reverse transcription was amplified using a Perkin Elmer 9600 thermal cycler and Amplitaq DNA polymerase (Perkin Elmer Cetus, Weiterstadt, FRG). The oligonucleotide primers used for the amplification of Flt-1 cDNA were 5Ј-GATGTCGACG GTATAAATAC ACATGTGCTT-3Ј and 5Ј-CTATG-GAAGA TCTGATTTCTTACAGT-3Ј. The primers used for amplification of Flk-1/KDR cDNA were 5Ј-CAGGGATCCT GAAATTACT-3Ј and 5Ј-CTGTCGACGT TTGAGAACCT CAC-3Ј. The cDNA was amplified in 35 cycles, each cycle consisting of 30 s at 94°C, 30 s at 50°C, and 60 s at 72°C. Aliquots of 15 l were separated in a 1.2% agarose gel and stained with ethidium bromide.
Protein Expression and Purification-Recombinant human PlGF (PlGF-2) was expressed using the baculovirus system and purified as recently described (20). Briefly, supernatants from SF158 cells infected with baculovirus, containing the cDNA for PlGF, were adjusted to 0.4 M NaCl, sterile-filtered, and applied to heparin-Sepharose. After washing the column with 0.4 M NaCl, PlGF was eluted with 2 M salt. The fractions containing PlGF, as determined with a Western dot-blot anal-ysis employing a specific antiserum made against an N-terminal 20mer peptide, were eluted. A second round of purification was performed by using a TSK-Heparin column (Toso Haas). PlGF was eluted with an ascending salt gradient from 0.4 to 2 M NaCl. Purity of PlGF was checked by SDS-gel electrophoresis and Western blot analysis with specific rabbit serum against PlGF (20) and quantified with the BCA assay (Pierce).
VEGF Labeling and Binding Studies-Human recombinant VEGF was expressed in baculovirus and purified as described recently (28). Labeling was performed using the chloramine T method as modified for VEGF (15), and specific activities of 3-4 ϫ 10 5 cpm/ng were achieved. For binding competition studies, 125 I-VEGF was mixed with various concentrations of either unlabeled VEGF or PlGF in RPMI containing 20 mM Hepes and 0.1% gelatin. Medium from cells was removed and replaced by the samples with 125 I-VEGF with or without the unlabeled VEGF or PlGF. After a 3-h incubation on ice, cells were washed 4 times with RPMI medium containing 0.1% gelatin. Bound 125 I-VEGF was removed from cells by lysis with 0.1% SDS, and counts were measured in a ␥ counter (Packard, Frankfurt, FRG).
Preparation of Antibodies against the Soluble KDR Receptor-KDR was cloned from cDNA from a human placenta cDNA library. A 2.3kilobase BglII-fragment coding for the first seven IgG-like extracellular domains was transferred into the vector pBacPAK9 (Clontech, Heidelberg, FRG) for baculovirus expression. After transfection into Sf 158 insect cells clones were obtained by screening for the presence of KDR message. These clones were used for virus amplification and overexpression of soluble KDR (sKDR). Insect cell supernatants were concentrated by ultrafiltration, dialyzed against 10 mM Tris (pH 8.0), 150 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 and passed 4 times through a 5-ml concanavalin A-Sepharose column. Bound protein was eluted with the same buffer containing 0.5 M methyl-D-mannopyranoside and 0.5 M methy-D-glucopyranoside. The fractions containing sKDR, as checked by SDS-gel electrophoresis, were adjusted to 0.8 M ammonium sulfate and applied to a 5-ml phenyl-Sepharose HP column. Proteins were eluted with a decreasing ammonium sulfate gradient. Final purification was obtained by gel filtration on a Superdex S200 column. Purity was greater than 95% as estimated by SDS-polyacrylamide gel electrophoresis. The identity of the purified protein with an apparent molecular mass of 110 kDa with sKDR was shown by N-terminal sequence analysis using Edman degradation chemistry on an automatic gas Sequencer (Applied Biosystems, Weiterstadt, FRG). 0.5 mg of purified protein were injected into the rabbit, after an additional two boosts with each 0.2 mg of protein, the animal was bled. The serum revealed a half-maximal titer at a 1:7500 dilution in an ELISA for sKDR.

PlGF-VEGF Binding Studies with Endothelial Cells Expressing Either One of the VEGF Receptors and Isolated Human
Monocytes-In radioligand binding studies, PlGF has previously been shown to bind to a soluble form of the Flt-1 receptor but not of the KDR receptor (12,21). Using a cellular model, we studied the competitive binding of PlGF or VEGF with 125 I-VEGF in PAE cells which were transfected with either the KDR or the Flt-1 receptor (11). PlGF was found to compete with 125 I-VEGF for binding to Flt-PAE cells at similar concentrations as that observed with unlabeled VEGF (Fig. 1A). In contrast, no competition of PlGF with 125 I-VEGF occurred at any of the concentrations tested in KDR-PAE cells, although unlabeled VEGF showed the expected concentration-dependent competition (Fig. 1B). When primary endothelial cells were used, which endogenously express both VEGF receptors (Flk-1/KDR and the Flt-1), PlGF only competed up to a maximum of 50% of that achieved with unlabeled VEGF (data not shown). Employing isolated human monocytes, PlGF competed for binding with 125 I-VEGF in a similar dose-response as noniodinated VEGF. Similar to the results obtained with Flt-PAE cells, VEGF competed at a 2-4-fold lower dose than PlGF for 125 I-VEGF binding to monocytes (Fig. 1C). Expression of Flt-1 mRNA in Peripheral Human Blood Monocytes-Previous studies with iodinated 125 I-labeled VEGF have revealed one specific binding site on monocytes (18). The results described above suggest that Flt-1 is the monocyte VEGF receptor. To further investigate this hypothesis, we analyzed monocyte mRNA for the presence of Flt-1 transcripts. RNA isolated from monocytes and HUVE cells share a transcript of 7.5 kilobases in size ( Fig. 2A) when probed with a specific Flt-1 cDNA as described (3). The presence of transcripts coding for the Flt-1 receptor in monocytes was also confirmed by RT-PCR analysis with specific primers for human Flt-1 cDNA. Using the same PCR approach, transcripts for KDR could not be detected in cDNA derived from the same monocyte preparation (Fig. 2B).
Comparison of PlGF-and VEGF-mediated Tissue Factor Induction on Monocytes and Endothelial Cells-VEGF has been shown to induce procoagulant activity on the surface of human monocytes and HUVE cells by causing the de novo synthesis of tissue factor (8). As assessed by binding studies, HUVE cells express lower numbers of Flt-1 than KDR receptors (11,16), in contrast to monocytes which exclusively express Flt-1 receptors. Therefore, we decided to compare the actions of PlGF and VEGF, in both monocytes and HUVE cells, in terms of their potential to induce tissue factor production. When isolated human monocytes were exposed to increasing concentrations of human recombinant PlGF and VEGF, an equivalent, dose-dependent induction of tissue factor was observed with both agents. The half-maximal response occurred at a concentration of 0.3 nM with VEGF and 0.5 nM with PlGF. Incubation of the cells with a monoclonal antibody specific to human tissue factor prior to the assessment of tissue factor abolished all of the tissue factor-inducing activity of PlGF and VEGF (Fig. 3A). When HUVE cells were incubated with PlGF, a dose-dependent induction of tissue factor, with a half-maximal response at 0.5 nM, was observed. VEGF-stimulated tissue factor expression appeared to be half-maximal at 1 nM, and maximal tissue factor production by VEGF is induced to a greater extent as compared to similar concentrations of PlGF (Fig. 3B).
Chemotactic Response of Peripheral Human Blood Monocytes to PlGF and VEGF-Using a modified Boyden chamber assay, VEGF has been shown to attract monocytes in a dose-dependent manner. Checkerboard analysis revealed that the response of monocytes to VEGF was a result of chemotaxis and not of chemokinesis (8). Due to the selective affinity of PlGF to Flt-PAE cells and our finding that Flt-1, but not KDR, is expressed on freshly isolated human peripheral blood monocytes, we hypothesized that PlGF may specifically induce monocyte migration. When recombinant PlGF was added to the lower well of a modified Boyden chamber, the number of migrating monocytes was similar to that induced by the addition of VEGF as a chemotactic stimulus (Fig. 4). Migration of monocytes was observed with a PlGF concentration of 100 pM, the response was maximal at 300 pM. The extent of chemotaxis in response to PlGF and VEGF was equivalent to that observed with the potent formylated chemotactic tripeptide fMet-Leu-Phe (29) which was used as a positive control. The chemotactic activity of PlGF was abolished by heating the peptide at 95°C for 10 min, prior to the addition to the lower well of the Boyden chamber (Fig. 4).
Inhibition Studies with a Neutralizing Serum against KDR-In order to further distinguish between the two VEGF receptors, a polyclonal serum against the KDR receptor was developed. Rabbits were immunized with a 110-kDa protein coding for the extracellular seven IgG-like domains of the KDR receptor. In binding competition assays, serum (r-212) from one of these rabbits inhibited binding of 125 I-VEGF to KDR-PAE FIG. 2. A, detection of Flt-1 transcripts in peripheral blood monocytes by Northern blot analysis. Human monocytes were isolated and RNA was extracted as described under "Experimental Procedures." Northern blots with RNA from monocytes (Monocytes) and from HUVE cells (HUVEC) were probed with radiolabeled Flt-1 cDNA as described in the text. The blot was visualized by using a PhosphorImager. B, RT-PCR analysis of human monocyte RNA, revealing the presence of Flt-1-, but not of Flk-1/KDR-specific transcripts in monocytes. Total RNA from human monocytes was prepared, and RT-PCR analysis was performed using oligonucleotide primers specific for human Flt-1 (lane 1) and for KDR (lane 3). A specific 1080-bp Flt-1 cDNA, but no Flk-1 cDNA, was amplified from the monocyte cDNA preparation. Cloned full-length cDNAs of human Flt-1 (lane 2) and KDR (lane 4) were used as templates in positive control amplification reactions.

FIG. 1. Binding competition studies of VEGF and PlGF on Flt-PAE cells (A), KDR-PAE cells (B), and monocytes/macrophages (C).
125 I-VEGF was incubated together with increasing concentrations of non radioactive VEGF (ⅪOOⅪ) or PlGF (Ç---Ç) for 3 h at 0°C. After incubation, cells were removed and bound radioactivity was determined as described in the text. The mean of duplicates is shown. The deviation was smaller than 10% for all values of this figure. cells (Fig. 5). A 40-fold dilution neutralized binding of labeled VEGF to the same extent as a 50-fold excess of nonradioactive VEGF, whereas the same dilution of the preimmune serum did not affect binding of 125 I-VEGF to KDR-PAE cells. In contrast, 125 I-VEGF binding to human monocytes (which do not express KDR) was not effected by a 40-fold dilution of serum r-212 (Fig.  5).
Serum r-212 was assessed for its ability to inhibit the production of tissue factor induced by either VEGF or PlGF in HUVE cells. Although the PlGF-induced tissue factor production is much less than the one induced by VEGF at the same concentration, it is clearly above the values obtained with media alone. KDR neutralizing serum r-212 had no effect on this PlGF-mediated tissue factor induction, which is in concordance with the assumption that the effect of PlGF is solely dependent on Flt-1. However, when a 2 nM VEGF concentration was applied in the presence of serum r-212, tissue factor production was reduced to values obtained with a 2 nM concentration of PlGF alone. A control rabbit serum did not affect the tissue factor production achieved with 2 nM VEGF or PlGF (Fig. 6).

FIG. 3. Induction of monocyte (A) and endothelial (B) tissue factor production by VEGF and PlGF.
Cells, prepared and cultivated as described in the text, were incubated together with increasing concentrations of either VEGF (ⅪOOⅪ) or PlGF (Ç---Ç) for 6 h at 37°C. Cells were scraped from the plate, and tissue factor was assessed as described in the text. Results show mean Ϯ S.E. of at least 3 independent experiments. Errors smaller than symbol size are not shown. A, dose dependence of VEGF (solid line) or PlGF (dashed line) induced monocyte tissue factor production. In one experiment, PlGF (1 nM) was heated for 10 min at 95°C prior to the addition to monocytes (*). For one concentration of PlGF, cells were incubated with a neutralizing antibody to tissue factor for 15 min (after the usual incubation time with factor), and tissue factor was determined (Q). B, dose dependence of VEGF-(solid line) or PlGF-(dashed line) induced endothelial tissue factor production. HUVE cells were incubated with increasing concentrations of factors, and tissue factor was determined as described in the text.

FIG. 4. Induction of monocyte migration by PlGF and VEGF.
Peripheral blood monocytes, prepared as described in the text, were added to the upper compartment of a modified Boyden chamber. The indicated concentrations of VEGF, PlGF, heat-treated 1 nM solution of PlGF (1 nM heat), medium alone (medium), or the chemotactic tripeptide fMet-Leu-Phe (10 Ϫ7 M) were added to the lower compartments, the chambers were incubated for 3 h at 37°C, and then migrated cells were stained and counted as described in the text. Results show mean Ϯ S.E. of 4 independent experiments. In each experiment, cells were counted in at least 4 representative high power fields.

FIG. 5. Effects of polyclonal antibodies to the KDR receptor (serum r-212) on binding of 125 I-labeled VEGF to KDR-PAE cells and monocytes. KDR-PAE cells (open bars) and monocytes (hatched bars)
were incubated with 125 I-VEGF in the presence or absence (none) of 50-fold excess of unlabeled VEGF (50-fold VEGF), serum r-212 (r-212) directed against KDR and preimmune rabbit serum (Co-serum) for 3 h at 0°C. After the incubation, cells were detached and bound radioactivity was detected as described in the text. The mean of at least duplicates (deviation was smaller than 10%) is shown for all values of this figure.
According to the observation that monocytes express the Flt-1 but not the Flk-1/KDR receptor, serum r-212 should not affect VEGF-mediated tissue factor induction in monocytes. Indeed, when serum r-212 was added to monocytes in the presence of a 1 nM concentration of VEGF, no reduction in tissue factor production was observed as compared to treatment with the same dose of VEGF in the absence of serum r-212 (Fig. 6).

DISCUSSION
The close sequence homology between PlGF and VEGF (53% based on amino acid comparison) led to the suggestion that these polypeptides may mediate similar biological activities. However, when PlGF was compared to VEGF in terms of their abilities to stimulate endothelial proliferation, PlGF was found to be much less active than VEGF (20,30). This may be due to the preferential binding of PlGF to Flt-1 in comparison to Flk-1/KDR, as shown in studies using soluble VEGF receptors (12,21). In this study, we focussed on two questions. Firstly, is Flt-1 the preferential receptor for PlGF in living cells, and, secondly, can Flt-1 binding of VEGF or PlGF give rise to biological functions. To address the first question we used genetically engineered porcine aortic endothelial cells expressing either the KDR or Flt-1 receptor exclusively. The second question we investigated by the use of neutralizing antibodies to the KDR receptor and by testing two biological responses of monocytes, which express only the Flt-1 receptor, to VEGF or PlGF.
When PlGF binding to cellularly expressed VEGF receptors was tested by competition studies with iodinated VEGF, we found selective PlGF binding to the PAE cells transfected with the Flt-1 receptor but not to those transfected with the KDR receptor. In isolated human peripheral blood monocytes, a dosedependent binding competition for 125 I-VEGF (similar to that described in Flt-PAE) was observed with both PlGF and VEGF. This strongly supports the assumption that monocytes express only one VEGF receptor, which is Flt-1, and that PlGF is also a ligand for this receptor. Further support for this hypothesis came from expression studies probing mRNA isolated from monocytes or HUVE cells for either KDR or Flt-1 transcripts. Again, only expression of Flt-1 could be detected in freshly isolated human blood monocytes. These findings concur with those of an earlier study demonstrating only one specific binding site on monocytes/macrophages for VEGF (18), in contrast to binding studies on endothelial cells which revealed two distinct receptor binding sites (11,15,16). In addition, crosslinking studies with 125 I-VEGF on isolated human monocytes clearly demonstrated only one specific band of a molecular mass greater than 200 kDa (18). These data can only be reconciled by a model in which the receptor-mediating biological activity in monocytes is promiscuous, i.e. it is activated by both VEGF and PlGF, which is the situation in monocytes for the Flt-1 receptor.
A more complex situation, however, emerges in respect to HUVE cells. By means of cross-linking studies with 125 I-VEGF, two novel receptors for VEGF 165 were identified in a breast cancer cell line by Soker et al. (31). These two novel receptors, which are also found in HUVE cells, can be distinguished from Flk-1/KDR and Flt-1 on the basis of cross-linking and immunoprecipitation experiments. The relative molecular masses of the novel receptor complexes are 165 and 175 kDa in comparison with molecular masses greater than 200 kDa for the complexed forms of the Flk-1/KDR and Flt-1 receptors (31). The binding affinities of these receptors for VEGF are similar to the binding affinity of VEGF for Flk-1/KDR (1963), making the distinction of these receptor types difficult by means of binding studies. While the two novel receptors may mediate an as yet undefined biological activity, it is highly unlikely that they contribute to tissue factor induction by PlGF in HUVE cells and monocytes or chemotaxis in monocytes alone as they do not bind soluble PlGF (31).
Although it has been shown that Flt-1-transfected fibroblasts respond to VEGF in terms of intracellular signaling, VEGF-dependent mitogenic activity was not observed (32). Similarly, VEGF-induced [ 3 H]thymidine incorporation and chemotaxis could not be shown in the endothelial Flt-PAE transfectants in comparison with KDR-PAE cells (11). The Flt-1 receptor, however, appeared to be active, since intracellular signaling was observed with Flt-PAE cells after stimulation with VEGF. One possible explanation for this phenomenon was that PAE cells lack an important intracellular pathway required for Flt-1 signaling. One other possibility is that Flt-1 functions as a decoy receptor, analogous to the interleukin-1 receptor type I as suggested recently (21). We believe this report is the first which confirms that the Flt-1 receptor mediates signaling leading to distinct biological responses (which are tissue factor induction and chemotaxis in monocytes) and does not act as a decoy receptor.
In order to exclude the possibility that Flt-1 can only transduce signaling in monocytes, we determined whether PlGF was also able to activate endothelial cells known to express Flt-1. After stimulation of HUVE cells with PlGF, a dose-dependent tissue factor induction could be observed, although this response was less prominent when compared to VEGF. The halfmaximal induction of tissue factor production occurred at a 2-fold lower concentration with PlGF as compared with VEGF. This is consistent with the lower K d values described for Flt-1 in comparison with KDR (11). In contrast to monocytes (which can be stimulated to a similar maximal production of tissue factor by both PlGF and VEGF), maximal tissue factor produc- without or with anti-KDR serum r-212 (r-212) or normal rabbit serum (Co-serum). Peripheral blood monocytes (hatched bars) were also incubated together with 1 nM VEGF in the absence (1 nM VEGF) or presence of anti-KDR-serum r-212 (r-212) or with no addition (medium). The cells were incubated for 6 h at 37°C, scraped, and tissue factor was determined as described in the text. The mean Ϯ S.E. of at least 3 independent experiments are shown unless they are too small to be graphically placed. tion with VEGF in HUVE cells is greater than with PlGF. There are at least two possible explanations for the difference in tissue factor production induced by equivalent concentrations of VEGF and PlGF in HUVE cells. The KDR receptor could either evoke a stronger biological response or may be present in higher amounts than Flt-1. The latter possibility is favored by binding studies which show significantly more binding sites on HUVE cells for the receptor with the lower affinity for 125 I-VEGF (16), that is, the KDR receptor (11).
The hypothesis that Flt-1 is also a functional receptor in HUVE cells is further supported by studies using neutralizing antibodies (serum r-212) to the KDR receptor. These antibodies reduce tissue factor induction by VEGF to the same values as observed with PlGF alone. In contrast, this serum cannot reduce endothelial tissue factor production by PlGF, which supports the hypothesis that PlGF-mediated tissue factor production is dependent on Flt-1 signaling. This signaling may be expected to occur upon Flt-1 homodimerization, since the effect of PlGF is independent of the anti-KDR serum. However, these results cannot exclude the possibility that VEGF also employs KDR-Flt-1 receptor heterodimerization, as the antibodies should not distinguish between KDR homodimers and KDR-Flt-1 heterodimers.
PlGF has been hypothesized to play a role in placental development and angiogenesis (19). PlGF and Flt-1 expression, as determined by in situ hybridization, co-localizes with the spongiotrophoblast region of the placenta, whereas VEGF and Flk-1/KDR are coexpressed predominately at the embryonic sites. 2 It has been reported that Flt-1 "knockout" mouse embryos lack functional vascular tubes, suggesting that Flt-1 is important for endothelium lumenal differentiation or interactions of endothelium with the extracellular matrix (33). Whether this defect in Flt-1 knockout mice is linked to endothelial functions or to enzymatic or other functions of monocytes/macrophages is unknown. Monocytes/macrophages have been proposed as critical players in the process of angiogenesis and wound healing (34). The strong placental expression of PlGF could contribute to the increased demand for angiogenesis in the growing placenta which may be partially mediated by chemoattraction of peripheral blood monocytes.
We show here a new feature of PlGF, the induction of procoagulant tissue factor production in endothelial cells and monocytes. Expression of tissue factor in the endothelium of tumor vessels and within the tumor cells correlates with the switch to the malignant and angiogenic phenotype of breast carcinomas (35). Tissue factor expression can also regulate VEGF expression as shown by tissue factor antisense transfection studies. Tumors derived from tissue factor antisense expressing cells exhibit reduced angiogenesis (36). In addition, tissue factor knockouts in mice reveal an impaired pattern of extraembryonic angiogenesis during early embryogenesis. 3 Furthermore, tissue factor production by PlGF-and/or VEGF-stimulated monocytes/macrophages may contribute to a common pathological complication of pregnancy, the increased risk of thrombosis. Further studies should reveal whether production and secretion of PlGF in placenta con-tributes to the consistently observed placental thrombosis during pregnancy.