Preferential Inhibition of Platelet-derived Growth Factor-stimulated DNA Synthesis and Protein Tyrosine Phosphorylation by Nordihydroguaiaretic Acid*

Nordihydroguaiaretic acid (NDGA), a reportedly spe- cific lipoxygenase inhibitor, was found to selectively inhibit platelet-derived growth factor (PDGF)-stimulated DNA synthesis in Swiss 3T3 cells. Maximal inhibition of PDGF-induced [SH]thymidine incorporation (96%) was observed using 4 w NDGA (ICso = 1.5 PM). No effect of NDGA was observed upon DNA synthesis stimulated with either fetal bovine serum, bombesin, or epidermal growth factor (EGF) in the presence of insulin, or with the potent mitogen Pasteurella multocida toxin. The se- lective inhibition of PDGF-stimulated DNA synthesis by NDGA was also observed in diploid murine cells, rat, and human fibroblasts. Furthermore, 4 1.1~ NDGA also inhibited PDGF-stimulated anchorage-independent colony growth of rat-1 cells by 76%. Using Swiss 3T3 cells, we found that PDGF-stimulated arachidonic acid mobilization and prostaglandin Ez production was abolished by NDGA in a dose-dependent manner. Inhibition of PDGF-stimulated arachidonic acid mobilization by NDGA could not, however, explain its potent inhibitory effect upon PDGF-stimulated DNA synthesis. Our results showed that NDGA also selectively inhibited PDGF receptor tyrosine phosphorylation in a dose- dependent manner in intact cells. Protein tyrosine phosphorylation stimulated by EGF or bombesin was not altered by NDGA treatment. Crucially, NDGA inhibited in vitro the tyrosine kinase activity of anti-phosphoty-rosine and anti-PDGF receptor immunoprecipitates prepared from cultures stimulated with PDGF. This inhibition of receptor tyrosine phosphorylation in a cell- free system confirmed that NDGA acts directly at the level of the PDGF receptor tyrosine kinase domain. These results suggest that the potent and selective inhibitory effect of NDGA on PDGF-stimulated DNA syn- thesis results from its inhibitory action on tyrosine phosphorylation. PBS containing 3% non-fat milk as described (36). Immunoreactive bands were visualized using 1251-labeled sheep anti-mouse IgG followed by autoradiography. Assay ofprotein Kinase Activity-Immunoprecipitates prepared from 1.5 x lo6 cells as described above were washed three times with lysis buffer and twice with 50 m~ HEPES, pH 7.4,O.l m~ EDTA, 0.01% Brij, 75 m~ NaCl, and 100 Na3V0, (kinase assay buffer). NDGA was added to the immunoprecipitates in 1 ml of kinase assay buffer, and the mixture was incubated for 15 min at 4 "C. After this time, the immu- noprecipitates were resuspended in 20 pl of kinase buffer, and the reaction was initiated by the addition of 10 m~ MgClz and 100 p~ [y-3zPl ATP (10 pCi) in a total volume of 30 pl at 30 "C. ARer a 10-min incubation, immunoprecipitates were washed twice with lysis buffer and analyzed by SDS-PAGE followed by autoradiography.

Preferential Inhibition of Platelet-derived Growth Factor-stimulated DNA Synthesis and Protein Tyrosine Phosphorylation by Nordihydroguaiaretic Acid* (Received for publication, October 20, 1993) J a n Domin,Theresa Higgins,and Enrique RozengurtS From the Imperial Cancer Research Fund,44 Lincoln's Inn Fields,London WC2A 3PX,United Kingdom Nordihydroguaiaretic acid (NDGA), a reportedly specific lipoxygenase inhibitor, was found to selectively inhibit platelet-derived growth factor (PDGF)-stimulated DNA synthesis in Swiss 3T3 cells. Maximal inhibition of PDGF-induced [SH]thymidine incorporation (96%) was observed using 4 w NDGA (ICso = 1.5 PM). No effect of NDGA was observed upon DNA synthesis stimulated with either fetal bovine serum, bombesin, or epidermal growth factor (EGF) in the presence of insulin, or with the potent mitogen Pasteurella multocida toxin. The selective inhibition of PDGF-stimulated DNA synthesis by NDGA was also observed in diploid murine cells, rat, and human fibroblasts. Furthermore, 4 1.1~ NDGA also inhibited PDGF-stimulated anchorage-independent colony growth of rat-1 cells by 76%. Using Swiss 3T3 cells, we found that PDGF-stimulated arachidonic acid mobilization and prostaglandin Ez production was abolished by NDGA in a dose-dependent manner. Inhibition of PDGFstimulated arachidonic acid mobilization by NDGA could not, however, explain its potent inhibitory effect upon PDGF-stimulated DNA synthesis.
Our results showed that NDGA also selectively inhibited PDGF receptor tyrosine phosphorylation in a dosedependent manner in intact cells. Protein tyrosine phosphorylation stimulated by EGF or bombesin was not altered by NDGA treatment. Crucially, NDGA inhibited in vitro the tyrosine kinase activity of anti-phosphotyrosine and anti-PDGF receptor immunoprecipitates prepared from cultures stimulated with PDGF. This inhibition of receptor tyrosine phosphorylation in a cellfree system confirmed that NDGA acts directly at the level of the PDGF receptor tyrosine kinase domain. These results suggest that the potent and selective inhibitory effect of NDGA on PDGF-stimulated DNA synthesis results from its inhibitory action on tyrosine phosphorylation.
PDGFl is a potent mitogen for cells of mesenchymal origin and has been implicated in wound healing, development, and inflammation in addition to the etiology of many disease processes including atherosclerosis, rheumatoid arthritis, and oncogenesis (1,2). PDGF is a disulfide-linked dimer of two related * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence and reprint requests should be addressed. Tel.: 071-269-3454; Fax: 071-269-3094.
The abbreviations used are: PDGF, platelet-derived growth factor; DMEM, Dulbecco's modified Eagle's medium; EGF, epidermal growth factor; FBS, fetal bovine serum; NDGA, nordihydroguaiaretic acid; PAGE, polyacrylamide gel electrophoresis; PGEz, prostaglandin Ez; TMEF, tertiary mouse embryo fibroblasts. polypeptide chains which are assembled either as homodimers or as a heterodimer. Binding of PDGF to its cell surface receptor causes receptor dimerization and transphosphorylation of specific residues along the PDGF receptor polypeptide chain (1,3,4). These phosphorylated tyrosine residues serve as attachment sites for intracellular proteins, an interaction mediated through SH2 domains present in the effector proteins (5). Substrates for the PDGF receptor include phospholipase Cy, the ras GTPase activating protein, the p85 regulatory subunit of phosphatidylinositol 3'-kinase, members of the pp60"" family of protein tyrosine kinases, Nck, GRBB, and the phosphotyrosine phosphatase Syp (6)(7)(8)(9). Once bound to the receptor, many of these substrates are themselves phosphorylated on tyrosine residues by the receptor kinase activity. PDGF thus triggers a diverse array of downstream early signaling events (10)(11)(12)(13)(14). Despite the role of PDGF in a variety of disorders characterized by excessive cell proliferation and the increasing understanding of the molecular events involved in PDGFstimulated signal transduction, few pharmacological agents exist which selectively inhibit the ability of PDGF to induce DNA synthesis.
The proliferation of Go-arrested cells can be triggered by multiple signal transduction pathways that act in a combinatorial and synergistic fashion (11,15). The release of arachidonic acid from the sn-2 position of membrane phospholipids is increasingly implicated as one of the signals involved in this process (16)(17)(18). PDGF stimulates a striking biphasic mobilization of arachidonic acid in Swiss 3T3 cells (19). The mechanisms by which arachidonic acid participates in mitogenic signal transduction, however, remain unclear. Arachidonic acid could either act directly as a second messenger (20)(21)(22)(23) or alternatively serve as the substrate for the production of a variety of biologically active eicosanoids (24). To evaluate the action of eicosanoids in many biological systems, specific inhibitors of both cyclooxygenase and lipoxygenase enzymes have been utilized. NDGA, a reportedly specific lipoxygenase inhibitor (25-27), has been used extensively to test the role of the lipoxygenase pathway in the action of many growth factors and cytokines including EGF, basic fibroblast growth factor, and tumor necrosis factor (28-31).
In the present study, we have examined the ability of NDGA to attenuate PDGF-stimulated DNA synthesis in an effort to determine the contribution of leukotriene biosynthesis to this process. We found that this phenolic plant lignan dramatically inhibited PDGF-stimulated DNA synthesis in a selective manner. Surprisingly, the mechanism by which this effect is achieved involves inhibition of PDGF receptor protein tyrosine kinase activity.

EXPERIMENTAL PROCEDURES
Cell Culture-Stock cultures of Swiss 3T3 cells were propagated as previously described (32). For experimental purposes, cells were sub-8260 cultured in 33-or 90-mm Nunc dishes with DMEM supplemented with 10% FBS, 100 unitslml penicillin, and 100 pg/ml streptomycin. Cultures were incubated in a humidified atmosphere of 10% C02, 90% air at 37 "C. Cells were rendered quiescent by incubation under these conditions for 6-8 days prior to use.
Tertiary passage mouse embryo fibroblasts (TMEF), rat-1 cells, and human foreskin fibroblasts used at the 18th passage were seeded at lo5 cells per 33-mm dish in D~M supplemented with 1OYo FBS. After 3 days, the cultures were switched to DMEM containing 0.5% FBS. The cultures were confluent and quiescent 4-6 days later. PHIThymidine Incorporation Assay-Determination of DNA synthesis was performed as previously described (33). Briefly, cultures were washed twice with DMEM and incubated with D M E~a~o u t~s medium (1:l viv) containing K3Hlthymidine (1 pcitml; 1 w) and various additions as described. After 40 h, the cultures were washed twice with phosphate-buffered saline and incubated in 5% trichloroacetic acid for 30 min at 4 "C. Trichloroacetic acid was then removed, and the cultures were washed twice with ethanol and extracted in 1 ml of 2% Na2C03, 0.1 M NaOH, 1% SDS. Incorporation was determined by scintillation counting in 10 ml of Picofluor. Autoradiography ofLabeled Nuclei-Labeled nuclei were determined by autoradiography as described previously (34). Cultures were washed twice with DMEM and incubated in 2 ml of D M E~a~o u t~s medium (1:l v/v) containing [3Hlthymidine (5 pCi/mll and various additions. After 40 h, the medium was removed, and the cultures were washed twice with Tris saline, pH 7.4, at 4 "C, fixed twice (for 5 and 2 min) with 5% trichloroacetic acid at 4 "C, and washed three times with 70% ethanol. Autoradio~phic film was put on the cultures and developed 7 days later. After staining with Giemsa, the labeled nuclei were counted.
Clonogenic Assay-Stock cultures of rat-1 cells, 3-4 days after passage, were trypsinized and gently dispersed to ensure a single-cell suspension. Cells were added to DMEM supplemented with 2.5% FBS and other factors, as described for each experiment, to give 5 x 1D" cells per ml. Agarose was added at a dilution of 0.3%. An aliquot (1 ml) of this mixture was plated onto 33-mm plastic dishes containing a 2-ml base layer of hardened 0.3% agarose. In all cases, the composition of the base layer was identical with that of the top layer, as indicated. The cultures were incubated in an atmosphere of 10% GO,, 90% air at 37 "C for 14 days and then stained for 18 h with the vital stain nitro blue tetrazolium. Colonies with diameters >200 pm were counted using a BioTranII automated colony counter (New Brunswick Scientific).
Release of Radioactiuity from P H l~a c h i d o~t e -l a~l e d CeUs-Confluent and quiescent cultures of Swiss 3T3 cells in 33-mm dishes were incubated for 16 h with L5, 6,8,9,11,12,14,15-sHlarachidonic acid (1 pCi/ml, 211 Citmmol). One hour prior to the start of the experiment, NDGA was added as indicated. After this time, the cells were washed twice with DMEM and incubated in 1 ml of the same medium in the absence or presence of PDGF and NDGAat 37 "C. ARer 1 h, the medium was removed and centrifuged at 16,000 x g for 5 min, and the radioactivity in the supernatant was determined by scintillation counting.
The kinetics of arachidonic acid mobilization were determined in the following manner. Labeled cultures incubated in the absence or presence of NDGA for 1 h prior to the start of the experiment were washed twice with DMEM and incubated at 37 "C in 2 ml of DMEM with or without PDGF and NDGA. At various times, aliquots of the extracellular medium (50 pl) were removed and processed as described above. The volume of medium removed was replaced with fresh medium at 37 "C containing the corresponding factors.
Production ofPGE-Cultures pretreated for 1 h in the absence or presence of NDGA were washed twice with DMEM and incubated with PDGF and NDGA at 37 "C in 1 ml of DMEM. After 1 h, the medium was removed, and the PGE, released into the medium was determined by radioimmunoassay as follows. Samples were diluted in assay buffer consisting of 0.15 M NaCI, 10 m~ EDTA, 0.3% bovine y-globin, 0.005% Triton X-100,0.05% sodium azide, and 25 m~ phosphate buffer, pH 6.8. Rabbit anti-PGE, antibody, lZ6I-PGEz tracer, and the samples were incubated for 24 h at 4 "C. Antibody-bound tracer was precipitated by the addition of 16% polyethylene glycol, 0.05% sodium azide, and 50 m M phosphate buffer, pH 6.8, for 30 min at 4 "C. After this time, the samples were centrifuged at 4 "C (3,000 x g, 30 min), and the supematants were aspirated. The radioactivity in the pellets was determined by scintillation counting.
Zrnmunoprecipitations-Quiescent cultures of Swiss 3T3 cells were incubated in the absence or presence of NDGA for 1 h. ARer this time, the cells were washed twice with DMEM and treated with factors for 10 min at 37 "C in the absence or presence of NDGA as indicated. The cells were then lysed at 4 "C in 1 ml of 10 m~ TrisiHCl, pH 7.6,5 m~ EDTA, 50 m M NaC1,30 m~ sodium pyrophosphate, 50 m~ NaF, 100 PM Na3V04, 1% Triton X-100, and 1 m M phenylmethyisulfonyl fluoride (lysis buffer). Lysates were clarified by centrifugation at 13,000 x g and precleared by incubation with albuminlagarose for 1 h at 4 "C. After removal of albuminlagarose by a brief (10 s) centrifugation, the supernatants were transferred to a fresh tube, and phosphotyrosyl proteins were immunoprecipitated for 16 h at 4 "C with agarose-linked anti-phosphot~osine monoclonal antibody (Ab-1). For the immunopre~pi~tion of p120, the 4B12 monoclonal antibody was precoupled to goat anti-mouse IgG-agarose. Lysates were also incubated with anti-GTPase activating protein, anti-phospholipase Cyl, and anti-a-PDGF receptor (PDGFR-7) antiserum (35) for 16 h at 4 "C. At the end of this time, the immune complexes were immunoprecipitated with protein A-agarose for 4 h at 4 "C.
Assay ofprotein Kinase Activity-Immunoprecipitates prepared from 1.5 x lo6 cells as described above were washed three times with lysis buffer and twice with 50 m~ HEPES, pH 7.4,O.l m~ EDTA, 0.01% Brij, 75 m~ NaCl, and 100 Na3V0, (kinase assay buffer). NDGA was added to the immunoprecipitates in 1 ml of kinase assay buffer, and the mixture was incubated for 15 min at 4 "C. After this time, the immunoprecipitates were resuspended in 20 pl of kinase buffer, and the reaction was initiated by the addition of 10 m~ MgClz and 100 p~ [y-3zPl ATP (10 pCi) in a total volume of 30 pl at 30 "C. ARer a 10-min incubation, immunoprecipitates were washed twice with lysis buffer and analyzed by SDS-PAGE followed by autoradiography.

NDGA Selectively Inhibits PDGF-stimulated DNA Synthesis-Confluent and quiescent cultures of Swiss 31'3 cells were stimulated with PDGF in the absence or presence of in-
creasing concentrations of NDGA. At the end of 40 h, the incorporation of L3Hlthymidine into cellular DNA was determined. Fig. 1 shows that NDGA caused a dramatic inhibitory effect upon PDGF-mediated mitogenesis. Maximal inhibition of PDGF [3HJthymidine incorporation (96%) was observed at 4 m NDGA (ICso = 1.5 )1~). Parallel cultures were stimulated with either bombesin or EGF in the presence of insulin or with the potent mitogen Pasteurella rnultocida toxin (37). No effect of NDGA was observed upon DNA synthesis stimulated by these mitogens over the same concentration range of NDGA which completely blocked the stimulation by PDGF. In addition, no effect of NDGA at 4 p~ was observed upon FBS-and insulinstimulated DNA synthesis while the response to vasopressin and insulin and phorbol 12,13-dibutyrate and insulin was attenuated by 22% and 34%, respectively.
O n the basis of these results, we proceeded to examine the effect of NDGA upon PDGF-induced DNA synthesis in greater detail. Fig. 2A shows the dose-dependent stimulation of DNA synthesis by PDGF in the absence and presence of NDGA. In control cultures, maximal stimulation of DNA synthesis by PDGF was obtained using 7.5 ng/ml (ECao = 3.5 ng/ml). This response was dramatically inhibited (>97%) by 4 p~ NDGA. At PDGF concentrations greater than 5 ng/ml, however, some stimulation was still retained. It is known that insulin markedly increases the potency of PDGF for inducing DNA synthesis (38). Fig. 2B shows the synergistic stimulation of DNA synthesis by PDGF and insulin. In the presence of 1 pg/ml insulin, the PDGF mitogenic dose response is shifted to the left. Under these conditions, maximum L3H1thymidine incorporation was observed a t 2.5 ng/ml PDGF (EC50 = 0.5 ng/ml). NDGA caused a 15-fold displacement of this response (ECS0 = 7.5 ng/ml). Stimulation of DNA synthesis by 10 ng/ml PDGF together with insulin in the absence and presence of 4 VM NDGA was 101% and 85% of FBS, respectively. EGF in the presence of a fixed concentration of insulin (0.5 pg/ml) also produced a dose-dependent increase in L3H]thymidine incorporation (Fig. 2C). Maximal stimulation of DNA synthesis was obtained by 1.25 ng/ml EGF (EC50 = 0.15 ng/ml). In the presence of NDGA (4 p~) , a slight stimulation of L3H1thymidine incorporation was observed with insulin alone, and, at maximal EGF concentrations, only a slight inhibitory effect was observed (18%).
To confirm that NDGA was inhibiting PDGF-stimulated DNA replication rather than changing the specific activity of the ["Hlthymidine precursor pool, quiescent cultures of Swiss 3T3 cells were treated in the absence or presence of NDGA together with various factors, and the incorporation of ["HJthymidine into DNA subsequently was quantified by autoradiography of labeled nuclei. Fig. 3A shows the dose-dependent enhancement of the labeling index by PDGF. Maximum stimulation by PDGF was observed using 10 ng/ml (EC50 = 3.5 ng/ml). These values are in good agreement with the measurement of ["Hlthymidine incorporation. Stimulation of PDGFinduced DNA synthesis was severely inhibited (>98%') by 4 PM NDGA. The selective nature of NDGA inhibition as assessed by autoradiography of labeled nuclei is shown in Fig. 3B. A slight synergistic effect of NDGA with insulin was observed (control 2.370, NDGA 15.4%) while NDGAinhibited the effect o f P multocida toxin, bombesin and insulin, EGF and insulin, and vasopressin and insulin by 13%#, 28%#, 34%, and 14%, respectively. Such inhibitions are similar to those observed when the incorporation of ["Hlthymidine stimulated by these mitogens was determined (Fig. 1). The results of these experiments confirm a marked inhibitory effect of NDGA upon PDGF-stimulated DNA synthesis in Swiss 3T3 cells.
In view of the selective effects of NDGA in preventing PDGFstimulated DNA synthesis in Swiss 3T3 cells, we determined whether similar effects of NDGA could be observed in diploid mouse fibroblasts as well as in cells from other species. Quiescent cultures of TMEF cells, rat-1 cells, and human foreskin fibroblasts were stimulated with PDGF and other growth-promoting factors either in the absence or presence of 2 V M NDGA. Cumulative [3H]thymidine incorporation was then measured after 40 h of incubation. NDGA abolished the PDGF stimulation of DNA synthesis in each cell type (Fig. 4). No significant effect of NDGA was observed upon the DNA synthesis stimulated by a range of other mitogens in these cells with the exception of an inhibition (36%) of the mitogenic effect of EGF and insulin in TMEF cells.
We also examined the effect of NDGA upon PDGF-stimulated anchorage-independent growth of rat-l cells. As shown in Table  I, PDGF and EGF both induced colony formation of this cell type causing a 9-fold and 6-fold increase in the number of colonies per dish, respectively. When both factors were added together, a synergistic stimulation was observed (39-fold). Addition of 4 p~ NDGA markedly inhibited the PDGF effect (76% inhibition) while no inhibition of EGF-stimulated colony formation was seen. NDGA also inhibited the synergistic effect of PDGF and EGF (79% inhibition) to that observed in the presence of EGF alone. The results of these experiments demonstrate that the marked inhibitory effect of NDGA upon PDGFstimulated DNAsynthesis is not restricted to the Swiss 3T3 cell line but is also observed in murine diploid cells (TMEF), rat-1 cells, and human fibroblasts. P D G F -s t i~u~a t e~ A r a c~~~o n i c Acid ~o b i l~a t i o~ and PGE, Release Is Also ~n~~~~~e d by ~~G A -P D G F , unlike EGF, stimulates a sustained mobilization of arachidonic acid in Swiss 3T3 cells (19). Once mobilized, the metabolism of this fatty acid in these cells is directed toward PGE2 biosynthesis (10,14,19). If NDGA inhibits eicosanoid production by acting as a potent lipoxygenase inhibitor (25-27), i t should not interfere with either arachidonic acid mobilization or PGE2 production in response to PDGF stimulation. To test this, cultures were labeled with [3H]arachidonic acid for 16 h followed by pretreatment for 1 h with or without NDGA as indicated. After this time, the cells were incubated for 1 h in the absence or presence of PDGF and NDGA. Fig. 5A shows that PDGF causes a 34-fold increase in arachidonic acid release after 1 h of stimulation. Suprisingly, NDGA produced a dose-dependent inhibition of this effect (IC50 = 1.4 PM). Maximum inhibition (94%) was achieved by 5 p NDGA. PDGF stimulated arachidonic acid mobilization in a dose-dependent manner (Fig. 5 3 1. Maximal arachidonic acid mobilization was achieved by 12 ng/ml PDGF (ECEO = 7.5 ng/ ml). NDGA (4 p~) potently inhibited PDGF-mediated arachidonic acid release over the entire dose-response curve. The inhibition of PDGF-stimulated arachidonic acid mobilization by NDGA was not due to a shift in the kinetics of the response. Fig. 5C shows the biphasic pattern of PDGF-mediated arachidonic acid release previously described in Swiss 3T3 cells (19). NDGA (4 p M j inhibited the arachidonic acid release stimulated by PDGF (25 ngfml) over a 1-h period. In Swiss 3T3 cells, the predominant arachidonic acid metabolite is PGE2.
PDGF (25 ng/mi) produced a 36-fold increase in PGE, production after 1 h of stimulation (Fig. 511). This effect was inhibited by NDGA in a dose-dependent manner. Half-maximal inhibition was achieved with 0.9 PM NDGA, and maximal inhibition (97%) was obtained by 5 p NDGA.
The data shown in Fig. 5 demonstrate that NDGA severely inhibits PDGF-stimulated arachidonic acid mobilization, and, consequently, PGE2 production in a dose-dependent manner in Swiss 3T3 cells.
PDG?-stimulated Qrosine Phosphorylation in Intact Cells Is In.hibited by NDGA-Inhibition of arachidonic acid mobilization by NDGA could not account for the dramatic effect of NDGA on PDGF-stimulat~d DNA synthesis over its entire dose response (14). We reasoned that other signaling events were also affected. Since stimu~ation of PDGF receptor tyrosine kinase activity is a prerequisite for the induction of DNA synthesis by this ligand (1,3,4), we examined whether exposure of Swiss 31'3 fibroblasts to NDGA inhibited PDGF-stimulated tyrosine phosphorylation. Quiescent cultures were pretreated for 1 h with or without NDGA and then stimulated with PDGF, EGl?, or bombesin in the absence or presence of NDGA as in-

The egect of NDGA on PDGFand E e F -s t~~u l a~d colony fornation by rat-1 cells
NDGA was used at concentration of 4 m, PDGF and EGF were both used at 25 ng/ml. Colony formation (mean t S.E., n = 5) of rat-1 cells was determined in 0.3% agarose containing DMEM and 2.5% FBS in the absence or presence of growth factors as indicated. NDGA-mediated inhibition of PDGF receptor phosphorylation was dose-dependent (Fig. 6A). Total inhibition of receptor autophosphorylation stimulated by 10 ng/ml PDGF was achieved using 3 p~ NDGA (IC50 = 0.85 PM). When the concentration of PDGF was increased to 25 ng/ml, maximal inhibition of receptor phosphorylation by NDGA was 60%.

Colonieddish
To complement the Western blot analysis, protein tyrosine kinase activity of ti-phosphotyrosine immunoprecipitates obtained from cells stimulated in the absence and presence of PDGF and NDGA was also determined. Cultures incubated with or without NDGA for 1 h were stimulated in the presence and absence of PDGF and NDGA for 10 min. After cell lysis and immunoprecipitation with anti-phosphotyrosine monoclonal antibodies, the resulting immune complexes were incubated with [y-32Pl ATP for 10 min in the absence or presence of NDGA as indicated, and the products were analyzed by SDS-PAGE. Immunoprecipitates prepared from cultures stimulated with 10 nglml and 25 ng/ml PDGF produced a marked phosphorylation of the PDGF receptor migrating with a n M , of 180,000-190,000 (Fig. 6B). Addition of NDGA caused a dosedependent inhibition of receptor phosphorylation stimulated by both concentrations of PDGF with maximal inhibition achieved at 4 p~ and SO p~ NDGA, respectively.

NDGA Inhibits PDGF-stimulated Qrosine Phosphorylation of Distinct Receptor Substrates in Intact
Cells-Ligand activation and autophosphorylation of the PDGF receptor is known to result in the tyrosine phosphorylation of multiple receptor substrates including GTPase activating protein, p120, and phospholipase Cyl (6). We examined the degree to which tyrosine phosphorylation of these substrates was inhibited by NDGA. Quiescent cultures of Swiss 3T3 cells were incubated with PDGF in the absence or presence of NDGA which was added presence (closed circles) of 25 n g h l PDGF together with NDGA as indicated for 1 h at 37 "C. At the end of this time, the extracellular medium was removed and centrifuged at 13,000 x g for 5 min, and the supernatant was subjected to a specific radioimmunoassay for PGEz as described under "Experimental Procedures." Where used, NDGA was added to the cultures 1 h prior to the start of the experiment and was present throughout. 1 h prior to stimulation. At the end of this time, cells were lysed, and proteins were immunoprecipitated with either anti-phosphotyrosine-agarose monoclonal antibodies, agarose-linked anti-pl20 monoclonal antibody, anti-GTPase activating protein, or anti-phospholipase C y , antisera. The immunoprecipitates were analyzed by SDS-PAGE followed by immunoblotting with anti-phosphotyrosine monoclonal antibodies. Fig. 7 shows that 6 p~ NDGA markedly inhibits the PDGF-stimulated tyrosine phosphorylation of GTPase activating protein (60%), pS20 (88%), and phospholipase Cy, (40%). Each antiserum also immunoprecipitated a band migrating with a M , = 180,000-190,000 which is most likely associated PDGF receptor. The anti-GTPase activating protein antiserum constitutively immunoprecipitated a band of M, = 180,000-185,000.

NDGA (pM)
FIG. 6. NDGA inhibits PDGF stimulation of tyrosine phosphorylation in intact cells. Upperpanel, quiescent and confluent cultures of Swiss 3T3 cells were incubated for 1 h in the absence or presence of 10 p~ NDGA as indicated. The cultures were then washed twice with DMEM and incubated for 10 min with either 10 ng/ml or 25 ng/ml PDGF, 10 ny bombesin, or 5 ng/ml EGF in the absence or presence of 10 p~ NDGA. Cells were lysed, and anti-phosphotyrosine immunoprecipitates were prepared as described under "Experimental Procedures." Tyrosine-phosphorylated proteins were detected by Western blot analysis using anti-phosphotyrosine monoclonal antibodies. Partel A, in intact cells, NDGA inhibited PDGF tyrosine phosphorylation in a dosedependent manner. Quantitation of the NDGA-induced inhibition of PDGF receptor phosphorylation by 10 ng/ml PDGF (closed circles) and 25 ng/ml PDGF (open circles) was achieved by scanning densitometry. Panel B , confluent and quiescent cultures were incubated for 1 h in the absence or presence of NDGAas shown. The cultures were washed with DMEM and incubated for 10 min with either 10 ng/ml PDGF (closed circles) or 25 ng/ml PDGF (open circles) in the presence or absence of NDGA. Following cell lysis, anti-phosphotyrosine immunoprecipitates were prepared as described under "Experimental Procedures." Protein kinase activity of the immune complexes was assessed during a 10-min incubation with I y'"PIATP at 30 "C in the absence or presence of NDGA.
The products of this reaction were analyzed by SDSPAGE and autoradiography. Phosphorylation of the PDGF receptor was quantified by scanning densitometry. and immunoprecipitates from lysates of these cells were prepared using anti-phosphotyrosine monoclonal antibodies. These immunoprecipitates were then preincubated in the absence or presence of NDGA for 15 min at 4 "C. After this time, the kinase activity of the immunoprecipitates was determined. Fig. 8 (left and center panels) shows that NDGA produces a dose-dependent inhibition of PDGF-stimulated kinase activity in the immunoprecipitates. Maximal inhibition was achieved using 10 p~ NDGA (ICao = 3.8 PM).
The effect of NDGA in vitro upon the protein tyrosine kinase activity of immunoprecipitates prepared using an a-PDGF receptor monoclonal antibody was also examined. Confluent and quiescent cultures of Swiss 3T3 cells were stimulated with 10 ng/ml PDGF for 10 min at 37 "C. The monolayer was lysed, and the extract was incubated with an a-PDGF receptor antibody (35) for 16 h at 4 "C. The immunoprecipitates were incubated in the absence or presence of 6 PM NDGA for 15  NDGA inhibits PDGF-stimulated tyrosine phosphorylation of GTPase activating protein, p120, and phospholipase C y in intact cells. Cultures of Swiss 3T3 cells were incubated for 1 h in the absence or presence of 6 p~ NDGA as indicated. The cultures were then washed twice with DMEM and incubated with 25 ng/ml PDGF in the absence or presence of 6 p~ NDGA for 10 min. The cells were then lysed and immunoprecipitated with specific antibodies to either phosphotyrosine ( P -Y ) , GTPase activating protein, p120, or phospholipase Cyl.
Following SDSPAGE, immunoprecipitates were analyzed by immunoblotting with anti-phosphotyrosine monoclonal antibodies and autoradiography.
Treatment of intact cells with PDGF resulted in the association and subsequent tyrosine phosphorylation of a set of proteins of M , = 70,000-80,000 (Fig. 8, vertical bar). NDGA abolished the phosphorylation of these PDGF receptor substrates.
The result of this series of experiments (Figs. [6][7][8] demonstrates that NDGA selectively inhibits PDGF-stimulated tyrosine phosphorylation in intact cells and cell-free preparations.

DISCUSSION
The results of the present study demonstrate that NDGA produces a potent inhibition of PDGF-mediated DNA synthesis in a variety of cell types. NDGA, however, exerted little or no consistent effect upon the mitogenic response to a wide range of other growth factors including EGF, a ligand whose receptor also possesses intrinsic tyrosine kinase activity. Indeed, a small synergistic stimulation of DNA synthesis was observed when NDGA was added together with insulin. NDGA maintained its potent inhibition of PDGF-mediated DNA synthesis in the presence of insulin. This selective inhibition of PDGF-stimulated DNA synthesis was not confined to Swiss 3T3. Our results demonstrate a selective inhibition of PDGF stimulated DNA synthesis in diploid murine cells, rat, and human fibroblasts. We have also shown that NDGA inhibits PDGF stimulation of PDGF --+ + + + + + F~~  8. NDGA inhibits PDGF-stimulated tyrosine kinase activity in a cell-free system. Leflpanel, quiescent and confluent cultures were stimulated with 25 ng/ml PDGF for 10 min a t 37 "C. Following cell lysis, anti-phosphotyrosine immunoprecipitates were prepared as described under "Experimental Procedures." The immunoprecipitates were then incubated in the absence or presence of various concentrations of NDGA for of NDGA. The products of the reaction were analyzed by SDSPAGE and autoradiography. Center panel, in the cell-free system, NDGA inhibited 15 min a t 4 "C. After this time, protein kinase activity was determined by incubation with [y3*P]ATP a t 30 "C for 10 min in the absence or presence PDGF tyrosine phosphorylation in a dose-dependent manner. Quantitation of the NDGA-induced inhibition of PDGF receptor phosphorylation by 25 nglml PDGF (open circles) was achieved by scanning densitometry. Right panel, cultures of Swiss 3T3 cells were incubated in the presence or absence of 10 ng/ml PDGF for 10 min a t 37 "C. After cell lysis, immunoprecipitates were prepared using anti-a-PDGF receptor antisera (PDGFR-7). The immunoprecipitates were then incubated in the absence or presence of 6 p~ NDGA for 15 min a t 4 "C. After this time, protein kinase activity was determined bv incubation with W*PIATP a t 30 "C for 10 min in the absence or presence of 6 p~ NDGA. The products of the reaction were analyzed by SDSkAGE and autoradiography.
anchorage-independent growth of rat-1 cells while EGF-stimulated colony formation was unaltered. This finding further illustrates the selective nature of the NDGA inhibition of PDGFstimulated cell proliferation.
PDGF induces a striking mobilization of arachidonic acid and production of PGE2 in Swiss 3T3 cells (10,14,19). Since NDGA has been used ubiquitously as a selective inhibitor of lipoxygenase activity, we were surprised to find that it also potently inhibited PDGF-stimulated arachidonic acid mobilization and, consequently, PGE2 production in a dose-dependent manner. These inhibitory effects were observed a t concentrations which gave near-maximal inhibition of PDGF-stimulated DNA synthesis. It is known that NDGA has nonselective effects when used at high concentrations (44,45). Our results demonstrate that even a t low concentrations (<lo PM), NDGA can no longer be regarded a s a selective inhibitor of the lipoxygenase pathway.
Previous studies demonstrated that vasopressin pretreatment also potently inhibits PDGF-stimulated arachidonic acid mobilization in Swiss 3T3 cells (14). PDGF-mediated DNA synthesis was, however, only inhibited by this pretreatment when subsaturating concentrations of PDGF were used. At concentrations of PDGF greater than 12-15 ng/ml, the inhibitory effect of chronic vasopressin desensitization on DNA synthesis was overcome. Since NDGA dramatically inhibits DNA synthesis even at saturating concentrations of PDGF, we reasoned that NDGA inhibits additional signaling events.
Stimulation of quiescent fibroblasts by PDGF is accompanied by the activation of protein tyrosine kinase activities resident in the cytoplasmic domains of the PDGF receptor (1,3,4). Kinase inactive mutants of the PDGF receptor cannot mediate most ligand-dependent biological responses showing that tyrosine phosphorylation is necessary for signal transduction (46). The PDGF receptor tyrosine kinase is characterized by having its catalytic domain split by an insert of approximately 100 amino acids (2)(3)(4). Binding of PDGF results in the autophosphorylation of the PDGF receptor on specific tyrosine residues, thereby facilitating the association of numerous substrates. To date, 8 tyrosine residues along the P-PDGF receptor have been identified as sites of receptor autophosphorylation and intracellular substrate interaction (47).
The results of our studies produced several lines of evidence indicating that NDGA potently inhibits PDGF receptor tyrosine kinase activity. Western blot analysis demonstrated that NDGA specifically inhibits PDGF-stimulated tyrosine phos-phorylation in intact cells in a dose-dependent manner. Since the PDGF receptor is the major substrate of its own tyrosine kinase activity, this was taken as the reference for the inhibitory effect of NDGA. Analysis of protein tyrosine kinase activity of anti-phosphotyrosine immunoprecipitates prepared from cells stimulated with PDGF in the presence or absence of NDGA also revealed an inhibitory effect of this compound. Western blot analysis for specific PDGF receptor substrates demonstrated that NDGA inhibited the tyrosine phosphorylation of GTPase activating protein, phospholipase Cy, and p120. Since the inhibition of p120 was quantitatively the most striking of the three substrates examined, these results suggest a differential inhibition of PDGF receptor substrate tyrosine phosphorylation by NDGA. Protein tyrosine phosphorylation stimulated by either EGF or bombesin was unaltered by NDGA treatment. Crucially, NDGA inhibited the ability of anti-phosphotyrosine and anti-a-PDGF receptor immunoprecipitates prepared from cultures treated with PDGF to stimulate tyrosine phosphorylation in vitro. These results suggest that the potent and selective inhibitory effect of NDGA on PDGF-stimulated DNA synthesis results from its inhibitory action on tyrosine phosphorylation.
In the search for antagonists of PDGF action, a number of low molecular weight compounds have emerged as promising candidates. One of the earliest examples was suramin (48). This compound was, however, subsequently shown to act in a nonspecific manner (49). The aminoglycoside neomycin was also found to affect certain PDGF responses (50). Its use is limited, however, because of the high concentrations required to achieve the effect on receptor binding. Peptides derived from the primary sequence of either PDGF or the extracellular portion of the PDGF receptor have also been shown to act as receptor antagonists (51,52) as have antibodies to PDGF which inhibited autocrine stimulation of Simian Sarcoma virus-transformed cells (53). Since stimulation of receptor-mediated phosphorylation is the primary event following binding of many polypeptide growth factors to their receptors, one approach in the development of selective antagonists has been directed against this protein tyrosine kinase activity. The tryphostins are a group of synthetic compounds which exhibit potent inhibitory effects upon both membrane-bound and cytosolic protein tyrosine kinases (54). These compounds are themselves modelled upon natural compounds shown to inhibit phosphorylation including erbstatin, the flavone quercetin, the isoflavone genistein, and herbimycin A (55).
Our findings suggest that NDGA may provide a novel structural motif to generate tyrosine kinase inhibitors with selectivity for PDGF. In contrast to other PDGF receptor antagonists or inhibitors of tyrosine phosphorylation, NDGA has already been shown to possess low toxicity in animal models (56). Identification of such preferential inhibitory effects, novel mechanism of action, and low toxicity raises the possibility that NDGA or its derivatives may have important pharmaceutical potential. NDGA could play a n important role in the treatment of various disease states which implicate PDGF as a causative proliferative agent including atherosclerosis, fibrotic conditions, and cancer.