Placenta growth factor stimulates MAP kinase and ... - Nature

Oncogene (1998) 16, 359 ± 367  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00

Placenta growth factor stimulates MAP kinase and mitogenicity but not phospholipase C-g and migration of endothelial cells expressing Flt 1 Eva Landgren1, Petter Schiller1, Yihai Cao2 and Lena Claesson-Welsh1 1

Dept. of Medical and Physiol. Chemistry Biomedical Center, Box 575, S-751 23 Uppsala, Sweden, and 2Department of Cell and Molecular Biology, Karolinska Institute, Stockholm, Sweden

Vascular endothelial growth factor (VEGF) and placenta growth factor (PlGF) are structurally related growth factors for endothelial cells. VEGF binds to the related receptor tyrosine kinases Flt 1 and KDR/Flk 1 with high anity, whereas PlGF binds only to Flt 1. Ligandstimulated KDR is known to transduce signals for cellular activity such as proliferation and migration, whereas weak or no responses have been recorded for Flt 1. We examined VEGF and PlGF for their capacity to stimulate signal transduction in porcine aortic endothelial cells expressing Flt 1 or KDR. VEGF had essentially no eect on Flt 1 expressing cells, but induced DNA synthesis and migration of KDR expressing cells. PlGF on the other hand induced DNA synthesis but not migration of the Flt 1 cells. In agreement, MAP kinase, examined as a marker for DNA synthesis, was activated both by VEGF-stimulation of the KDR cells and by PlGF-stimulation of the Flt 1 cells. In contrast, phospholipase C-g (PLC-g), was tyrosine phosphorylated only in VEGF stimulated KDR cells, and not in the PlGF-stimulated Flt 1 cells, which is in agreement with a role for PLC-g in cellular migration. We furthermore examined induction of protein levels of plasminogen activator (PA), which was evident in the PlGF-stimulated Flt 1 cells, but not in the VEGF-stimulated KDR cells. These data show that Flt 1 is able to mediate an array of biological signals when appropriately stimulated and that the pattern of responses of PlGF-stimulation of Flt 1 is distinct from the pattern of responses to VEGFstimulation of KDR. Keywords: VEGF; PLGF; endothelial cells; migration; proliferation

Introduction Vascular endothelial growth factor (VEGF) is a potent mitogen for endothelial cells. In addition to its growth promoting eect, VEGF induces increased permeability of vascular endothelial cells and stimulates angiogenesis (Neufeld et al., 1994). The expression of VEGF is upregulated by hypoxia and VEGF plays a role in many pathological processes such as tumour vascularization and proliferative retinopathy (Pierce et al., 1995; Plate et al., 1994; Shweiki et al., 1992). VEGF exerts its activity by binding to and activating cell surface receptor tyrosine kinases. Two related but

Correspondence: L Claesson-Welsh Received 12 June 1997; revised 2 September 1997; accepted 3 September 1997

distinct receptors for VEGF have been identi®ed, Flt 1 (VEGFR-1) (de Vries et al., 1992) and KDR/Flk 1 (VEGFR-2) (Terman et al., 1992). Both receptor types are expressed in endothelial cells, but expression of KDR has also been identi®ed in pancreatic duct precursor cells (OÈberg et al., 1994), and expression of Flt 1 has been demonstrated in monocytes (Clauss et al., 1996). The receptors have a similar structural organisation, with their extracellular parts composed of seven immunoglobulin-like loops and their intracellular tyrosine kinase domains interrupted by kinase inserts of about seventy amino acid residues. These VEGF receptors form a sub-family of receptor tyrosine kinases which also includes Flt 4 (VEGFR-3), which is expressed on lymphatic endothelium (Kaipainen et al., 1995). VEGF is the prototype for a number of recently discovered, structurally related growth factors which have been shown to, or are believed to, have important roles in the endothelial compartment. Thus, VEGF (VEGF-A), VEGF-B, VEGF-C and placenta growth factor (P1GF) show about 30 ± 53% similarity in their primary sequences. In particular, eight cysteine residues are found at conserved positions in these molecules; a structural organisation also found in the plateletderived growth factor. The receptor for VEGF-B has not yet been identi®ed, but VEGF-C binds to KDR/ Flk1 and Flt 4 (Joukov et al., 1996). PlGF binds only to Flt 1 (Park et al., 1994). Ligand-binding to receptors with tyrosine kinase activity typically induces oligomerization and autophosphorylation of the receptor molecules. Phosphorylation of receptor kinases is known to regulate their intrinsic kinase activity and to present binding sites for signal transduction molecules. Thereby a signal transduction cascade is initiated, which eventually results in the generation of speci®c cellular responses. Based on their structure, it is reasonable to assume that both KDR and Flt 1 are activated in agreement with the general scheme for receptor tyrosine kinases. KDR has been shown to autophosphorylate in response to VEGF (Millauer et al., 1993; Terman et al., 1992) and four autophosphorylation sites (Y951, Y996, Y1054 and Y1059) have been identi®ed in vitro, in the KDR intracellular domain (Dougher-Vermazen et al., 1994). In contrast, VEGF stimulation leads only to very weak Flt 1 autophosphorylation (de Vries et al., 1992; Waltenberger et al., 1994). A number of src homology 2 (SH2) domain-containing substrates, such as phospholipase C-g (PLC-g), Ras GTPase activating protein (RasGAP), phosphatidyl inositol 3-kinase (PI3kinase) and NcK have been implicated in VEGFstimulated signalling in endothelial cells (Guo et al., 1995). It has, however, been dicult to show which

Different effects of PIGF and VEGF E Landgren et al

360

receptor these substrates bind to, since most cells examined express more than one type of VEGF receptor. VEGF-induced phosphorylation of PLC-g and GAP has been detected in sinusoidal endothelial cells and in Flt 1 transfected NIH3T3 ®broblasts (Seetharam et al., 1995). The regulatory p85 subunit of PI3-kinase has been shown to bind to Flt 1 using the yeast two-hybrid system (Cunningham et al., 1995). The biological responses to VEGF stimulation appear to be dierent in KDR and Flt 1 expressing cells. While KDR is well accepted as a mediator of biological activity (Millauer et al., 1993; Quinn et al., 1993), the function of Flt 1 has been uncertain. Activation of KDR expressed in porcine aortic endothelial cells (PAE/KDR) by VEGF, leads to proliferation and migration, whereas PAE/Flt 1 cells are unable to undergo proliferation and chemotaxis in response to VEGF (Waltenberger et al., 1994). Furthermore, NIH3T3 cells expressing Flt 1 fail to undergo proliferation in response to VEGF treatment (Seetharam et al., 1995). Recently however, Flt 1 endogenously expressed in monocytes was reported to mediate induction of tissue factor production and chemotaxis in monocytes upon VEGF- and PlGFstimulation (Clauss et al., 1996). In addition, PlGF has been shown to have a weak growth stimulatory eect on human umbilical vein endothelial (HUVE) cells (Sawano et al., 1996). In order to further elucidate the role of Flt 1 in endothelial cells, we have studied PAE cells expressing Flt 1 or KDR. We have examined the capacity of the two cell types to respond to PlGF or VEGF in a number of biological assays, and in the activation of signal transduction molecules. Our data show that PlGF stimulation activates Flt 1 more eciently than VEGF stimulation, and that there are distinct patterns of cellular responses to Flt 1 and KDR activation, respectively.

Results Expression of VEGF receptors in porcine aortic endothelial cells Porcine aortic endothelial cells were isolated and allowed to undergo spontaneous transformation (Miyazono et al., 1987). The resulting cell line was transfected with cDNA encoding either of the two VEGF receptors, Flt 1 and KDR (Waltenberger et al., 1994). To characterise expression of the receptors, we ®rst performed Northern blot analysis. RNA was isolated from PAE/Flt 1 and PAE/KDR, eletrophoresed under denaturing conditions, transferred to a ®lter and hybridised with 32P-labelled probes corresponding to Flt 1 or KDR (Figure 1). When using Flt 1 cDNA as a probe, a signal was obtained, which showed expression in the Flt 1 transfected cells, as well as a low level of endogenous expression in the KDR cells. In both cases, two bands were seen, one of about 4.5 kbp and one of 2.6 kbp. In the KDR transfected PAE cells, a KDR-hybridising band of 5.5 kbp was evident. We did not detect any KDR signal in the Flt 1 transfected cells, even after prolonged exposure of the blot. We conclude that the transfected PAE cells contained increased levels of Flt 1 and KDR mRNA, respectively. In addition,

Probe: PAE cell:

Flt 1 Flt 1 KDR

KDR Flt 1 KDR

EtBr Flt 1 KDR

—

—

—

—

—

—

Figure 1 Flt 1 and KDR mRNA levels in transfected porcine aortic endothelial cells. RNA from PAE/Flt 1 and PAE/KDR cells was prepared and analysed by Northern blotting (20 mg RNA/lane) as described under Materials and methods. Hybridisation was performed with 32P-labelled Flt 1 (left panel) and KDR (middle panel) cDNA probes. Arrows indicate migration positions of the 4.5 and 2.6 kbp Flt 1-reactive transcripts (left panel) and the 5.5 kbp KDR-reactive transcript (middle panel), respectively. Ethidium bromide (EtBr) staining of the agarose gel shows equal loading of RNA and indicates the migration positions of 28S and 18S ribosomal RNA (right panel)

there appears to be a small level of endogenous Flt 1 mRNA, but not KDR mRNA, in the PAE cells. Scatchard analyses of receptor binding of VEGF to the PAE cells expressing Flt 1 and KDR was previously performed as described (Waltenberger et al., 1994). High anity binding sites for VEGF were found on both cell lines, and the numbers of receptors per cell were estimated to 506103 (PAE/Flt 1) and 1506103 (PAE/KDR). Tyrosine phosphorylated proteins in VEGF and PlGFstimulated cells In order to ensure that the receptors were functional tyrosine kinases, we stimulated the cells and analysed the induction of tyrosine phosphorylation (Figure 2). VEGF-stimulation was performed on 35S-methionine labelled PAE/Flt 1 and PAE/KDR cells, followed by immunoprecipitation with anti-phosphotyrosine monoclonal antibodies or anti-receptor antibodies. PAE/ KDR cells contained increased levels of phosphotyrosine-proteins after VEGF stimulation. A component of about 200 kDa was likely to correspond to the receptor, as judged from the similar migration rate of the receptor protein analysed in parallel (Figure 2, right panel). As shown before, VEGF stimulation of PAE/Flt 1 cells led to a very small increase in the level of tyrosine phosphorylation. We therefore tested the possibility that PlGF, which has been shown to bind to Flt 1, would have a more potent eect. As seen in Figure 2, PlGF treatment did indeed increase the level of tyrosine phosphorylation in Flt 1 cells to a higher extent than VEGF treatment. Thus, a number of components were more strongly phosphorylated in PlGF-treated cells, than in untreated cells. These included a 180 kDa species, which as judged from the pattern of an anti-Flt 1 receptor immunoprecipitate, run in parallel, could correspond to the Flt 1 receptor protein. On the other hand, both VEGF receptor types have been found to vary in the extent of N-linked glycosylation, and therefore in size, between dierent cell types (Takahashi and Shibuya, 1997). We can therefore not exclude the possibility that the 200 kDa tyrosine phosphorylated component in the PlGF-stimulated PAE/Flt 1


PAE/Flt-1 Immunoprec: Rec Ptyr Ptyr Ptyr Stimulation: – – PIGF VEGF kDa 200 —

PAE/KDR Rec Ptyr Ptyr Ptyr – – PIGF VEGF —

— KDR

— Flt-1

97 —

added on the opposite side of the ®lter. As a control, 10% foetal bovine serum was used as a chemoattractant. After 4 h incubation at 378C, the number of cells which had moved through the ®lter was counted. As seen in Figure 3, an ecient chemotactic response was achieved when stimulating PAE/KDR cells with VEGF, whereas PlGF-stimulation had no eect on these cells. PAE cells expressing Flt 1 failed to undergo chemotaxis towards VEGF as well as PlGF.

—

PlGF stimulates DNA synthesis in Flt 1 expressing PAE cells 69 —

—

Figure 2 PlGF-induced tyrosine phosphorylation in Flt 1expressing PAE cells. PAE/Flt 1 and PAE/KDR cells were metabolically labelled with 35S-methionine and stimulated or not with 100 ng/ml of PlGF or VEGF for 7 min at 378C. Labelled cells were lysed and immunoprecipitated with the dierent antibodies against the receptors and against phosphotyrosine. The migration positions of the receptors are indicated, as well as migration positions of marker proteins (myosin, 200 kDa; phosphorylase b, 97 kDa; bovine serum albumin, 69 kDa) run in parallel. Receptor speci®c (open and closed triangles) and common phosphoproteins (diamond) are indicated as described in the text

cells, represents the Flt 1 receptor. The Flt 1 receptor immunoprecipitate also contained a prominent 100 kDa band, the identity of which is unclear, but which could represent degraded receptor (see Kanda et al., 1996). It is relevant to compare the pattern of tyrosine phosphorylated proteins in PAE/Flt 1 cells, stimulated with PlGF, with that of PAE/KDR cells stimulated with VEGF, as shown in Figure 2. It appears that phosphorylation of certain proteins is induced in both cell types upon PlGF- and VEGF-stimulation, respectively. An example is the 130 kDa band, indicated by an open diamond in Figure 2. There are also proteins that appear to be tyrosine phosphorylated in a ligand/ receptor speci®c manner, such as the 30, 65, 80, 145 and 200 kDa components in the PAE/Flt 1 cells (indicated by open arrows) and the 45, 140 and 150 kDa components in the PAE/KDR cells (indicated by closed arrows). KDR, but not Flt 1, is able to mediate chemotaxis VEGF has been implicated in endothelial-speci®c biological signalling, such as induction of endothelial cell proliferation and angiogenesis. To study whether the PlGF-induced increase in tyrosine phosphorylation in PAE cells represents an increased capacity for PlGFstimulated signal transduction, we performed dierent assays for biological signalling. First, we analysed the eect of PlGF and VEGF on migration of PAE cells using the modi®ed Boyden chamber technique. PAE cells expressing either Flt 1 or KDR were suspended in serum-free Ham's F12 medium and placed above a 8 mm-thick nitrocellulose ®lter. Ten ng/ml of either VEGF or PlGF in serum-free Ham's F12 medium was

The capacity of PlGF and VEGF to stimulate DNA synthesis in PAE cells was investigated using a [3H]thymidine incorporation assay. PAE/Flt 1 and PAE/KDR cells were serum-starved for two days and then stimulated with 1 and 10 ng/ml of VEGF or PlGF (Figure 4a and b). VEGF-stimulation of PAE/KDR cells resulted in a more than twofold incorporation of [3H]thymidine relative to unstimulated cells, whereas PAE/Flt 1 cells showed no increase in [3H]thymidine incorporation in response to VEGF. However, stimulation of cells expressing Flt 1 with PlGF led to a signi®cant increase in DNA synthesis. A less pronounced response to PlGF was observed in PAE cells expressing KDR; this response was probably mediated by Flt 1 endogenously expressed in the PAE cells. In a similar assay, using a broader range of concentrations of PlGF, PAE/Flt 1 cells responded to low concentrations in a dose-dependent manner with a maximal response at 1 ng/ml (Figure 4c). Furthermore, PlGF treatment induced signi®cant increases in DNA synthesis in four independent PAE/Flt 1 cell clones (data not shown). PlGF induces MAP kinase activity in PAE cells Since activation of MAP kinase is known to be an important step in the mitogenic signalling by receptor tyrosine kinases, we investigated whether PlGF is able to induce activation of MAP kinase in endothelial cells. For this purpose, PAE cells expressing KDR or Flt 1 were treated with 100 ng/ml of either VEGF or PlGF for 30 min at 48C, followed by 5 min incubation at 378C. The cells were lysed and MAP kinase was immunoprecipitated and subjected to a kinase reaction in which myelin basic protein (MBP) was present as an exogenous substrate. Incorporation of 32P-radioactivity into MBP, as analysed by subsequent SDS ± PAGE and autoradiography, was taken as a measure of MAP kinase activity. Treatment of PAE/Flt 1 with PlGF resulted in a twofold increase in MAP kinase activity whereas VEGF had essentially no eect (Figure 5). KDR expressing PAE cells showed a clear induction of MAP kinase activation upon VEGF-stimulation. The weaker activation of MAP kinase after PlGF stimulation of PAE/KDR cells is consistent with the result of the mitogenic assay presented above. Phospholipase C-g (PLC-g) is a signal transduction molecule that is known to interact with receptor tyrosine kinases and which has been suggested to be involved in mitogenic as well chemotactic signalling. We analysed the phosphorylation of PLC-g in PAE cells expressing Flt1 or KDR. As shown in Figure 6, VEGF stimulation of PAE/KDR led to receptorassociation and tyrosine phosphorylation of PLC-g,

361


362

Figure 3 Eects of VEGF and PlGF on migration of PAE cells expressing KDR or Flt 1. PAE/Flt 1 and PAE/KDR cells were analysed for their migration towards 10 ng/ml of VEGF or PlGF in a modi®ed Boyden chamber as described under Material and methods. Addition of 10% FBS as a chemoattractant, served as a control. In each experiment, the number of migrating cells were counted in at least two representative high power ®elds of the microscope. The results show mean+s.e.m. of three experiments

while Flt 1 on the other hand apparently failed to mediate phosphorylation of PLC-g, irrespective of whether VEGF or PlGF was used for stimulation of the PAE/Flt 1 cells. PlGF enhances the expression level of plasminogen activator An early and important step in angiogenesis is believed to be induction of plasminogen activator (PA). Angiogenic factors such as VEGF and bFGF have been shown to stimulate levels of expression as well as enzymatic activity of plasminogen activator in dierent cell lines. In order to examine whether PlGF has a role in PA induction, we examined the protein levels of urokinase-type plasminogen activator (u-PA) in PAE cells before and after VEGF- or PlGF-stimulation. PAE cells expressing Flt 1 or KDR were treated or not with VEGF or PlGF for 20 h. [35S]methionine was present in the culture medium the last 8 h of the incubation. The conditioned medium was collected and incubated with anti-human urokinase antibodies. The immunoprecipitates were analysed by SDS ± PAGE (Figure 7). Interestingly, PlGF was a more potent inducer of u-PA than VEGF. Moreover, the eect of VEGF was transduced only by Flt 1 and not by KDR. Discussion We show in this paper, that PlGF rather than VEGF, stimulates Flt 1 mediated signal transduction and that

the cellular responses transduced by Flt 1 and KDR dier. Activation of MAP kinase and increased protein levels of plasminogen activator were to some extent detected also in the VEGF-stimulated Flt 1 cells, indicating that this ligand-receptor complex is functional, but probably inecient in eliciting downstream signal transduction. PlGF shows a limited expression pattern, mainly restricted to the placenta (Maglione et al., 1993), whereas VEGF is expressed in a wide variety of cell types (Shibuya , 1995). Thus, our data indicate a role for PlGF-stimulated Flt 1 signal transduction during foetal development. High anity binding of growth factors, such as that of VEGF to Flt 1, should result in dimerization of receptors, activation of the receptor tyrosine kinase and tyrosine phosphorylation of receptor molecules. It appears that VEGF fails to properly induce this series of events, in spite of the high anity binding to Flt 1 expressed in PAE cells (Waltenberger et al., 1994). Receptor binding of the isoform of VEGF used in this study, VEGF165, has been shown to be in¯uenced by heparan-containing proteoglycans (Tessler et al., 1994). This is in parallel to the ®broblast growth factor receptor system, where heparan-containing proteoglycans have been proposed to present the growth factor to cells such that dimerization of receptor molecules is facilitated (Wilkie et al., 1995). Thus, it is possible that the PAE cells lack a speci®c proteoglycan, critical for ligand-induced dimerization of receptor molecules, but not for receptor-binding per se. The isoform of PlGF we have used, PlGF-1, lacks a heparin-binding site (Park et al., 1994). On the other hand, other cell types,


363

a

b

c

Figure 4 PlGF-stimulated DNA synthesis in PAE cells expressing Flt 1. PAE/Flt 1 and PAE/KDR cells were grown in serum-free medium for 48 h and then incubated for 24 h in serum-free medium containing 0.5 mCi [3H]thymidine and 0.2% BSA and dierent concentrations of VEGF or PlGF (1 or 10 ng/ml; panels a and b). The incubation was terminated by trichloroacetic acid-precipitation, and the precipitated radioactivity was measured by scintillation counting. In c, PAE/Flt 1 cells were exposed to a broader range of concentrations of PlGF

expressing Flt 1 alone, have also responded poorly or not at all to VEGF (Seetharam et al., 1995). Seetharam et al. have suggested that Flt 1 is functional as a tyrosine kinase only in endothelial cells; our data, however indicate that the biological eect is determined by the growth factor rather than the cell type. Although Flt 1 contains a tyrosine kinase domain with the expected structural elements, it is very poorly autophosphorylated in response to ligand stimulation as shown previously (de Vries et al., 1992; Waltenber-

ger et al., 1994) and in this work. However, this does not seem to aect the magnitudes of the biological responses. The slightly weaker mitogenic response in Flt 1/PAE relative to KDR/PAE is in agreement with the somewhat lower number of receptors in this cell line. PlGF has previously been shown to bind to Flt 1, but with lower anity than VEGF-binding to this receptor (Clauss et al., 1996; Park et al., 1994). The binding site for both PlGF and VEGF has been


364

determined to the second immunoglobulin-like domain of Flt 1 (Davis-Smyth et al., 1996). We have not directly proven that PlGF exerts its eect by binding to Flt 1 on the PAE cells, which express this receptor protein to some extent already before transfection. In experiments where stimulation of cells were performed using a mix of PlGF and VEGF, the eect on cellular responses expected to arise from PlGF alone, was attenuated (data not shown). This result indicates that PlGF and VEGF compete for binding to the same receptor on the PAE cells; it is most likely that this receptor is Flt 1. Heterodimers of PlGF and VEGF have been shown to exist in vivo and to be mitogenically active, to the same extent as VEGF homodimers, on HUVE cells (DiSalvo et al., 1995). We examined the eect of VEGF/PlGF heterodimers on PAE/Flt 1 and PAE/KDR; the heterodimers had no eect on either cell type (data not shown).

Ligand: kDa

–

PAE/Flt 1 VEGF PIGF

–

PAE/Flt 1 VEGF PIGF (ng/ml) (ng/ml)

PAE/KDR VEGF PIGF

21.5 —

The three dierent cellular responses examined in this paper, induction of plasminogen activator, migration and proliferation, all contribute to the angiogenic process. In addition, angiogenesis requires dierentiation of endothelial cells. Using an in vitro model of endothelial cells cultivated in the middle of two collagen layers, we examined whether PlGF could induce tube-formation of the endothelial cells. Neither PlGF nor VEGF were, however, active in this process (N Ito and L Claesson-Welsh; data not shown), using murine brain-derived endothelial cells expressing Flt 1, but lacking KDR (Kanda et al., 1996). Recent reports on the targeted inactivation of Flt 1 (Fong et al., 1995), KDR (Shalaby et al., 1995) and VEGF (Carmeliet et al., 1996; Ferrara et al., 1996) genes, show fundamental roles for this ligand/receptor family in angiogenesis (formation of blood vessels from pre-existing ones) and vasculogenesis (establishment of the vascular system)

Ligand: kDa 66 —

–

1 10

1 10

PAE/KDR VEGF PIGF (ng/ml) (ng/ml) –

1 10 1 10

MBP 46 —

Figure 5 Eects of VEGF and PlGF on MAP kinase activation. PAE/Flt 1 and PAE/KDR cells were stimulated or not with 100 ng/ml of VEGF or PlGF for 7 min at 378C, lysed and immunoprecipitated (ip) with antiserum against MAP kinase. The immunoprecipitated samples were subjected to in vitro kinase assay in the presence of the exogenous substrate MBP, followed by SDS ± PAGE and autoradiography. The migration position of MBP is indicated by an arrow. The histogram in the lower part of the ®gure shows the relative intensities of the 32P-containing MBP bands in the dierent lanes, quantitated by use of a Fuji Image analyzer and with the band in the unstimulated samples set to 100%, for each cell line

a Ligand: kDa 200 —

Figure 7 Eects of VEGF and PlGF on expression levels of plasminogen activator in PAE/KDR or PAE/Flt 1 cells. Cells were metabolically labelled with 35S-methionine and stimulated or not with VEGF or PlGF. Conditioned medium was collected and immunoprecipitated with an anti-uPA antibody. The samples were subjected to SDS ± PAGE under non-reducing conditions followed by ¯uorography. The histogram in the lower part of the ®gure shows the relative intensities of the 35S-labelled uPA bands in the dierent lanes, quantitated using a Fuji Image Analyzer, and with the band in the unstimulated sample set to 100%, for each cell line

b

PAE/Flt 1 PAE/KDR – VEGF PIGF – VEGF PIGF

Ligand: KDR

uPA

PAE/Flt 1 PAE/KDR – VEGF PIGF – VEGF PIGF

kDa 200 —

PLC-γ

PLC-γ

97 —

97 —

ip: anti-PLC-γ ib: anti-phosphotyrosine

ip: anti-PLC-γ ib: anti-PLC-γ

Figure 6 Tyrosine phosphorylation of PLC-g in PAE/KDR, but not in PAE/Flt 1 cells. Cells were stimulated or not with 100 ng/ml of VEGF or PlGF, lysed and immunoprecipitated (ip) with antiserum against PLC-g. The samples were separated by SDS ± PAGE, transferred to nitrocellulose ®lters and immunoblotted (ib) with phosphotyrosine antibody (a) or antiserum against PLC-g (b)


Figure 8 Schematic ®gure showing ligand-bound and dimerized Flt 1 and KDR. The dierent signal transduction properties and results of biological assays examined in the paper are summarised. Strong response (+), weak response (+) and no response (7) are indicated

during development. Mice lacking the KDR and VEGF genes died in utero, at day 8.5 ± 9. The embryos lacked blood islands and there were no or few endothelial cells. In the Flt 1 7/7 animals, which also died in utero, endothelial cells were detected which were able to form vessels. However, these were disorganised, with endothelial cells in their lumen. These data point to critical roles for these proteins also after birth. Thus, in a rat tumour model, suppression of functional KDR expression (by expression of a dominant-negative truncated KDR) was shown to suppress tumour formation; probably by interference with neo-vascularisation of the tumour (Millauer et al., 1994). One can infer from these studies, that Flt 1 and KDR have dierent roles in angiogenesis and therefore that the receptors activate dierent signal transduction pathways. We show that both receptor types are functional, since both mediated induction of MAP kinase activity and DNA synthesis. The signal transduction pathways initiated by the receptors dier, however, since KDR failed to mediate induction of plasminogen activator, and since PlGF did not stimulate migration of the Flt 1 expressing cells. Furthermore, immunoprecipitation with phosphotyrosine antibodies from stimulated Flt 1 and KDR expressing cells indicated receptor-speci®c phosphorylation of certain proteins. Induction of DNA synthesis and proliferation by receptor tyrosine kinases has been shown to be mediated by many dierent signal transduction molecules that physically interact with and become activated by the receptors. Several of these signalling molecules channel their signal through the Ras pathway and the downstream MAP kinase cascade, whereas other are Ras-independent (Roche et al., 1996). The signal transduction pathway leading to directed migration towards platelet-derived growth factor has been characterised in detail and shown to involve phosphatidyl inositol 3-kinase (PI3-kinase) and PLC-g (Kundra et al., 1995; WennstroÈm et al., 1994). Although we previously reported lack of tyrosine phosphorylation of PLC-g in the KDR expressing PAE cells (Waltenberger et al., 1994), we have now, by

use of a more sensitive antibody, identi®ed complexformation between PLC-g and KDR and weak tyrosine phosphorylation of PLC-g. In the Flt 1 cells, tyrosine phosphorylated PLC-g could not be detected, in agreement with the lack of migration of these cells towards VEGF as well as PlGF. However, VEGF stimulation of NIH3T3 ®broblasts transfected with Flt 1 has been shown to induce tyrosine phosphorylation of PLC-g (Seetharam et al., 1995). We can not exclude the possibility that PLC-g is tyrosine phosphorylated in PAE/Flt 1 cells although at signi®cantly lower levels relative to the PLC-g phosphorylation detected in PAE/KDR cells. Whether PI3-kinase is activated by these growth factors is disputed (Guo et al., 1995) and we have previously reported our negative ®ndings on this matter (Waltenberger et al., 1994). In addition, it is likely that migration can be modulated by other signal transduction pathways, potentially in a cell speci®c manner. Indeed, PlGF has been reported to induce migration of monocytes expressing Flt 1 (Clauss et al., 1996). It is important to continue the characterisation of the signal transduction properties of Flt 1 and KDR in order to understand their in vivo function. This task should not be limited to examination of known signal transduction molecules for their potential interaction with these receptors, since it is possible that transduction of endothelial cell-speci®c functions, such as dierentiation and control of the three-dimensional organisation, involves coupling to endothelial cell speci®c signal transduction molecules.

Materials and methods Cell culture cDNA's for human Flt 1 and human KDR were subcloned into the eucaryotic expression vectors pcDNA1/Neo and pcDNA1 (Invitrogen) respectively, and transfected into porcine aortic endothelial (PAE) cells by electroporation as described (Waltenberger et al., 1994). The PAE cells were cultured in Ham's F12 medium (Life Technologies, Inc.) supplemented with penicillin/streptomycin (Sigma) and 10% fetal bovine serum (Sigma). Analyses were performed on multiple individual PAE cell clones expressing Flt 1 or KDR. Ligands Recombinant VEGF165 expressed in the baculovirus system was a kind gift from D Gospodarowicz, Chiron Corp., Emeryville, CA. PlGF-1 was a kind gift from M Tsang, R & D systems. mRNA extraction and Northern blot analysis Flt 1 or KDR expressing PAE cells were grown in 175 cm2 tissue culture ¯asks to 100% con¯uency. RNA was prepared using the LiCl/urea method as described previously (Landgren et al., 1996) and denatured using formamide/formaldehyde at 558C for 15 min. The samples (20 mg/lane) were separated in 1% agarose gels, containing 2.2 M formaldehyde, and then blotted onto nitrocellulose ®lters (Maniatis et al., 1982). The ®lters were sequentially hybridized with a full-length KDR cDNA probe and a fulllength Flt 1 cDNA probe, which were 32P-labelled by use of the Megaprime DNA labelling system (Amersham). Hybridization was carried out at 428C in 50% formamide, 56SSC (16SSC=15 mM sodium citrate, pH 7.0,

365


366

150 mM NaCl), 56Denhart's solution (2% each of Ficoll, polyvinylpyrrolidone and bovine serum albumine; BSA), 0.1% SDS and 0.1 mg/ml salmon sperm DNA, followed by high stringency washes; 3620 min in 0.16SSC and 0.1% SDS, whereafter the ®lters were exposed on Hyper®lm (Amersham). The ®lters were stripped between the hybridizations, by incubation at 958C in 0.016SSC and 0.1% SDS for 15 min.

and replaced with serum-free medium containing 0.2% BSA followed by incubation with 0.5 mCi [3H]thymidine/ml and VEGF or PlGF at dierent concentrations for 24 h at 378C. Cells were then washed and high molecular [3H]radioactivity was precipitated using 5% trichloroacetic acid. The precipitate was washed and solubilized, and [3H]radioactivity was determined by liquid scintillation counting. Several independent PAE cell clones were examined with similar results.

Antibodies A rabbit antiserum speci®cally reacting with KDR was raised against a synthetic peptide corresponding to a sequence in the kinase insert of KDR (Waltenberger et al., 1994). The Flt 1 rabbit antiserum was a kind gift from N Ferrara, Genentech, CA and the antibody against PLC-g was kindly provided by J Schlessinger, Dept of Pharm. New York University Medical Center. A rabbit antiserum reacting with MAP kinase was raised against a carboxyterminal MAP 2 kinase peptide (EETARFQPGYRS). Metabolic labelling and immunoprecipitation PAE/Flt and PAE/KDR cells were grown to subcon¯uency in 75 cm2 tissue culture ¯asks. Cells were cultured for 6 h in methionine-de®cient MCDB 104 medium (SVA, Uppsala, Sweden) containing 0.2% BSA and supplemented with 50 mCi/ml [35S]methionine (Amersham), and then stimulated or not with 100 ng/ml of either VEGF or PlGF for 7 min at 378C in the presence of 100 mM Na3VO4. The monolayers were rinsed once with icecold PBS containing 250 mM Na3VO4 before lysis in 0.5% Triton X-100 (Merck), 0.5% deoxycholate (Sigma), 20 m M Tris-HCl pH 7.5, 150 mM NaCl, 10 mM EDTA, 1% Trasylol (Bayer), 1 mM phenylmethyl sulfonyl ¯uoride (PMSF) and 250 mM Na3VO4. The lysates were cleared by centrifugation at 10 000 g for 10 min and then incubated with the PY 20 antibody (AFFINITI) for 1 h at 48C, followed by incubation for 30 min with Immunosorb A (ECDiagnostics AB, Uppsala, Sweden). Immune complexes were collected by centrifugation and washed once with lysis buer, once with high salt buer (0.5 M NaCl, 1% Triton X-100 and 20 mM Tris-HCl pH 7.5) and once with PBS. Immunoprecipitates were denatured by boiling for 5 min in sample buer (4% SDS, 0.2 M Tris-HCl pH 8.8, 0.5 M sucrose, 5 mM EDTA, 0.01% bromophenol blue and 2% bmercaptoethanol), and resolved by SDS ± PAGE in a 7% polyacrylamide gel. The gel was treated with Amplify2 (Amersham), dried and exposed to X-ray ®lm (Fuji). Chemotaxis assay The assay was performed in a modi®ed Boyden chamber as described (Auerbach et al., 1991) using micropore nitrocellulose ®lters (8 mm-thick, 8 mm-pore) coated with type-1 collagen solution at 100 mg/ml (Vitrogen 100; Collagen Corporation). Cells were trypsinized and resuspended at a concentration of 16106 cells per ml in serumfree medium containing 0.2% BSA. The cell suspension was placed in the upper chamber and serum-free medium containing 0.2% BSA and 10 ng/ml of either VEGF or PlGF or 10% FBS, was placed below the ®lter in the lower chamber. After 4 h at 378C, the medium was removed and the cells sticking to the ®lter were ®xed in 99% ethanol and stained with Giemsa solution. The number of cells that had migrated through the ®lter were counted. All experiments were performed in triplicate. [3H]thymidine incorporation assay Cells were grown to 50% con¯uency in 12-well plates. Forty-eight h before stimulation, the medium was removed

MAP kinase assay After treatment of cells with 100 ng/ml of either VEGF or PlGF for 30 min at 48C, followed by 7 min incubation at 378C, the cells were lysed in lysis buer and subjected to immunoprecipitation using an anti-MAPK rabbit antiserum, as described above. The immune complexes were washed three times with lysis buer and twice with kinase buer (20 mM HEPES pH 7.5, 10 mM MgCl2, 2 mM MnCl2, 100 mM Na3VO4 and 1 mM DTT) and then incubated for 15 min at 308C in 25 ml kinase buer containing 10 mg myelin basic protein (MBP, Sigma), 1 mM protein kinase inhibitor (PKI, Sigma) and 5 mCi [g-32P]ATP (Amersham). The kinase reaction was terminated by addition of 40 ml of 8% SDS, 0.4 M Tris-HCl pH 8.8, 1 M sucrose, 10 mM EDTA, 0.02% bromophenol blue and 4% b-mercaptoethanol and the samples were boiled for 5 min and resolved by SDS ± PAGE in a 15% SDSpolyacrylamide gel. After ®xation in methanol/acetic acid, the gels were dried and analysed by autoradiography. For quanti®cation of radioactivity, a Bio-Imaging analyser BAS2000 (Fujix) was used. PLC-g phosphorylation Cells were stimulated with the appropriate ligand and lysed in lysis buer as described above. Immunoprecipitation was performed using PLC-g antiserum. Proteins were resolved by SDS ± PAGE and electrophoretically transferred onto nitrocellulose membranes (Hybond-C extra, Amersham), which were incubated with anti-phosphotyrosine antibody (4G10, UBI) or PLC-g antiserum, followed by incubation with peroxidase-conjugated sheep antimouse immunoglobulins or peroxidase-conjugated swine anti-rabbit immunoglobulins (Amersham). After washing, bound antibody was visualized using the ECL Western blotting detection system. Analysis of uPA protein levels PAE/Flt and PAE/KDR cells were seeded at 3.86105 cells per well in 6-well plates. When cells reached con¯uency the medium was changed to serum-free Ham's F-12 medium containing 0.2% BSA. Cells were stimulated or not with VEGF or PlGF. After 8 h incubation, cells were labelled for 12 h in methionine-de®cient MCDB 104 medium (SVA, Uppsala, Sweden) supplemented with 30 mCi/ml [35S]methionine (Amersham), 2% FBS and growth factors as above. The conditioned media were collected and centrifuged in order to remove cells and debris. One ml conditioned medium was immunoprecipitated with 25 mg anti-human urokinase antibody (American Diagnostica Inc.) followed by adsorption to Immunosorb A (EC Diagnostics AB, Uppsala, Sweden). The beads were washed three times with PBS and once with doubledistilled water. The samples were then heated and analysed by SDS ± PAGE (10% acrylamide) under nonreducing conditions. Gels were ®xed for 20 min in 10% acetic acid, 40% methanol, followed by a 20 min incubation in Amplify2 (Amersham). Gels were then dried and exposed to Hyper®lm MP2 (Amersham). Quantitation was performed using a Fuji Image Analyzer.


Acknowledgements We thank Michael Cross for critical reading of the

manuscript. This work was in part supported by a grant to Lena Claesson-Welsh from the Swedish Cancer Society.

References Auerbach R, Auerbach W and Polakowski I. (1991). Pharmac. Ther., 51, 1 ± 11. Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M, Fahrig M, Vandenhoeck A, Harpal K, Eberhardt C, Declercq C, Pawling J, Moons L, Collen D, Risau W and Nagy A. (1996). Nature, 380, 435 ± 439. Clauss M, Weich H, Breier G, Knies U, RoÈckl W, Waltenberger J and Risau W. (1996). J. Biol. Chem., 271, 17629 ± 17634. Cunningham SA, Waxham MN, Arrate PM and Brock TA. (1995). J. Biol. Chem., 270, 20254 ± 20257. Davis-Smyth T, Chen H, Park J, Presta LG and Ferrara N. (1996). EMBO J., 15, 4919 ± 4927. de Vries C, Escobedo JA, Ueno H, Houck K, Ferrara N and Williams LT. (1992).Science, 255, 989 ± 991. DiSalvo J, Bayne ML, Conn G, Kwok PW, Trivedi PG, Soderman DD, Palisi TM, Sullivan KA and Thomas KA. (1995). J. Biol. Chem., 270, 7717 ± 7723. Dougher-Vermazen M, Hulmes JD, BoÈhlen P and Terman BI. (1994). Biochem. Biophys. Res. Commun., 205, 728 ± 738. Ferrara N, Carver-Moore K, Chen H, Dowd M, Lu L, O'Shea KS, Powell-Braxton L, Hillan KJ and Moore MW. (1996). Nature, 380, 439 ± 442. Fong G-H, Rossant J, Gertsenstein M and Breitman ML. (1995). Nature, 376, 66 ± 70. Guo D, Jia Q, Song H-Y, Warren RS and Donner DB. (1995). J. Biol. Chem., 270, 6729 ± 6733. Joukov V, Pajusola K, Kaipainen A, Chilov D, Lahtinen I, Kukk E, Saksela O, Kalkkinen N and Alitalo K. (1996). EMBO J., 15, 290 ± 298. Kaipainen A, Korhonen J, Mustonen T, van Hinsbergh VWM, Fang G-H, Dumont D, Breitman M and Alitalo K. (1995). Proc. Natl. Acad. Sci. USA, 92, 3566 ± 3570. Kanda S, Landgren E, LjungstroÈm M and Claesson-Welsh L. (1996). Cell Growth and Dierentiation, 7, 383 ± 395. Kundra V, Anand-Apte B, Feig LA and Zetter BR. (1995). J. Cell Biol., 130, 725 ± 731. Landgren E, Eriksson A, WennstroÈm S, Kanda S and Claesson-Welsh L. (1996). Exp. Cell Res., 223, 405 ± 411. Maglione D, Guerriero V, Viglietto G, Ferraro MG, Aprelikova O, Alitalo K, Del Vecchio S, Lei K-J, Chou JY and Persico MG. (1993). Oncogene, 8, 925 ± 931. Maniatis T, Fritsch EF and Sambrook J. (1982). Molecular Cloning. A laboratory manual. Cold Spring Harbor, N.Y. Millauer B, Shawver LK, Plate KH, Risau W and Ullrich A. (1994). Nature, 367, 576 ± 579.

Millauer B, Wizigmann-Voos S, SchnuÈrch H, Martinez R, Mùller NPH, Risau W and Ullrich A. (1993). Cell, 72, 835 ± 846. Miyazono K, Okabe T, Urabe A, Takaku F and Heldin C-H. (1987). J. Biol. Chem., 262, 4098 ± 4103. Neufeld G, Tessler S, Gitay-Goren H, Cohen T and Levi BZ. (1994). Prog. Growth Factor Res., 5, 89 ± 97. Park JE, Chen HH, Winer J, Houck KA and Ferrara N. (1994). J. Biol. Chem., 269, 25646 ± 25654. Pierce EA, Avery RL, Foley ED, Aiello LP and Smith LEH. (1995). Proc. Natl. Acad. Sci. USA, 92, 905 ± 909. Plate KH, Breier G, Weich HA, Mennel HD and Risau W. (1994). Int. J. Cancer, 59, 520 ± 529. Quinn TP, Peters KG, De Vries C, Ferrara N and Williams LT. (1993). Proc. Natl. Acad. Sci. USA, 90, 7533 ± 7537. Roche S, McGlade J, Jones M, Gish GD, Pawson T and Courtneidge SA. (1996). EMBO J., 15, 4940 ± 4948. Sawano A, Takahashi T, Yamaguchi S, Aonuma M and Shibuya M. (1996). Cell Growth & Di., 7, 213 ± 221. Seetharam L, Gotoh N, Maru Y, Neufeld G, Yamaguchi S and Shibuya M. (1995). Oncogene, 10, 135 ± 147. Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu X-F, Breitman ML and Schuh AC. (1995). Nature, 376, 62 ± 66. Shibuya M. (1995). Adv. Cancer Res., 67, 281 ± 316. Shweiki D, Itin A, Soer D and Keshet E. (1992). Nature, 359, 843 ± 845. Takahashi T and Shibuya M. (1997). Oncogene, 14, 2079 ± 2089. Terman BI, Dougher-Vermazen M, Carrion ME, Dimitrov D, Armellino DC, Gospodarowicz D and Bohlen P. (1992). Biochem. Biophys. Res. Commun., 187, 1579 ± 1586. Tessler S, Rockwell P, Hicklin D, Cohen T, Levi BZ, Witte L, Lemischka IR and Neufeld G. (1994). J. Biol. Chem., 269, 12456 ± 12461. Waltenberger J, Claesson-Welsch L, Siegbahn A, Shibuya M and Heldin C-H. (1994). J. Biol. Chem., 269, 26988 ± 26995. WennstroÈm S, Siegbahn A, Yokote K, Arvidsson A-K, Heldin C-H, Mori S and Claesson-Welsh L. (1994). Oncogene, 9, 651 ± 660. Wilkie AOM, Morriss-Kay GM, Jones EY and Heath JK. (1995). Curr. Biol., 5, 500 ± 507. OÈberg C, Waltenberger J, Claesson-Welsh L and Welsh M. (1994). Growth Factors, 10, 115 ± 126.

367