Basic fibroblast growth factor modulates the mitogenic potency of the platelet-derived growth factor (PDGF) isoforms by specific upregulation of the PDGF alpha receptor in vascular smooth muscle cells.

Platelet-derived growth factor AA (PDGF AA), in contrast to PDGF AB and BB, is a poor mitogen for smooth muscle cells (SMC). However, together with basic fibroblast growth factor (bFGF) it acts synergistically on DNA synthesis of these cells. Northern blot analysis revealed that bFGF selectively increases the PDGF-receptor alpha subtype (PDGF-R alpha) mRNA level without a significant effect on the PDGF-R beta mRNA level. The amount of PDGF-R alpha protein is also selectively increased after stimulating SMC with bFGF as shown by immunoprecipitation of lysates from SMC with anti-PDGF-R alpha antibodies. The number of binding sites for 125I-PDGF AA is more than doubled after bFGF-treatment, whereas the specific binding for PDGF AB and BB increased only by approximately 30 and 20%, respectively. The increase in the number of PDGF-R alpha renders the SMC responsive for PDGF AA as demonstrated by the induction of the proto-oncogene c-fos as well as by an increased cell proliferation. The enhanced PDGF binding after bFGF treatment may in fact explain the observed synergistic behavior. These data are discussed with regard to a possible role of growth factor-induced transmodulation of receptor expression during atherogenesis.

The distinct binding characteristics of the PDGF isoforms led to the identification of two different PDGF receptor subtypes (Hart et al., 1988Heldin et al., 1988. The a receptor subtype (PDGF-Ra) binds either a B chain or an A chain of PDGF, whereas the p receptor subtype (PDGF-Rp) can bind only a B-chain . Many cell types like human fibroblasts and human smooth muscle cells express both receptor subtypes (Heldin and Westermark, 1990), but there are also examples for cells having only one type of PDGF receptor (Smits et al., 1989;. After binding of ligands to the PDGF-RP a dimerization of receptor molecules takes place that is closely associated with receptor kinase activation Bishayee et al., 1989). Also PDGF a receptors (PDGF-Ra) undergo dimerization, and ap heterodimers can be formed as well , Kanakarj et al., 1991. Recent findings suggest that the responsiveness of cells to PDGF depends on the availability of the ligand as well as on the expression of the corresponding receptors (Heldin and Westermark, 1990). The expression and the availability of the two PDGF receptor types can be modulated by regulatory molecules (Gronwald et al., 1989;Battegay et al., 1990). For example, bFGF can modulate the PDGF-driven differentiation of 0-2A progenitor cells into oligodendrocytes by up-regulating the expression of functional PDGF-Ra (McKinnon et al., 1990).
Basic fibroblast growth factor (bFGF) has a number of biological properties comparable with PDGF; for example it is a potent smooth muscle cell mitogen (Gospodarowicz et al., 1981;. It is also mitogenic for a wide variety of cells of mesodermal and neuroectodermal origin (Folkman and Klagsbrun, 1987;Rifkin and Moscatelli, 1989). The wide distribution of bFGF in cells and tissues suggests a role in the regulation of normal tissue function. An important feature of this growth factor is that it lacks a signal peptide for secretion (Abraham et al., 1986) and, in contrast to PDGF, appears to be cell-associated rather than secreted (Vlodavsky et al., 1987). bFGF has been found in the extracellular matrix (ECM), and in the basement membrane (Folkman et al., 1988), bound to heparan sulfate proteoglycans and glycosaminoglycans. It was proposed that the mitogen can be mobi-lized when needed by remodeling of the basement membrane or ECM by activation of hydrolases, e.g. heparinases (Vlodavsky et al., 1987).
By analyzing the effects of combinations of different growth factors on DNA synthesis and proliferation of smooth muscle cells, we found that many combinations act strongly synergistic. Given this synergism among growth factors, we hypothesized that this may be the result of the up-regulation of their growth factor receptors. In this report we demonstrate that bFGF up-regulates PDGF-Ra expression in a nontransient way. The increase of PDGF-Ra transcription shown by Northern blot analysis was time-and dose-dependent and decreased after removal of bFGF. The increased amount of PDGF-Ra was demonstrated by an increase of binding sites for the lz5I-PDGF isoforms, in particular for lzSI-PDGF AA. This may explain the observed particular increase in mitogenicity for the PDGF AA isoform after bFGF treatment.

EXPERIMENTAL PROCEDURES
Materials-PDGF AA, AB, and BB are recombinant proteins expressed in Escherichia coli (Hoppe et al., 1989. Recombinant human bFGF was a generous gift of the Takeda Chemical Company (Osaka, Japan) or was obtained from Progen (Heidelberg, Federal Republic of Germany (FRG)). EGF was from Amgen (Thousand Oaks, CA) and IGF-1 from Boehringer (Mannheim, FRG). All growth factors were reconstituted after the manufacturer's instructions, aliquoted, and stored a t -20 or -70 "C. After thawing, aliquots were diluted and immediately used. The PDGF-Ra antibodies were a generous gift from Dr. Lea Eisenbach, Weizman Institute of Science, Israel. Heparin was purchased from Sigma (Munich, FRG) and Suramin from Mobay Chemical (New York).
Cultivation of Cells-Bovine aortic smooth muscle cells (BSMC) were explanted from bovine aorta according to the method of Voyta et al. (1984) and grown in Dulbecco's modified Eagle's medium (DMEM, GIBCO) containing 10% calf serum (HyClone, Logan, UT). Cells were identified as smooth muscle cells by their "hill and valley" growth pattern typical for vascular smooth muscle cells and by staining with a monoclonal antibody specific for SMC a-actin (Skalli et al., 1986). Human vascular smooth muscle cells (HSMC, obtained from Joachim Rupp, University of Tubingen, FRG) were explanted from the media of the aortic arch from a young male donor, isolated by enzymatic disaggregation and grown in Waymouth/Ham's F-12 medium (l:l, v/v) containing 15% fetal calf serum (Boehringer). HSMC were identified to be SMC as described for BSMC. BSMC and HSMC between passage 3 and 6 were used for all experiments.
PHlThymidine Incorporation Assay-BSMC were plated into a 24-well tray (20,000 cells/well) in culture medium supplemented with 10% calf serum. After 5 days cells reached confluence and were made quiescent by incubation in DMEM containing 1% calf serum for another 2 days. Stimulation with growth factors was carried out in 0.5 ml/well defined medium (DMEM + 100 pg/ml BSA) for 24 h.
After 18 h 1 pCi/well [3H]thymidine was added. 6 h later the stimulation was stopped by removing the radioactive media, washing the cells with phosphate-buffered saline, and fixing them with ice-cold methanol (0.5 ml/well for 2 X 5 min). Then cells were incubated with 5% trichloroacetic acid (2 X 10 min). Finally the acid-insoluble fraction was lysed in 0.5 ml/well 0.3 M NaOH and the radioactivity determined by liquid scintillation counting.
Proliferation Assays-To measure BSMC proliferation cells were plated sparsely (5000 cells/well) in a 24-well tray in DMEM, 10% calf serum. About 24 h after seeding medium was replaced by DMEM containing 1% calf serum, and growth factors were added. Medium and growth factors were changed every other day. After 2, 4, and 6 days cells were dissociated with trypsin/EDTA and counted in triplicate using a cell counter (Coulter Electronics).
Induction Experiments with BSMC and HSMC-For induction experiments with subsequent RNA isolation BSMC and HSMC were grown in 75 or 150-cmZ tissue culture flasks (Costar). After cells reached confluence, they were brought to quiescence as described for the thymidine incorporation assay and then stimulated with growth factors. After the indicated time periods cells were washed once with PBS and used for RNA preparation. In some experiments bFGF was removed after 12 h from low and high affinity binding sites by washing twice with 200 pg/ml heparin and 1 mM suramin for 10 min. Low affinity binding sites are defined as sites from which bFGF can be removed by treatment with 2 M NaCl or with heparin (Moscatelli, 1987;Vlodavsky et al., 1987). High affinity binding sites (FGF receptors) are defined as sites from which bFGF can be removed by treatment with suramin, which blocks the interaction between the growth factor and the growth factor receptor (Yayon and Klagsbrun, 1990).
Preparation of RNA and Northern Blot Hybridization-Total RNA was prepared according to the method of Chirgwin et al. (1979), except that the protease digestion step was omitted. 10 pg total RNA were denatured at 65 "C in a solution containing 2.2 M formaldehyde, 50% (v/v) formamide, and ethidium bromide. The RNA was electrophoresed in a formaldehyde containing 1.25% agarose gel that was prepared by a surface-tension method (Rosen et al., 1990). Following transfer of RNA to a nitrocellulose membrane (Schleicher & Schull) using 10 X SSC (1.5 M NaC1,0.15 M sodium citrate, pH 7.0) the filter was baked in a vacuum oven at 80 "C for 2 h. The blot was hybridized at 65 "C in 2 X SSC, 10 X Denhardt's solution (Denhardt, 1966), 2.5% dextran sulfate, 0.1% SDS, 0.1% sodium pyrogenphosphate, 2 mM EDTA, and 30 pg/ml denatured salmon sperm DNA. The cDNA probes were labeled with [32P]dCTP (3000 Ci/mmol, Amersham) using an oligolabeling kit (Pharmacia LKB Biotechnology Inc.). The filters were washed 2 X 15 min in 2 X SSC, 0.1% SDS, 0.1% sodium pyrogenphosphate, 2 mM EDTA, then 2 X 15 min in 1 X SSC, 0.1% SDS, 2 mM EDTA, and finally 2 X 15 min in 0.4 X SSC, 0.1% SDS, 2 mM EDTA at 65 "C.
Densitometric Analysis-For quantification of mRNA induction after Northern blot analysis, blots were exposed without an intensifying screen using Kodak X-Omat AR-films. If not overexposed, the development of silver grains in x-ray films by the autoradiographic signal is proportional over a wide range (Bonner and Laskey, 1974). The relative autoradiographic signal was then quantified using an LKB laser densitometer (Ultroscan 2202). The arithmetic mean of several measurements was used to determine the relative autoradiographic signal.
Immunoprecipitation and Western Blot-Confluent serum-starved BSMC were stimulated with 5 ng/ml bFGF in defined media (DMEM + 100 pg/ml BSA) for 18 h. Control cells remained unstimulated. lo' cells were washed twice with PBS, scraped with a rubber policeman, and centrifuged at 900 rpm for 10 min. The cell pellet was lysed in 1 ml ice-cold RIPA buffer (20 mM Tris/HCl, pH 7.5, 150 mM NaCl, 0.1% (w/v) sodium deoxycholate, 0.1% (v/v) Triton X-100, 0.1% (v/ v) SDS, 1 mM phenylmethylsulfonyl fluoride, 10 pg/ml aprotinin, 10 pg/ml leupeptin, 10 pg/ml pepstatin). Lysates were homogenized by 10 strokes in a Dounce homogenizer and centrifuged at 10,000 rpm for 15 min. After equilibration of the protein concentrations, the supernatants were precleared using rabbit nonimmune serum and Staphylococcus aureus Cowan strain A (Sigma). An immunoprecipitation with rabbit antiserum R9 (1:250), made against a peptide in the C terminus of PDGF-Ra,' was carried out overnight. Protein A-Sepharose CL-4B (Pharmacia) was used to precipitate the immune complexes. Beads were washed 4 times in RIPA buffer, and the immune complexes were eluted by boiling the beads for 5 min in SDS-sample buffer containing 2% mercaptoethanol. Proteins were electrophoresed by SDS-PAGE and transferred to 0.1-pm nitrocellulose filters (Schleicher & Schull) using semi-dry blotting. Filters were blocked by incubating in TBS (0.05 M Tris/HCl, pH 7.5,0.15 M NaCl, 0.03% (v/v) NaN3) containing 2% BSA. Then blots were incubated for 3 h with rabbit antiserum R5, made against a peptide in the interkinase region of PDGF-Ra (Do et al., 1992), diluted 1:250 in antibody solution (TBS + 2% BSA + 0.2% Nonidet P-40). The primary antibody was removed and the blots washed three times for 5 min in washing solution (TBS + 0.5% BSA + 0.2% Nonidet P-40).
To detect antibody reactions blots were incubated for 1 h with antirabbit horseradish peroxidase-conjugated IgG (Dianova, FRG) diluted 1:100,000 in antibody solution, washed three times in washing solution, and then developed using a chemoluminescence detection system (Amersham).
Metabolic Labeling of Cells and Immunoprecipitation-Confluent HSMC were made quiescent by incubation in culture medium containing 1% FCS for 2 days. Then medium was replaced by defined media (Waymouth medium/Ham's F-12 medium, 1:l (v/v) + 100 pg/ ml BSA) and cells were stimulated with 5 ng/ml bFGF for 18 h. Control cells remained unstimulated. After 10 h of stimulation, cell Eagle's MEM (GIBCO), and medium was changed to 4 ml/plate layers were washed two times with methionine-and cysteine-free C. Fitzer-Attas, unpublished results. . After another 8-h incubation, cells were washed two times in cold PBS and then lysed in 1.5 ml of ice-cold RIPA buffer. After 30 min the cell layers were scraped with a rubber policeman and the lysates centrifuged at 10,000 rpm for 15 min. After equilibration of the protein concentrations the supernatants were incubated with 0.5 ml of WGA-Sepharose (Pharmacia) for 3 h. The binding glycoprotein fraction was eluted with 1 ml of 0.2 M acetylglucosamine in RIPA buffer and precleared by incubation with nonimmune serum for 1 h followed by incubation with formalin-fixed S. aurew Cowan strain A in PBS (Sigma) for another 30 min. Then an immunoprecipitation with rabbit anti-PDGF-RP antiserum PDGFR-3 (Claesson-Welsh et al., 1989), diluted 1:200, was carried out. Immune complexes were precipitated using protein A-Sepharose CL-4B. After washing four times in RIPA buffer, immune complexes were eluted by boiling the beads for 5 min in SDS-sample buffer containing 2% mercaptoethanol. SDS-PAGE was done in 7% gels, according to the method of Laemmli (1970). For fluorography gels were soaked in Amplify (Amersham), dried, and exposed to Kodak X-Omat AR-films.
Preparation of "'I-PDGF Zsoform-PDGF-AA, PDGF-AB and PDGF-BB were iodinated as described by Hunter and Greenwood (1962). Phe-23 of PDGF-BB was replaced by a tyrosine (Hoppe et a l ,  1989). 1 mCi of Na'*'I was mixed with 10 pg of each PDGF isoform in 50 pl of phosphate buffer, pH 7.4, and then 20 pg chloramine T in 10 p1 of H20 was added. After 60 s 20 p1 of tyrosine solution (1 mg/ ml; dissolved in 1 M NaOH and neutralized) were given to the reaction. Then, 10 GI of KI (0.1 M) and 10 91 BSA (10 mg/ml) were added. The entire reaction mixture was loaded an a G-25 column (Pharmacia), equilibrated with 0.1 M acetic acid, and eluted with the same buffer. Fractions (250 pl) containing the labeled protein were pooled, aliquoted, and stored at -30 "C. Purity of iodinated protein was controlled by SDS-PAGE and autoradiography. Specific binding activity was checked by competition experiments with the corresponding unlabeled PDGF isoforms.
Binding Assays-BSMC were seeded into 12-or 24-well trays in DMEM, 10% calf serum and grown to confluence (1-3 X lo5 cells/ well). They were made quiescent by incubating them in DMEM containing 1% calf serum for 2 days. Cells were stimulated with 5 ng/ ml bFGF in defined media (DMEM + 100 pg/ml BSA) over 12 h.
Control cells remained unstimulated. Cells were placed on ice and washed two times with binding buffer (DMEM, 25 mM Hepes, 1 mg/ ml BSA, pH 7.4). Increasing amounts (1-40 ng/ml) of "'I-PDGF isoforms (specific activity 25,000-30,000 cpm/ng) were added to the wells: for each concentration of "'1-PDGF the amount of unspecific binding was estimated separately by the addition of 10 pg/ml of unlabeled ligand. The plates were incubated at 4 "C for 4 h, washed five times with washing buffer (PBS with eaz+ and M e , 1 mg/ml BSA), and lysed in 500 pl of lysis buffer (1% Triton X-100, 10% glycerol in 20 mM Tris/HCl, pH 7.5). 400 pl of the lysate was used for counting in a gamma-counter.

RESULTS
bFGF Enhances PDGF-induced DNA Synthesis in SMC-The proliferation of vascular SMC can be stimulated by a variety of growth factors. We investigated the effect of the growth factors PDGF (AA, AB, and BB) and bFGF individually as well as combinations of bFGF with individual PDGF isoforms on DNA synthesis of BSMC (Fig. 1). We found that PDGF AA is a poor mitogen for SMC, whereas PDGF AB and BB strongly induce DNA synthesis of these cells (Fig.   lA). bFGF which is known to be a potent mitogen for SMC caused only a small increase of ['Hlthymidine incorporation into BSMC used for these experiments which was comparable with that obtained with PDGF AA. However, the combination of 1 ng/ml bFGF together with 20 ng/ml PDGF AA or AB or 16 ng/ml PDGF BB induced DNA synthesis to an extent that was more than additive as compared with the stimulation obtained by the respective growth factor alone ( Fig. 1, B-D). This synergistic effect was particularly evident for the combination PDGF AA and bFGF, even in lower concentration ranges of PDGF AA (Fig. 1B).
In order to gain more information about the mechanism of the observed synergistic interactions, we investigated whether preincubation with bFGF shows the same effect on the stimulation of DNA synthesis by PDGF as obtained by simultaneous application of the growth factors. In a series of experiments, BSMC were pretreated for 24 h with 5 ng/ml bFGF or were left unstimulated. Then cells were stimulated with 30 ng/ml of the individual PDGF isoforms, and DNA synthesis was determined. We found that pretreatment with bFGF enhanced the PDGF AA-induced DNA synthesis more than &fold, whereas the PDGF AB-and BB-induced DNA synthesis increased only approximately %fold as compared with the FIG. 1. Effect of combinations of bFGF together with PDGF on the stimulation of DNA synthesis on BSMC. BSMC were seeded into 24-well trays in culture medium supplemented with 10% calf serum and grown until they reached confluence. They were made quiescent by incubation in DMEM, 1% BCS for 2 days. Then the cells were stimulated for 24 h with different concentrations of PDGF AA, AB, BB (A), and combinations of the different PDGF isoforms with bFGF (B-D) under serum-free conditions. Incorporation of [3H]thymidine into DNA was measured 18-24 h after beginning of the stimulation as described under "Experimental Procedures." The experiment shown was repeated three times with similar results.  (Table I).
These data indicate that under the influence of bFGF the responsiveness of SMC for PDGF, especially for PDGF AA, is significantly increased.
bFGF Increases PDGF-induced Proliferation of SMC-To investigate whether the synergistic interaction between bFGF and PDGF can also be observed at the level of SMC proliferation, we carried out proliferation assays under low serum conditions. Cells were stimulated with either one of the PDGF isoforms alone or in combination with bFGF, and the cell number was determined over 6 days (Fig. 2). We could show that combinations of bFGF with one PDGF isoform exerts a slight but significant synergistic effect on SMC proliferation. This synergistic effect on cell proliferation was not as significant as the effect on DNA synthesis. Nevertheless, the proliferation data are in agreement with the earlier observation that the responsiveness of cells to PDGF is increased under the influence of bFGF ( Fig. 1 and Table I).
Taken together these data show that bFGF is able to enhance effectively the mitogenic potency of each PDGF isoform with the most pronounced effect for PDGF AA.
bFGF Increases the Level of PDGF-Ra mRNA-We determined whether the bFGF-induced increase of PDGF mitogen- To compensate the changes in basal level of thymidine incorporation due to the bFGF preincubation, data were presented as procentual values (+S.D.) related to the respective positive control (DMEM + 10% calf serum). The positive control without bFGF treatment corresponds to 52,700 cpm/well and the positive control after bFGF treatment to 120,750 cpm/well. Each data point in the experiment represents the mean of three determinations. The experiment was reneated once with similar results. icity could be related to a modulation of PDGF receptor mRNA levels. Total RNA extracted from cells grown in defined media with or without bFGF was used for Northern blot analysis. Using cDNA probes specific for PDGF-Ra and -@, we investigated the effect of bFGF on the transcription of both PDGF receptor genes. Time course experiments indicated that preincubation with bFGF selectively increases PDGF-Ra mRNA levels in BSMC (Fig. 3A). The increase was maximal (about 7-10-fold) after 12-18 h of preincubation and remained substantially unchanged during 36 h of incubation (not shown). In contrast to these results, the levels of PDGF-RP transcripts remained unchanged within the time frame of the experiments (Fig. 3A). The analysis of PDGF receptor transcription after bFGF treatment in HSMC led to comparable results (Fig. 3B).
The increasing effect of bFGF on PDGF-Ra mRNA expression is concentration-dependent (Fig. 4). The Northern blot data indicate that bFGF concentrations of 0.75 ng/ml already maximize the the PDGF-Ra mRNA level. Preincubation of cells with higher concentrations of bFGF leads to a smaller but constant level of PDGF-Ra mRNA. As in the experiments shown in Fig. 3, PDGF-RP mRNA levels are not effected upon preincubation of BSMC with increasing concentrations of bFGF (Fig. 4).
In order to investigate whether the removal of bFGF from pretreated cells could reverse the increase of PDGF-Ra mRNA levels, we dissociated bFGF from low and high affinity binding sites by treatment with 200 Ng/ml heparin and 1 mM suramin. At different time points after this treatment, cells were analyzed for their PDGF-Ra gene transcription (Fig. 5). Removal of bFGF does not significantly decrease PDGF-Ra mRNA levels for 24 h. However, 48 h after dissociation of bFGF, the level of PDGF-Ra transcripts was comparable with that of normal unstimulated BSMC. Hence the removal of bFGF leads to a reversion of PDGF-Ra mRNA up-regulation, and these data suggest that the modulating effect of bFGF requires the constitutive presence of the ligand and therefore probably the constant down-regulation of the FGF-receptor. We made the observation that glyceraldehyde-3-phosphate dehydrogenase gene transcription was influenced by long time bFGF-incubations (see also Fig. 9). Therefore we additionally show the ethidium bromide gel used for RNA blotting to demonstrate equal sample loading (Fig. 5).
In addition we determined whether the observed synergistic cooperativity of bFGF with PDGF in SMC could also depend Time-dependent increase of PDGF-Ra m R N A levels after stimulation with bFCF i n RSMC a n d HSMC. BSMC and HSMC were seeded into culture flasks ( 7 5 cm') and grown to confluence. After a 2-day starvation in hasal medium containing 1% BCS, medium was changed to DMEM + 100 pg/ml BSA for BSMC and Waymouth/Ham's F-12 + 100 pg/ml BSA for HSMC and cells were st.imulated for 0-12 h with 5 ng/ml bFGF. After the indicated time period total RNA was isolated as described under "Experimental Procedures" and subjected to Northern blot analysis. In each lane 10 pg total RNA from BSMC ( A ) and HSMC ( E ) were fractionated on a 1.25% agarose gel and transferred to nitrocellulose filters. The 6.5kh PDGF-Ra transcripts were detected by hybridization with a 0.75kb AccI-EcoRI cDNA fragment that covers the tyrosine kinase region of the PDGF -Ra (top panel). PDGF-R@ transcripts (5.7 kh) hyhridized with a 2.8-kb EcoRI fragment. representing the coding region of the PDGF-R@ cDNA. To control sample loading and RNA integrity radiolabeled DNA was stripped and the filter reprohed using a human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA fragment (bottom panel).
on an increased expression of the FGF-receptor (flg-type). We found that treatment of cells with PDGF AA and BB had no significant effect on the flg-receptor gene expression (not shown). bFGF increased FGF receptor mRNA levels only by a low extent (Fig. 6).
bFCF Stimulates Synthesis of PDCF-Ra Protein-In order to investigate whether elevated levels of PDGF-Ra mRNA 18 s -GAPDH FIG. 5. Reversal of I'I)CF-Ha mRNA up-regulation after treatment with heparin and suramin. I3SMC were seeded into culture flasks (7.5 cm") and grown t o rontluencc in D " I containing 10% calf serum followed by a starvation in I)MEM, 1"; calf serum for 2 days. Then medium WAS changed to DMEM + 1 0 0 pg/ml HSA. and cells were stimulated for 12 h with 5 ng/ml hFGF. After this period, hFGF was dissociated from low and high affinity hinding sites by washing the cells two times for 10 min with 200 pg/ml heparin and 1 mM suramin in DMEM + 100 pg/ml HSA. After the indicated periods of time total RNA was isolated. and Northern hlot analysis was carried out as described in the legend to Fig. 3 . Equal sample loading was demonstrated hy reprohing the filter using a human glyceraldehyde-3-phosphate dehydrogenase ((;A/'I)H) fragment and hy showing the ethidium hromide gel cErh. Nr. 1 used for R N A hlotting.  2 and 4 ) of 5 ng/ml bFGF total cell lysates were extracted. Glycoprotein-enriched fractions (WGA fractions, lanes I and 2) and WGA flow-through (lanes 3 and 4 ) were prepared and immunoprecipitated with an antibody against the PDGF-RP (PDGFR-3). The arrow shows PDGF-RP protein after SDS-PAGE and autoradiography. after bFGF-treatment were also translated into increased amounts of PDGF-Ra protein, lysates from bFGF-treated or untreated BSMC were used for immunoprecipitation with antibodies against PDGF-Ra. The results demonstrate that treatment with bFGF significantly increases the amount of PDGF-Ra protein in SMC (Fig. 7A). In contrast, the amount of PDGF-RP protein is not influenced by bFGF preincubation (Fig. 7 B ) .
These results together with the Northern blot data from Fig. 3 indicate that bFGF is able to up-regulate PDGF-Ra in smooth muscle cells on the mRNA level as well as on the level of protein synthesis.
bFGF Differentially Modulates the Binding of the Three PDGF Isoforms-In order to determine if increased levels of PDGF-Ra protein also result in an elevated number of PDGF binding sites on the cell surface, we studied the binding of ""I-PDGF AA, I2'I-PDGF AB, and "'I-PDGF BB to BSMC after preincubation with bFGF in comparison with untreated cells. As shown in Fig. 8, pretreatment with bFGF leads to an increase of specific binding for ""I-PDGF AA by about 150% (upper panel). The specific binding for ""I-PDGF AB and "'I-PDGF BB was only increased to a lower extent (middle and lowerpunels). Curve fitting of these binding data revealed that the observed increases reflect an elevation of the number of binding sites rather than an increased affinity of the ligands FIG. 8. Binding of different I2"I-PDGF isoforms to BSMC after treatment with bFGF. Confluent quiescent BSMC (1-2 X IO5 cells/well) were left unstimulated (open squares) or were treated with 5ng/ml bFGF (closed circles) for 12 h a t 37 "C and the binding of "'I-PDGF AA, "'I-PDGF AB, or ""I-PDGF BB was determined a t 4 "C in the presence of increasing amounts of labeled isoforms. Bound ligand in the presence of 10 pg/ml unlabeled PDGF was substracted as unspecific binding. Fitting of curves was done using graFit software (Erithacus, Ltd.) for the receptor numbers as indicated. The data points represent the means of three determinations.
to the receptors. The number of binding sites per cell for PDGF AA was more than doubled (approximately 18,000-42,500), whereas the number of binding sites for PDGF AB increased only by approximately 30% and for PDGF BB by approximately 20%, respectively.
Preincubation with bFGF Increases the PDGF "induced c-fos Gene Transcription-To further explore the functional properties of the bFGF-up-regulated PDGF-Ra, we investigated the effect of the PDGF isoforms AA and BB on the transcription of early genes. The PDGF AA and BB stimulated activation of the proto-oncogene c-fos, an immediate early mediator of mitogenic and differentiation stimuli, in bFGF-treated and untreated cells was determined. We found that in the absence of bFGF (Fig. 9, lunes 1-3) PDGF BB Significantly induces the c-fos transcription (lune 3 ) , whereas PDGF AA does not show a detectable effect as compared with unstimulated cells (lune 2). In contrast, in bFGF preincubated cells (Fig. 9, lunes 4-6), PDGF AA significantly induces transcription of c-fos (lune 5 ) . The PDGF BB-induced c-fos transcription does not show any difference between bFGFtreated or control cells (lunes 3 and 6 ) . These results demonstrate that bFGF-treated cells but not control cells are able to mediate a PDGF AA-induced c-fos response, being consistent with a stronger mitogenicity of PDGF AA for these cells.
Synergistic Inteructions between Other Growth Factors-In additional experiments we tried to find out whether synergistic interactions between growth factors are limited to the combination of bFGF and PDGF or whether they might be a  lanes 1 and 4 ) . Total RNA was isolated and subjected to Northern blot analysis as described in the legend to Fig. 3. c-fos mRNA transcripts were detected by hybridization with a 1.1-kb PstI v-fos fragment. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.   (Table 11). Additive as well as synergistic interactions between growth factors could be observed. The combination bFGF/EGF additively induces DNA synthesis as compared with the effect of the respective growth factors alone. In contrast, the combination PDGF/ IGF-1 acts slightly, the combinations EGF/IGF-1, PDGF/ EGF, and bFGF/IGF-1 are strongly synergistic on DNA synthesis of SMC, with bFGF/IGF-1 being the most potent combination. Thus our results indicate that synergistic interactions between growth factors are not restricted to the system bFGF/PDGF but seem to be a more general principle in cellular growth regulation.  Kanakarij et al., 1991). The dimeric PDGF-Raa binds all PDGF isoforms, dimeric PDGF-Rap binds PDGF AB and BB, and PDGF-RPP recognizes only PDGF BB. Thus PDGF AA binds only to the dimeric PDGF-Raa.
In cultured smooth muscle cells the expression of PDGF-R a is low compared with the expression of PDGF-R@ (Ferns et al., 1991a). As a consequence the number of available dimeric PDGF-Raa is also low (Sachinidis et al., 1990). This may explain the low mitogenic activity of PDGF AA for this cell type. We were able to confirm the low expression of PDGF-Ra as compared with PDGF-RP and the poor mitogenic potential of PDGF AA for the cultured SMC used for the experiments described here. However, we could demonstrate that this low mitogenic potential of PDGF AA was significantly increased by simultaneous application of bFGF.
Our results show that bFGF increases PDGF-Ra mRNA levels without effecting the amount PDGF-RP mRNA. The elevated PDGF-Ra transcript levels were translated into an increase of functional PDGF-Ra appearing at the surface of the SMC. Upon stimulating with bFGF, SMC reveal a substantial increase of immunoprecipitable PDGF-Ra with no change observed for the PDGF-RP. Binding experiments show that these findings coincide with a bFGF-induced significant increase of the number of binding sites for ""I-PDGF AA. Furthermore, upon incubation with bFGF, SMC become responsive to PDGF AA with respect to the transcription of the proto-oncogene c-fos. Thus the enhanced mitogenic potential of PDGF AA for SMC by simultaneous application of bFGF can be explained by the bFGF-dependent expression of functional PDGF-Ra. bFGF also enhances the responsiveness of SMC to PDGF AB and BB, however, to a much smaller extent. This observation can also be explained by the selective increase of the PDGF-Ra which is reflected by an increase of the number of binding sites for ""I-PDGF AB and BB, both being able to associate with dimeric PDGF-Raa and PDGF-Rap.
The synergistic effect of bFGF and PDGF described above could also be demonstrated for other combinations of growth factors. EGF was highly synergistic with IGF-1 and PDGF AB and bFGF together with IGF-1. In contrast, the combination bFGF and EGF was only additive with respect to the mitogenic potential of the individual growth factor. Whether these synergistic effects can also be explained by growth factor-dependent transmodulation of receptor expression remains to be clarified.
There are several reports showing transmodulation of receptor expression by other growth factors in other cell types as well. PDGF BB was found to up-regulate IL-1 receptor expression in Balb/c 3T3 fibroblasts (Chiou et al., 1989). IL-1 as well as other mitogenic factors enhance the expression of functional PDGF-Ra in an osteoblast-like cell line (Tsukamoto et al., 1991). TGF-/3 inhibits the expression of PDGF-RG in SMC (Battegay et al., 1990) and in Swiss mouse 3T3 fibroblasts (Gronwald et dl., 1989), whereas PDGF-R@ expression in Swiss mouse 3T3 fibroblasts is enhanced (Gronwald et al., 1989). bFGF up-regulates PDGF-Ra expression in 0-2A progenitor cells (McKinnon et al., 1990). It was also reported that retinoic acid leads to an enhanced expression of PDGF-Ra (Wang et al., 1990). Furthermore increased CAMP levels in rat Schwann cells caused up-regulation of PDGF-R/3 expression (Weinmaster and Lemke, 1990).
Thus transmodulation of receptor expression by growth factors seems to be a general phenomenon which is important to control coordinated growth and differentiation but which might also be involved in pathophysiological situations such as development of atherosclerosis. Majesky et al. (1990) showed that the PDGF A chain was strongly expressed in SMC along the luminal surface of a carotid artery, injured by balloon catheterization. In addition both PDGF-Ra and PDGF-RP were elevated (Majesky et al., 1990;Ferns et al., 1991b). SMC in human atherosclerotic plaques also express substantial amounts of PDGF A chain (Wilcox et al., 1991).
On the other hand bFGF was detected in large amounts in vascular smooth muscle cells and in the subendothelial basement membrane (Cordon-Cardo et d., 1990;Schulze-Osthoff et al., 1990). bFGF bound to the components of the ECM is proposed to be mobilized by remodeling of the basement membrane or the ECM by hydrolases or after injury of the endothelial cells (Vlodavsky et al., 1987).
Thus bFGF, in addition of being a mitogen for SMC itself, may up-regulate PDGF-Ra expression and may render the SMC responsive to PDGF AA. This is in agreement with results reported by Lindner et al. (1991) who could show that the first wave of SMC replication in balloon-injured vessels could be blocked by bFGF-neutralizing antibodies. Interestingly, FGF has been shown to induce PDGF A gene expression in SMC (Winkles and Gay, 1991). This observation supports the hypothesis that FGF, under conditions when it is available, may trigger an autocrine loop via the PDGF-Ra. In addition, the two isoforms PDGF AB and PDGF BB released from platelets are likely to mediate chemotactic migration of SMC and to contribute to the accumulation of neointimal SMC (Ferns et al., 1991b). The results presented in this paper suggest that transmodulation of PDGF-Ra expression by bFGF might be one of the mechanisms involved in neointimal formation and finally may participate in the development of atherosclerosis.