Induction of Platelet-derived Growth Factor a- and &Receptor mRNA and Protein by Platelet-derived Growth Factor BB*

We describe that stimulation of human fibroblasts with platelet-derived growth factor BB (PDGF-BB) induces a transient up-regulation of the PDGF a- and &receptor transcript and protein levels. The effect of PDGF-BB on the receptor transcript levels was more pronounced than those seen when other cytokines were used. Regulation of transcript levels by PDGF-BB was mediated through post-transcriptional mechanisms. No induction could be observed in a nuclear run-on analysis, but cycloheximide treatment attenuated the accumulation of both a- and &receptor transcripts induced by PDGF-BB. An increase in receptor protein levels was observed using two different experimental approaches. Increased amounts of receptor precursor forms could be immunoprecipitated from metabolically labeled cells, stimulated with PDGF-BB. In a second approach, cells were exposed to different concentrations of PDGF-BB, and, in a subsequent step, ligand binding analysis was performed. In this experiment, an initial down-regulation of receptors was followed by increased levels of the cell surface forms of the receptors. In conclusion, PDGF-BB, but not PDGF-AA, induces increased synthesis of both PDGF a- and 8-receptor protein; this constitutes a positive feed-back mechanism, which, for example, could serve to poten- tiate autocrine stimulation of growth.

cDNAs for both receptors have been isolated and characterized (9,10,(12)(13)(14). The two receptor proteins are organized in a similar fashion, each having five immunoglobulin-like domains in the extracellular part and a split tyrosine kinase effector domain intracellularly. The amino acid identity between the two receptors varies from 30% in the extracellular part to 87% in the most N-terminal part of the tyrosine kinase domain. Binding of PDGF to the extracellular part of either receptor type leads to dimerization of receptor molecules, followed by activation of the receptor tyrosine kinase (15)(16)(17)(18). As a consequence of activation and autophosphorylation, substrates for the receptor tyrosine kinase become stably associated with the receptors. Substrates for the PDGF Breceptor identified thus far include phospholipase C-y, GTPase activating protein, phosphatidylinositol-3'-kinase, cruf, and members of the src family (for a review, see Ref. 19). Phosphorylation of these, as well as of hitherto unidentified substrates, will eventually lead to induction of a number of genes; some are induced already a few minutes after receptor activation, others after several hours (20). Certain of these genes code for transcription factors like c-fos and c-myc, others code for factors that are not likely to be directly involved in triggering the growth response, like matrix or cytoskeletal components.
Positive feedback mechanisms in the growth response have been implicated by the finding that PDGF induces the expression of the A-chain of PDGF in a transient fashion (21). Several other growth modulating factors like EGF, IL-la, and TGF-P have also been found to induce the A-chain in certain cell types (22)(23)(24)(25); some of these factors have also been reported to induce the B-chain (22,26). Here we present additional evidence for positive feedback mechanisms in the prereplicative phase of human fibroblasts and demonstrate that stimulation with PDGF-BB is followed by enhanced accumulation of both PDGF a-and P-receptor mRNA as well as increased synthesis of both receptor proteins.
Research, Uppsala, respectively. Human recombinant IL-la was purchased from Genzyme Corp.
mRNA Preparation and Northern Blot Hybridization-RNA was prepared using the LiCl/urea method (28) by removing culture medium and disrupting the cells directly in the culture flasks by adding 5 + 2 ml per 175-cm2 flask of a solution composed of 3 M LiCl, 6 M urea, 0.2% SDS, 0.1% Antifoam A (Sigma). Total RNA was denatured using formamide/formaldehyde at 55 "C for 15 min, whereafter 20 pg was applied per lane and run in 1% agarose gels, containing 2.2 M formaldehyde, and blotted to nitrocellulose filters (29). The filters were baked at 80 "C under vacuum for 2 h before hybridization at 42 'C in 50% formamide, 5 X SSC (1 X ssc = 15 mM sodium citrate, pH 7.0, 150 mM NaCl), 5 X Denhart's solution (30), 0.1% SDS, 0.1 mg/ml salmon sperm DNA. The following receptor probes were used a 1.5-kilobase pair EcoRI-PstI fragment of the human PDGF areceptor cDNA (9) and a 0.7-kilobase pair PstI fragment of the human PDGF 8-receptor cDNA (10); the fragments were 32P-labeled by the Multiprime DNA labeling system (Amersham) and added at 1 X lo6 cpm/ml. Hybridization was also performed with a 32P-labeled glyceraldehyde-3-phosphate dehydrogenase cDNA probe (31) (obtained from Dr. R. Wu) added at 2 X lo5 cpm/ml, in order to assess the amount of RNA loaded in each lane. The filters were washed 3 X 20 min in 0.1% SSC, 0.1% SDS at 55 "C.
Proteinase K buffer (120 pl) containing 1.25% SDS, 25 mM EDTA, 25 mM Tris-HC1, pH 7.5, was added together with Proteinase K (200 pg), and the incubation was prolonged at 37 "C for 30 min. The samples were then phenol-extracted, precipitated, and redissolved in 100 pl of 10 mM Tris-HC1, pH 7.6, 40 mM MgClz to which 75 units of DNase I was added and incubated at 30 "C for 10 min. The samples were phenol-extracted again, precipitated twice in 1.5 M ammonium acetate, 70% ethanol, and resuspended in 10 mM Tris-HC1, pH 7.5, 1 mM EDTA. Equal amounts of 32P radioactivity were added to hybridization buffer (see "mRNA Preparation and Northern Blot Hybridization") containing 200 pg/ml salmon sperm DNA and 50 pg/ml Escherichia coli tRNA. Hybridization to denatured plasmid DNAs (10 pg) slot-blotted to nitrocellulose filters was performed at 42 "C for 40-64 h. The filters were washed for 10 min in 2 X SSC, 0.2% SDS at room temperature, followed by three washes for 30 min each in 2 X SSC, 0.2% SDS at 60 "C. The filters were air-dried and exposed to x-ray film at -70 "C. The plasmids used were: a 750-base pair fragment of the PDGF 8-receptor cDNA corresponding to the transmembrane part and the first part of the tyrosine kinase domain (10) in pUC19 and a 750-base pair fragment of the PDGF a-receptor cDNA corresponding to the tyrosine kinase part (9) in pUC19. A genomic c-fos plasmid, (pc-fos(human)-1), obtained from Dr. R. Muller, from which the coding sequences had been removed using NcoI and XhoI (33), served as a control that equal amounts of 32P radioactivity were applied to each filter.
Antibodies-The polyclonal rabbit antiserum PDGFR-1 was raised against purified porcine PDGF receptor preparations (34) and has previously been shown to recognize both PDGF a-and @-receptors (35). The rabbit peptide antiserum PDGFR-3 was raised against the synthetic peptide (QPNESDNDYIIPLP) located in the C terminus of the murine PDGF &receptor (12) and has been shown to be specific for the PDGF @-receptor (35). specific activities > 800 Ci/mmol and 600 Ci/mmol, respectively), was added. Cells were thereafter put on ice and washed once with icecold phosphate-buffered saline before lysis in 0.5% Triton X-100 (Merck), 0.5% deoxycholate (Sigma) in 20 mM Tris-HC1, pH 7.5, 150 mM NaCl, 10 mM EDTA, 1% Trasylol (Bayer, Leverkusen), and 1 mM phenylmethylsulfonyl fluoride (Sigma). The lysates were cleared by centrifugation at 10,000 X g for 10 min, and the supernatants were passed over a Lens culinaris lectin affinity column, previously equilibrated in lysis buffer. After washing the column, glycoproteins were specifically eluted with lysis buffer supplemented with 10% (w/v) a-D-mannoside (Sigma). Immunoprecipitation was performed sequentially by first adding 10 p1 of PDGFR-3; the samples were then incubated for 2-12 h at 4 "C. Protein A-Sepharose CL-4B (Pharmacia LKB Biotechnology Inc.) was used to precipitate the immune complexes. The beads were spun down after 30 min of end-over-end incubation. Supernatants were cleared on an extra round with protein A-Sepharose before addition of 10 p1 of PDGFR-1 antiserum. The precipitates were washed three times with RIPA buffer containing 1% Triton X-100, 1% deoxycholate, 0.1% SDS, 50 mM Tris-HC1, pH 7.5, 150 mM NaCl, 10 mM EDTA, and once with HzO. Immune complexes were eluted by heating the beads at 95 "C for 4 min in sample buffer containing 4% SDS, 0.2 M Tris-HC1, pH 8.8, 0.5 M sucrose, 5 mM EDTA, 0.01% bromphenol blue, and 2% 2-mercaptoethanol (Carl Roth, Karlsruhe, FRG).
Iz5I-PDGF BindirtP A~~~V -~~~I -P D G F -B B was labeled accordine to Bolton and Hunte; (37) specific activities of 14,000-18,000 cpm? ng and lZ5I-PDGF-AA by the chloramine T method (38) to 41,000-56,000 cpm/ng. The binding assay was performed using 35-mm dishes incubated in a water bath kept at 37 "C. Cultures were washed twice with MCDB 104 containing 20 mM Hepes, pH 7.5, and 1 mg/ml human serum albumin at 37 "C. Iodinated PDGF-BB or PDGF-AA (1 or 2 ng/ml as described in each figure legend) was then added to the cultures in 1 ml of the same buffer for 5 min at 37 "C, whereafter the dishes were immediately placed on ice and washed 5 times in icecold MCDB 104 supplemented as above. Cells were lysed in 500 p1 of 1% Triton X-100, 20 mM Hepes, pH 7.4, 10% glycerol, and 0.1 mg/ ml human serum albumin at room temperature, and cell-associated radioactivity was determined in a y spectrometer. Parallel cell cultures incubated and washed as above were trypsinized, and cells were counted in an electronic cell counter.
In order to determine the functional PDGF receptor down-modulating activity remaining in the medium at different time points after addition of the unlabeled recombinant PDGF-BB, AG1518 cell cultures serum-starved as above were incubated for 0-24 h with 20 ng/ ml PDGF-BB. The conditioned media from these cultures were then added to another set of serum-starved AG1518 cell cultures that were further incubated for 1 h at 37 "C before the binding assay described above was performed. The remaining PDGF receptor down-modulating activity was compared to that of different concentrations of PDGF-BB (2,5,10,20,50, and 100 ng/ml) incubated under the same conditions, i.e. in preconditioned medium from separate fibroblast cultures which were transferred onto fresh AG1518 cells. The remaining PDGF-receptor down-modulating activity is given as PDGF-BB equivalents (ng/ml) and represents the mean of two different experiments.

Effects of Different Growth Regulatory Factors on PDGF aand @-Receptor Transcript Levels-Serum-starved AG1518
human foreskin fibroblasts were incubated in the presence of different growth regulatory factors such as EGF, TGF-@, ILla, fetal calf serum, PDGF-AA, or PDGF-BB for 0, 3, and 8 h. Total cellular RNA was isolated. After electrophoresis and transfer to nitrocellulose, hybridization was performed with 32P-labeled PDGF a-and @-receptor cDNA probes. As seen in Fig. 1, A and B, the added factors affected PDGF a-and @-receptor mRNA levels in different ways. TGF-@ downmodulated the a-receptor mRNA level, but left the @-receptor transcript unaffected. These data are largely in agreement with what has been found by others using murine Swiss 3T3 fibroblasts (39, 40). The 8-receptor transcript levels were relatively stable, whereas the a-receptor mRNA levels underwent larger fluctuations in response to the different stimuli. PDGF-BB induced a marked increase in both PDGF a-and @-receptor mRNA levels; this was the most pronounced effect observed, in comparison with the responses to the other investigated cytokines. We therefore focused on this factor in our further studies. Analysis of PDGF a-and 8-Receptor mRNA Leveki after PDGF-BB Stimulation-The kinetics of the effect of PDGF-BB on PDGF a-and @-receptor mRNA levels was investigated (Fig. 2, A and B). An increase in both a-and 8-receptor mRNA was detected within a few hours after the addition of PDGF-BB, reaching a maximum between 6 and 10 h. From the plot in Fig. 2B, which shows the relative PDGF a-and @receptor mRNA levels after correction for differences in sample loading, it is apparent that the effect was more pronounced on the a-receptor than on the @-receptor transcript level.
The protein translation inhibitor cycloheximide has been shown to have different effects on induced mRNAs. Some mRNA species accumulate in the presence of cycloheximide, whereas the induction of other mRNAs is blocked by the drug (20). When cycloheximide was added for 6 h together with PDGF-BB to AG1518 fibroblasts, the induction of either PDGF receptor mRNAs was attenuated (Fig. 3). This suggests a need for newly synthesized proteins for the induction of both PDGF a-and 8-receptor mRNAs by PDGF-BB. Addi-

FIG. 3. Up-regulation of PDGF CY-and B-receptor mRNA by
PDGF-BB is inhibited by cycloheximide. Total RNA was isolated from AG1518 fibroblasts cultured for 6 h in the presence or absence of 20 ng/ml PDGF-BB and with or without 10 pg/ml cycloheximide. After electrophoresis and transfer to nitrocellulose, hybridization was performed with PDGF a-and 0-receptor specific fragments. The blot was analyzed by autoradiography.
tion of cycloheximide alone had no effect on the receptor transcript levels.
Transcriptional Analysis of PDGF a-and @-Receptor Genes after PDGF-BB Stimulation-To gain further insight into the mechanism by which the alterations in PDGF receptor expression occur, nuclear RNA run-on assays were performed. Fibroblasts were incubated in the presence or absence of PDGF-BB for 6 h, and nuclei were isolated. The nuclei were incubated with [cx-~*P]UTP, and RNA was purified and hybridized to a-and @-receptor cDNAs immobilized on nitrocellulose filters. As seen in Fig. 4, only a marginal difference in the initiation of receptor RNA was found upon PDGF exposure for 6 h, as compared with unstimulated samples. The specificity of the hybridization at the conditions used is shown by lack of signal to the vector, pUC19, and the equal intensity of hybridization to genomic c-fos sequences reflects that equal amounts of "P radioactivity were applied to the filters. These results indicate that the changes evoked by PDGF-BB on the expression of PDGF a-and P-receptors cannot be attributed to increased levels of transcription.
Effect of PDGF-BB on PDGF a-and P-Receptor Protein Synthesis-To determine whether the PDGF-BB-induced increase in the levels of mRNA for the PDGF receptors was followed by an increase in receptor proteins, AG1518 human foreskin fibroblasts were treated with PDGF-BB for various periods of time. During the last hour of stimulation, [35S] methionine and [35S]cysteine were present at 100 pCi/ml. Lysates of the labeled cells were immunoprecipitated sequentially, first using PDGFR-3, specific for the PDGF P-receptor, and then using PDGFR-1, which recognizes both receptor types. As seen in Fig. 5, A and B, addition of 20 ng/ml PDGF-BB resulted in an increase in the level of both P-receptor (160 kDa) and a-receptor (140 kDa) protein precursor forms, which could be detected already after 2 h of incubation with PDGF-BB. The increase in the amount of a-and P-receptor precursor was 1.5-2-fold and 2-3-fold, respectively, based on densitometric scanning (Fig. 5C). Mature forms of the receptor proteins were detected to a much less extent than the precursors in this analysis, due to the short labeling time and to the rapid internalization and degradation of the cell surface expressed receptors in the presence of ligand.

Effect of PDGF on Cell Surface Levels of PDGF a-and P-
Receptors-In order to estimate the accumulation of cell surface receptors, ligand binding analysis on PDGF-BB-stimulated cells was performed. Incubation of AG1518 cells with 2, 20, or 100 ng/ml PDGF-BB induced a rapid down-modulation of lZ5I-PDGF-BB and lZ5I-PDGF-AA binding (Fig. 6, A and  B). After 8 h of incubation, a gradual increase in cell surface expressed receptors was seen on cells incubated with 20 ng/ ml PDGF-BB. This increase was less apparent in the presence of 2 ng/ml PDGF-BB, and it was concealed to a considerable extent, by the excess of ligand, when 100 ng/ml was used. Incubation with 20 ng/ml PDGF-AA induced a persistent min at 37 "C. After washing, cell-associated lZ5I radioactivity was determined. Parallel cultures incubated and washed as above were trypsinized and counted. C, conditioned media from AG1518 cultures treated with PDGF-BB (20 ng/ml) for 0-24 h were added to a new set of serum-starved cells and incubated for 1 h at 37 "C. In parallel, cells were incubated for 1 h with different concentrations of PDGF-BB. Binding of "'1-PDGF-BB for 5 min at 37 "C was performed to these cultures as described under "Experimental Procedures," and the amount of PDGF-BB remaining in the conditioned media was calculated from the obtained standard curve. The result represents the mean of two different experiments.
down-modulation of '251-PDGF-AA binding; no increase in the amount of cell surface expressed a-receptors were seen even after 24 h of incubation (Fig. 7). Furthermore, PDGF-AA did not affect lZ5I-PDGF-BB binding, which in human foreskin fibroblasts occurs preferentially to the @-receptor (41) (Fig. 7). To establish that the recovery of cell surface binding sites, starting at 8 h after addition of 20 ng/ml PDGF- AG1518 cells were exposed to 20 ng/ml PDGF-AA for 0-24 h at 37 "C. Cells were washed with medium, and binding of lZ5I-PDGF-BB (1 ng/ml) (B) or 9-PDGF-AA (1 ng/ml) (0) was performed for 5 min at 37 "C. After washing, cell-associated lZ5I radioactivity was determined. Parallel cultures treated as above were trypsinized and counted.
BB, was not due to depletion of ligand, the amount of PDGF-BB left in the conditioned media was determined. Thus, the ability of the conditioned media to down-modulate lZ5I-PDGF binding on fresh AG1518 cultures was compared with the degree of down-modulation seen with known concentrations of PDGF-BB. Fig. 6C shows that the down-modulating activities remaining at 8, 12, and 24 h were equivalent to 9, 10, and 6 ng/ml PDGF-BB, respectively. The fact that as much as about 50% of the added PDGF-BB remained in the medium after 8 and 12 h indicates that the recovery of lZ5I-PDGF-AA and '251-PDGF-BB binding, initiated at 8 h, is a result of the stimulatory effect of PDGF-BB on receptor synthesis and not only an effect of decreased PDGF-BB concentration in the medium during the incubation period. Moreover, parallel experiments showed that PDGF-BB at 10 ng/ml caused a pattern of down-modulation of @-receptors and a subsequent increase in expression levels, very similar to that seen for PDGF-BB at 20 ng/ml (data not shown).

DISCUSSION
This report describes the transient up-regulation of PDGF a-and @-receptor transcripts and proteins in human foreskin fibroblasts, in response to PDGF-BB. Stimulation with other growth factors or cytokines, for which there are receptors on fibroblasts, had only a marginal effect on the up-regulation of PDGF receptor transcript levels, as compared with PDGF-BB. Thus, we could not register any up-regulation of PDGF a-receptor after exposure of the cells to PDGF-AA. In fact, stimulation of the cells with PDGF-AA, which binds to the a-receptor with high affinity, but not to the @-receptor, resulted in a small decrease of a-as well as @-receptor mRNA levels. Our data do not allow a firm correlation as to whether PDGF-BB exerted its stimulatory effect via activation of areceptors or via a transduction mechanism mediated by breceptors. However, the fact that PDGF-AA treatment did not lead to increased levels of the a-receptor transcripts favors the notion that PDGF-BB transduced its stimulatory effect on the a-and @-receptors via the @-receptor.
As seen when comparing Figs. 2 A , 2B, 5, 6A, and 6B, the kinetics of response in PDGF receptor levels to PDGF-BB stimulation varied slightly depending on which experimental approach was used. The elevation of PDGF receptor mRNA levels peaked at 8 h and were down to prestimulation levels at 24 h after onset of stimulation. On the protein level, as determined in a metabolic labeling experiment, precursor forms of the receptors started to accumulate already a few hours after onset of stimulation. When the accumulation of cell surface expressed forms of the receptors were analyzed, a gradual increase was recorded at 8 h and reached its plateau at 12 h. The evaluation of the kinetics of the response in the different experiments is complicated by the fact that exposure of cells to PDGF-BB has many various effects. One such effect is the increased transport rate of newly synthesized receptors after ligand stimulation.' On the other hand, the receptors present on the cell surface at the point when the ligand is added will rapidly be internalized and degraded. This down-modulation is counteracted by the gradual increased synthesis of receptors, resulting in a net increase in the level of cell surface expressed receptors seen after PDGF-BB exposure (Fig. 6, A and E ) .
Results from cycloheximide treatment and nuclear run-on analysis indicated that the mechanisms behind the increased accumulation of PDGF receptor transcripts occurred mainly through post-transcriptional regulation and only to a minor extent by an accelerated rate of transcription. We cannot exclude that cycloheximide treatment attenuated the PDGF-BB-induced receptor transcript accumulation through a cycloheximide-induced net decrease in the number of cell surface PDGF receptors. However, since PDGF-BB treatment in itself leads to a rapid internalization of cell surface receptors (see e.g. Fig. 6, A and E ) , it is likely that the transduction of the stimulatory effect on the receptor transcript levels was initiated early after the addition of PDGF-BB. The possibilities for post-transcriptional regulation include, e.g. RNA processing, transport out of the nucleus, and RNA stability. The rate of degradation of transcripts appears to be mediated mainly through signals within the 3"untranslated (UT) region. Such signals have been identified as motifs consisting of AUUUA-or AU-rich sequences, which are frequently found in the 3"UT region of post-transcriptionally regulated transcripts (42). A survey of the available, incomplete 3'-UT cDNA sequences for the PDGF receptors reveal one AUUUA motif in the @-receptor and as many as seven in the a-receptor. It is, however, poorly understood exactly how these motifs mediate regulation of transcript stability. Thus, it remains to be shown whether the accumulation of PDGF receptor transcripts is due to increased mRNA stability, as a consequence of PDGF-induced activation of a labile, cycloheximide-sensitive stabilizing factor.
The expression levels of the PDGF a-and @-receptors have previously been shown to be regulated in a differential manner by exposure crf cells to various factors. The a-receptor undergoes down-modulation in Swiss mouse 3T3 fibroblasts, both on the protein and transcript levels, in response to TGF-@ (39, 40). Activated m u tyrosine kinase can down-modulate both PDGF a-and /3-receptor protein expression via decreased mRNA level^.^ Elevation of the CAMP levels in Schwann cells by forskolin treatment has been shown to induce stable expression of the @-receptor (43). Autoligand stimulation, in a fashion similar to what has been described here, has been shown to occur for other growth factor receptors. Stimulation by EGF leads to increased synthesis of EGF receptor transcript and protein, with a peak at 4-8 h after onset of stimulation (44,45).
Growth control is exerted through positive, as well as negative, feedback loops. Exposure of sparse cultures of human L. Claesson-Welsh, unpublished observations. Lehtola, L., NistBr, M., Holtta, E., Westermark, B., Alitulo, K. (1991) Cell Regul., in press. foreskin fibroblasts to PDGF has been shown to lead to transiently increased levels of TGF-B transcripts (24). TGF-@ has been shown to be a growth inhibitor for human fetal fibroblasts (46), NRK fibroblasts (47), and murine Swiss 3T3 fibroblasts (40); the inhibition is at least in part exerted through down-regulation of the PDGF a-receptor (39,40). On the other hand, PDGF stimulation of human foreskin fibroblasts has been shown to induce transient PDGF A-chain expression, which could lead to an amplification of the initial growth stimulus (21). The transient up-regulation of both PDGF receptor types in response to PDGF-BB could reflect an additional positive feedback mechanism. In cells expressing PDGF-BB as well as the PDGF receptor, the capacity for autocrine stimulation could thereby be potentiated. This could partly explain why overexpression of PDGF-BB in NIH 3T3 cells or Rat-1 cells gives a stronger transforming effect than overexpression of PDGF-AA (48, 49). Of relevance in this context is the recent finding that there is a high incidence of simultaneous expression of transcripts for the PDGF @receptor and PDGF B-chain in biopsies of benign fibroblastderived human skin tumors4 as well as in vascular proliferations in human malignant glioma (50).