Platelet-derived growth factor stimulates protein kinase A through a mitogen-activated protein kinase-dependent pathway in human arterial smooth muscle cells.

The abilities of platelet-derived growth factor (PDGF) and insulin-like growth factor (IGF-I) to regulate cAMP metabolism and mitogen-activated protein kinase (MAP kinase) activity were compared in human arterial smooth muscle cells (hSMC). PDGF-BB stimulated cAMP accumulation up to 150-fold in a concentration-dependent manner (EC50 approximately 0.7 nM). The peak of cAMP formation and cAMP-dependent protein kinase (PKA) activity occurred approximately 5 min after the addition of PDGF and rapidly declined thereafter. Incubating cells with PDGF and 3-isobutyl-1-methylxanthine (IBMX, a phosphodiesterase inhibitor) enhanced the accumulation of cAMP and PKA activity by an additional 2.5-3-fold, whereas IBMX alone was essentially without effect. The PDGF-stimulated increase in cAMP was prevented by addition of the cyclooxygenase inhibitor indomethacin, consistent with release of prostaglandins stimulating cAMP. PDGF, but not IGF-I, stimulated MAPK activity, cytosolic phospholipase A2 (cPLA2) phosphorylation, and cAMP synthesis which indicated a key role for MAP kinase in the activation of cPLA2. Further, PDGF stimulated the rapid release of arachidonic acid and synthesis of prostaglandin E2 (PGE2) which could be inhibited by a cPLA2 inhibitor (AACOCF3). Calcium mobilization was required for PDGF-induced arachidonic acid release and PGE2 synthesis but not for MAPK activation, whereas PKC was required for PGE2-mediated activation of PKA. In summary, these results demonstrated that PDGF increases cAMP formation and PKA activity through a MAP kinase-mediated activation of cPLA2, arachidonic acid release, and PGE2 synthesis in human arterial smooth muscle cells.

The proliferation of arterial smooth muscle cells is a key event in the formation and progression of lesions of atherosclerosis and in restenosis following angioplasty. This proliferation is most likely initiated and regulated by growth factors, such as the platelet-derived growth factors, PDGF-AA, 1 PDGF-BB, and PDGF-AB. In atherosclerotic lesions, a major source of PDGF-BB is activated macrophages, although smooth muscle, endothelial cells, and other cells can also express and secrete PDGF dimers (reviewed in Ref. 1).
Many growth factor receptors, including PDGF receptors, activate a signal transduction pathway that includes conversion of inactive Ras⅐GDP to active Ras⅐GTP, activation of Raf, MAP kinase kinase (MAPKK or MEK (2)), and MAP kinase (MAPK) (for review, see Ref. 3). Activation of the MAPK cascade results in the stimulation of DNA synthesis and cell proliferation, which can be inhibited by expression of a dominantnegative MAPKK (4,5). Conversely, expression of a constitutively active form of MAPKK can stimulate cell proliferation and transformation (4,6). A number of nuclear and non-nuclear proteins have been identified as substrates for MAPK. Among the latter is phospholipase A 2 , thereby providing a potential link to arachidonic acid metabolism (reviewed in Ref. 7).
Activation and translocation of the cytosolic 85-kDa phospholipase (cPLA 2 ), which catalyzes the release of arachidonic acid from the sn-2 position of phospholipids in the plasma membrane, is an important signal leading to prostaglandin synthesis (reviewed in Refs. 8 and 9). This enzyme can be distinguished from the low molecular weight forms of PLA 2 by insensitivity to disulfide-reducing agents or inhibition by arachidonic acid analogues (for reviews, see Refs. 10 and 11). Regulation of cPLA 2 appears to be critically dependent on the integration of multiple signals, including intracellular calcium, protein kinase C (PKC), and phosphorylation by MAP kinase (12)(13)(14)(15). However, in some cells, calcium-independent forms of cPLA 2 (16) and PKC-independent mechanisms of cPLA 2 activation have also been observed (15).
Although the coupling of hormonal and neurotransmitter receptors to cAMP synthesis is well established (reviewed in Ref. 17), the means by which growth factor receptor tyrosine kinases (e.g. the PDGF receptor) regulate cAMP metabolism is less well understood. At present, there are few examples of growth factors stimulating cAMP accumulation. Before elucidation of the MAP kinase cascade, it was reported that PDGF (in the presence of phosphodiesterase inhibitors) increased cAMP synthesis 6 -8-fold in Swiss 3T3 cells (18). In perfused rat hearts, epidermal growth factor (EGF) was found to stimulate cAMP accumulation (19). In epithelial cells overexpressing the EGF receptor (A431 cells), EGF alone did not affect cAMP accumulation, but did potentiate an increase in cAMP in response to cAMP-elevating agents (20).
Sustained elevation of cAMP inhibits the proliferation of many cell types, including smooth muscle cells (21). This phenomena may in part be explained by the fact that in many cell types, including human arterial SMC (hSMC), MAP kinase activation is inhibited by the cyclic AMP-dependent protein kinase (PKA) (22)(23)(24)(25)(26). The target for inhibition by PKA in the MAP kinase pathway may be Raf-1 (23,27), although other unidentified targets are also likely to play an important role (28).
To further evaluate how growth regulatory molecules may regulate cAMP metabolism in hSMC, we examined the effect of two major factors known to influence hSMC (i.e. PDGF and IGF-I). We report here that PDGF rapidly stimulates cAMP synthesis through a mechanism requiring intracellular calcium, PKC activity, MAPK-dependent phosphorylation of cPLA 2 , and activation of prostaglandin synthesis, an effect which culminates in increased PKA activity.

EXPERIMENTAL PROCEDURES
Materials-Recombinant human PDGF-BB and PDGF-AA were gifts from Hoffman-La Roche Inc. (Basel, Switzerland). Human recombinant IGF-I was obtained from Upstate Biotechnology Inc. (UBI, Lake Placid, NY). PKI peptide (TTYADFIASGRTGRRNAIHD), a specific peptide inhibitor of PKA (29), and Leu-Arg-Arg-Ala-Ser-Leu-Gly (Kemptide) (30) were synthesized at the Peptide Synthesis Facility, Howard Hughes Medical Institute (Seattle, WA). Recombinant rat Erk2 was a gift from Dr. M. Cobb (University of Texas, Southwestern, Dallas, TX), and anti-peptide antibodies to this protein were developed previously in this laboratory by Dr. R. Seger. Goat anti-rabbit alkaline phosphatase conjugate was obtained from Promega. Indomethacin, PMA, IBMX, and thapsigargin were obtained from Sigma, and forskolin was obtained from Calbiochem. These compounds were dissolved in dimethyl sulfoxide prior to use. AACOCF 3 , AACOCH 3 , and recombinant cPLA 2 were gifts from Dr. M. Gelb (University of Washington, Seattle, WA). In some instances, AACOCF 3 was obtained from Biomol (Plymouth Meeting, PA) as well as was the PKC inhibitor GF109203X. The arachidonic acid analogues were dissolved in ethanol or dimethyl sulfoxide prior to use. Polyclonal antibodies to cPLA 2 were prepared in rabbits immunized with human recombinant cPLA 2 as described previously (31).
Cell Cultures-Human newborn arterial smooth muscle cells were obtained from the thoracic aorta of infants following death due to congenital heart defects or sudden infant death. The cells were isolated by the explant method and cultured as described previously (32). Cells were used at passages 5 to 10, and were characterized as smooth muscle cells by morphologic criteria and by expression of smooth muscle ␣-actin. The cells were negative in mycoplasma assays and had a normal chromosome number. Subconfluent cell cultures were kept in Dulbecco's modified Eagle's medium (DMEM)/1% plasma-derived serum (PDS) for 2 days prior to experiments.
Measurement of Cyclic AMP-Smooth muscle cells in 10-cm dishes were stimulated with growth factors, tumor promoters, or inhibitors for the indicated periods of time. A typical experiment contained 3 ϫ 10 6 cells per 10-cm plate. The plates were washed 3 times with cold phosphate-buffered saline, and the cAMP was rapidly extracted with the addition of 1 ml of ice-cold ethanol (70%). After scraping the plates, the suspension was centrifuged at 13,000 ϫ g for 20 min at 4°C, the supernatant was collected, and 0.5 ml was evaporated in a Speed Vac centrifuge. Levels of cAMP were determined using a cAMP enzymelinked immunosorbent assay kit in prototypic stage; this was generously provided by Life Technologies, Inc.
Measurement of PKA Activity-PKA was assayed by measuring phosphorylation of the peptide substrate Kemptide (30), (0.17 mM) in the presence or absence of a 10 M concentration of the PKA inhibitor peptide (PKI) (29) as described (22). PKA activity was calculated as the amount of Kemptide phosphorylated in the absence of PKI peptide minus that phosphorylated in the presence of PKI peptide. The assay as described is influenced by the amount of cAMP carried over from the lysates into the kinase assay. In some instances, the activity ratio was determined for PKA as described by Corbin (33).
Measurement of MAPK and MAPKK Activity-MAPK and MAPKK activity was measured as described earlier (32). Cells in 10-cm dishes (approximately 3 ϫ 10 6 cells) were incubated with PDGF, IGF-I, or PMA for the indicated times. Immediately after stimulation, the cells were scraped and sonicated in buffer H (50 mM ␤-glycerophosphate, pH 7.4, 1.5 mM EGTA, 0.1 mM Na 3 VO 4 , 1 mM dithiothreitol, 25 g/ml aprotinin, 25 g/ml leupeptin, and 0.5 mM phenylmethylsulfonyl fluoride. For measurement of MAPKK activity, cell lysates were precleared on a DE52 mini column, and the phosphorylation of myelin basic protein by activated recombinant Erk2 was measured. MAPK activity was determined by the incorporation of radioactivity from [␥-32 P]ATP into myelin basic protein after a 15-min incubation at 30°C. Immunoblotting of MAP Kinase-Samples were immunoblotted for MAP (Erk1 and Erk2) after separation of samples on SDS-PAGE and transfer to Immobilon-P (Millipore). The antibody (7884) was raised against a carboxyl-terminal peptide of Erk2 and previously shown to detect both Erk1 and Erk2. The detection was completed by the use of goat-anti rabbit antibodies (alkaline phosphatase conjugate), and the color was developed as per the manufacturer's instruction (Promega).
Measurement of Arachidonic Acid Release and Prostaglandin E 2 (PGE 2 ) Release-For measurement of arachidonic acid release, cells in 6-well plates were labeled with [5,6,8,11,12,14, H]arachidonic acid (Dupont NEN) at 1 Ci/ml in DMEM, 1% human PDS for 24 h at 37°C according to Domin and Rozengurt (34). The cells were washed three times in new DMEM without [ 3 H]arachidonic acid and subsequently stimulated as indicated. The medium (1 ml) was collected, and the amount of released [ 3 H]arachidonic acid was determined by measuring the radioactivity in the medium. The radioactivity released prior to stimulation was subtracted from the amount released at the end of the stimulation. For measurement of PGE 2 synthesis, cells in 6-well plates were washed three times in DMEM (without PDS or bovine serum albumin), and then stimulated as indicated in 1 ml of new medium. PGE 2 released to the medium was measured after a 1:10 -1:30-fold dilution using a PGE 2 enzyme immunoassay (Amersham). The crossreactivity of this assay is less than 7% with PGE 1 and less than 5% with other related prostaglandins.
Phosphorylation of cPLA 2 -Cells in two maxi-plates (approximately 15 ϫ 10 6 cells/sample) were stimulated with vehicle (10 mM acetic acid, 0.25% bovine serum albumin), 1 nM PDGF-BB, or 10 nM IGF-I for 5 min. Cell extracts were prepared in Buffer H as described above and applied to a Mono Q (Pharmacia) column, which was developed with a linear gradient of 0 -400 mM NaCl in the same buffer. Fractions of 1 ml were collected and were assayed (5 l) for cPLA 2 phosphorylation by adding 1 g of recombinant cPLA 2 to a reaction mixture as described above for the MAP kinase assay. The recombinant cPLA 2 was confirmed by immunoblotting with antibodies to cPLA 2 . To control for nonspecific phosphorylation, in some instances cPLA 2 was omitted from the assay. The reaction was terminated by the addition of SDS-PAGE sample buffer, the samples were heated, separated on SDS-PAGE (10% acrylamide), and the gels were dried. Radioactive cPLA 2 was visualized by autoradiography. Mono Q fractions were also assayed for MAPK activity and immunoblotted for MAPK (Erk1, -2) protein as described above.
Immunoblotting of cPLA 2 -Lysates from cells incubated with PDGF-BB or IGF-I for 5 min were obtained by harvesting the cells in a buffer containing 50 mM Hepes, pH 7.4, 10% glycerol, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 200 M sodium orthovanadate, 10 mM tetrasodium pyrophosphate, 100 mM sodium fluoride, 3 M p-nitrophenyl phosphate, 10 g/ml aprotinin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride. The samples were added to an equal volume of SDS-PAGE sample buffer, boiled for 5 min, and were applied to a 10% SDS-polyacrylamide gel containing 1% bisacrylamide. Electrophoresis was carried out until the 80-kDa prestained standard migrated to 1 cm above the end of the gel. After being transferred to nitrocellulose, the membranes were incubated with polyclonal antibody to cPLA2 (1:1000) and the immunodetection was accomplished by the use of goat anti-rabbit horse radish peroxidase and enhanced chemiluminescence (ECL).

PDGF Stimulates a Transient Increase in cAMP and PKA
Activity-Incubation of hSMC with PDGF-BB (1 nM) stimulated a rapid increase in cAMP accumulation (approximately 50-fold) that was detected as early as 2.5 min after the addition of the growth factor and reached a maximal level by 5 min (Fig.  1A). The PDGF-stimulated increase in cAMP did not require the addition of phosphodiesterase inhibitors. In parallel with cAMP, PDGF-BB increased PKA activity up to 7-8-fold as determined by Kemptide phosphorylation (Fig. 1A). The PDGF-stimulated increase in cAMP and PKA activity was transient and disappeared within 15 min with no further increase in cAMP detected up to 1 h after PDGF addition. These results were confirmed in different isolates of human newborn arterial smooth muscle cells (data not shown). Incubation of hSMC with the phosphodiesterase inhibitor, IBMX (500 M), increased the PDGF-stimulated PKA activity by an additional 2-fold, whereas incubation with IBMX alone only slightly elevated the PKA activity (data not shown). Similar results were found with IBMX on cAMP accumulation and suggested that PDGF was stimulating cAMP synthesis, rather than inhibiting cAMP breakdown.
The stimulation of cAMP accumulation was dose-dependent, with up to a 150-fold increase over basal levels (10 -50 pmol/ml) seen at the highest concentration (10 nM) of PDGF-BB (Fig.  1B). The PDGF isoform, PDGF-AA, also increased cAMP levels in hSMC, although not to the same extent as PDGF-BB. This is likely a reflection of the 10-fold lower number of PDGF receptor ␣ subunits compared to PDGF receptor ␤-subunits in these cells (32). Interestingly, despite the fact that hSMC also contain functional IGF-I receptors (similar in number to PDGF-␣ receptors), IGF-I, at concentrations as high as 10 nM, did not increase cAMP synthesis, even in the presence of the phospho-diesterase inhibitor IBMX (Fig. 1, A and B, and data not shown).
Connection between cAMP Synthesis, MAP Kinase Activation, Intracellular Calcium, and PKC Activity-PDGF-BB stimulated MAPK activity approximately 2.5-fold, whereas IGF-I was without effect (Fig. 2). The PDGF-induced MAPK activation was transient, with a peak of activity occurring at 5 min after stimulation in hSMC and levels returning to basal 15 min later (32). Thus, activation of MAPK correlated with activation of PKA.
Since calcium is a key second messenger in the regulation of cAMP metabolism in many cell types (reviewed in Refs. 35 and 36), we investigated whether calcium was required for the stimulation of cAMP synthesis by exposing cells to the tumor promoter, thapsigargin. Thapsigargin stimulates the release of calcium from intracellular stores, resulting in an initial increase of intracellular calcium and later (at approximately 1 h), as the calcium is transferred to the extracellular space, a depletion of calcium from intracellular stores (37). Addition of PDGF-BB to hSMC rapidly increases intracellular calcium, an event that can be inhibited by prior incubation with 300 nM thapsigargin for 1 h. 2 Preincubation with thapsigargin (300 nM, 1 h) eliminated the synthesis of cAMP stimulated by PDGF-BB, demonstrating a requirement for intracellular calcium in this process ( Fig. 2A). The effect of this compound on the activation of MAPK by PDGF was examined. Neither an extended (1-h) (Fig. 2B) nor a brief (5-min) (data not shown) exposure of cells to thapsigargin affected the basal or PDGF-stimulated levels of MAP kinase activity in these cells (Fig. 2B). However, a brief incubation with thapsigargin (5 min) did increase PKA activity above basal levels (thapsigargin, 404 Ϯ 16 pmol/min/ml; basal, 80 Ϯ 2 pmol/min/ml) and cAMP accumulation (data not shown).
Protein kinase C (PKC) has also been implicated in the regulation of cAMP formation in many cell types (reviewed in Ref. 36). Therefore, the requirement for protein kinase C (PKC) activity in PDGF-stimulated cAMP synthesis was investigated. Incubating hSMC with PMA (1 M, 20 h) resulted in the complete loss of PMA-stimulated MAPK and MAPKK activity and in keeping with a known effect of this procedure for downregulation of PKC activity (data not shown). Down-regulation of PMA-sensitive PKC activity inhibited the PDGF-stimulated increase in cAMP by Ͼ90% ( Fig. 2A), without a significant influence on the activation of MAPK by the growth factor (Fig.  2B). Similarly, addition of the PKC inhibitor (1 M), bisindoylmaleimide (GF109203X) prevented the formation of cAMP by PDGF without inhibiting that stimulated by forskolin (data not shown). Interestingly, PMA alone did not increase cAMP accumulation, although it did activate MAP kinase as expected (Fig. 2B).
Prostaglandin E 2 Synthesis Is Required for cAMP Formation-Previously, Rozengurt reported that PDGF stimulates cAMP synthesis through a prostaglandin-dependent pathway in Swiss 3T3 cells (18). Since the rate-limiting step in prostaglandin synthesis is believed to be the cPLA 2 -catalyzed release of arachidonic acid (8,9), the ability of PDGF and IGF-I to stimulate arachidonic acid release in hSMC was examined. PDGF-BB stimulated the release of [ 3 H]arachidonic acid within 2.5-5 min of addition, after which the accumulation leveled off (between 10 and 20 min) (data not shown). As shown in Fig. 3A, both PDGF-BB and PDGF-AA (less potently) stimulated arachidonic acid release, whereas IGF-I was without effect. Incubation of hSMC with thapsigargin for 1 h prior to PDGF abolished the PDGF-stimulated increase of arachidonic acid release (Fig. 3A). Previously, it was shown that cPLA 2 activity could be inhibited by a class of arachidonic acid analogues (39,40). To confirm that the release of [ 3 H]arachidonic acid was due to cPLA 2 activation, we incubated hSMC with the analogue AACOCF 3 or the noninhibitory analogue AACOCH 3 prior to the addition of PDGF. The stimulation of [ 3 H]arachidonic acid release by PDGF was significantly inhibited by AACOCF 3 (Ͼ50%), whereas the analogue AACOCH 3 was without effect (Fig. 3A).
One of the major products of arachidonic acid metabolism in hSMC is prostaglandin E2 (PGE 2 ) (38), which is formed by the action of cyclooxygenases on arachidonyl precursors. To investigate the potential role of prostaglandin release in cAMP formation, hSMC were incubated with the cyclooxygenase inhibitor, indomethacin, prior to the addition of PDGF-BB. As seen in Table I, indomethacin (10 M, 30 min) completely inhibited the PDGF-stimulated increase in cAMP and PKA activity and slightly inhibited the basal levels of cAMP. Incubation with indomethacin did not affect the activation of MAPK, suggesting that the effects of this compound were specific to inhibition of cyclooxygenase activity. These results demonstrated that prostaglandin synthesis was required for PDGF-stimulated cAMP synthesis.
To specifically investigate whether PGE 2 was involved in stimulating the increase in cAMP, PGE 2 formation was measured in response to PDGF. Within 1 min of addition, PDGF increased PGE 2 synthesis in hSMC, which continued to accumulate with extended exposure to PDGF (Fig. 3B). PGE 2 increased PKA activity by 330% as early as 1 min after addition (187.6 Ϯ 0.6 to 622.6 Ϯ 12.0 pmol/min/ml). The peak of PKA 2 K. E. Bornfeldt, unpublished observations.

FIG. 3. PDGF stimulates arachidonic acid-and PGE 2 release via a cPLA 2 and calcium-dependent mechanism.
Human SMC in 6-well plates were incubated in DMEM with 1% PDS for 48 h. In A, the cells were labeled with 1 Ci/ml [ 3 H]arachidonic acid for the last 24 h. The cells were washed three times with fresh DMEM without [ 3 H]arachidonic acid, and release of [ 3 H]arachidonic acid to the medium during a 5-min stimulation with 1 nM PDGF-BB, 1 nM PDGF-AA, 1 nM IGF-I was measured. In some instances, intracellular calcium stores were depleted by preincubation with 300 nM thapsigargin, or cPLA2 activity was inhibited by a 30-min preincubation with 30 M AACOCF 3 . The AACOCF 3 analogue AACOCH 3 (30 M) was used as a control. In B, the cells were washed 3 times with DMEM and then stimulated with 1 nM PDGF-BB (Ⅺ) or vehicle (E, 10 mM acetic acid, 0.25% bovine serum albumin) for the indicated periods of time. PGE 2 release into the medium was measured using a PGE 2 enzyme immunoassay (Amersham). The results are expressed as mean Ϯ S.D. of triplicate samples. The experiment was repeated twice with similar results. activity occurred within 5 min of PGE 2 addition, in good agreement with the ability of PDGF to rapidly increase PKA activity through PGE 2 release (Table II). PDGF-induced PGE 2 release was dependent on intracellular calcium. Thus, both arachidonic acid release and PGE 2 release stimulated by PDGF were inhibited by depletion of intracellular calcium stores using thapsigargin ( Fig. 3A and data not shown).
Down-regulation of PKC inhibited the ability of PGE 2 to stimulate PKA activity from 6.5-fold in vehicle-treated cells to 1.4-fold in cells subjected to PKC down-regulation (Table II). In contrast, neither down-regulation of PKC nor addition of the PKC inhibitor, bisindolylmaleimide, inhibited the formation of cAMP stimulated directly by forskolin (data not shown). PKC down-regulation did not inhibit PDGF-induced PGE 2 release during a 5-min stimulation (11.1 Ϯ 0.6 to 30.6 Ϯ 0.7 ng of PGE 2 released/10 6 cells subjected to PKC down-regulation compared to 3.3 Ϯ 0.4 to 25.7 Ϯ 1.4 ng of PGE 2 released/10 6 vehicletreated cells).
PDGF Stimulates MAPK Activation and cPLA 2 Phosphorylation-Since the peak of PDGF-stimulated cAMP and PKA correlated with the maximal activation of MAPK in these cells (Ref. 32 and Fig. 1), we investigated whether MAPK was directly involved in the mechanism of PKA activation. To investigate the phosphorylation of cPLA 2 by MAPK, cell extracts were subjected to Mono Q chromatography and examined for their MBP phosphorylating activity. The fractions testing positive for MAPK activity were then characterized with respect to the presence of this enzyme as shown immunochemically and by cPLA 2 phosphorylating ability. As seen in Fig. 4A, PDGF-BB potently stimulated MAP kinase activity (MBP phosphorylation), whereas IGF-I did not increase this activity. Immunoblotting of the peak fractions from PDGF-treated cells confirmed that these samples contained phosphorylated and active MAPK (Erk1 and Erk2) as judged by band shift and further demonstrated that fractions from unstimulated or IGF-I-stimulated cells did not contain detectable MAPK activity (Fig. 4B). As shown in Fig. 4, fractions from PDGF-stimulated cells contained one major peak of MBP phosphorylating activity (Fig. 4A) that co-eluted with the peak of cPLA 2 phosphorylation (Fig. 4C). Fractions from IGF-I-stimulated cells did not contain cPLA 2 phosphorylating activity above basal levels (Fig.  4C). These results suggested that the deficiency in cAMP synthesis observed with IGF-I was due to the inability of this growth factor to significantly activate MAPK.
We further investigated the phosphorylation of endogenous cPLA 2 under conditions that led to the activation of cAMP synthesis in hSMC. Phosphorylation of cPLA 2 by MAPK results in a mobility shift on SDS-PAGE that correlates with the activation of this enzyme (13). Extracts of hSMC stimulated with PDGF or IGF-I were examined by immunoblotting for cPLA 2 . As seen in Fig. 5, the cPLA 2 from untreated hSMC was a doublet similar to the baculovirus-expressed human recombinant cPLA 2 standard which is partially phosphorylated in the SF9 cells (31). Samples from PDGF-treated cells showed that cPLA 2 mobility was shifted completely relative to the untreated or IGF-I-treated samples. The lack of effect of IGF-I is consistent with the finding that IGF-I did not activate MAPK in these cells and that cPLA 2 -dependent release of arachidonic acid was not stimulated by IGF-I.
Influence of PDGF-stimulated PKA on MAPK Activation-We previously demonstrated that PKA could inhibit PDGF-stimulated MAPKK and MAPK activity in hSMC, an effect that was observed with forskolin or PGE 2 (22). We therefore investigated whether the stimulation of PKA activity by PDGF could affect the activation of the MAP kinase cascade. Addition of indomethacin (10 M, 30 min) prior to PDGF abolished the growth factor-stimulated increase in cAMP and PKA activity without affecting the magnitude of MAPKK or MAPK activation by PDGF (Table I and data not shown). Furthermore, addition of indomethacin did not alter the time course of MAPKK or MAPK activation in response to PDGF in hSMC (data not shown). DISCUSSION PDGF initiates a multitude of biological effects through the activation of intracellular signal transduction pathways such as the MAP kinase cascade, phosphatidylinositol turnover, and calcium mobilization (reviewed in Ref. 41), and these effects are believed to contribute to smooth muscle cell proliferation and directed migration (32). Further, changes in eicosanoid metabolism can regulate smooth muscle cell growth and contraction through alterations in cAMP metabolism and calcium homeostasis (reviewed in Ref. 38). How these key signal transduction pathways are integrated is not presently well understood. Because of our interest in the cross-talk between the MAPK cascade and PKA, we investigated the effects of growth factors on cAMP metabolism in hSMC. We found that PDGF induces a strong and rapid formation of cAMP through a mechanism that includes MAP kinase-mediated activation of cPLA 2 , release of arachidonic acid, prostaglandin PGE 2 , and the subsequent activation of adenylyl cyclase. Although parts of these signaling TABLE II Prostaglandin E 2 rapidly stimulates PKA activity in human SMC Cells in 10-cm dishes were preincubated with or without thapsigargin or PMA as described in Fig. 2 and then stimulated with 1 nM PDGF-BB or 10 M PGE 2 (Biomol Research Laboratories Inc., Plymouth Meeting, PA) for 5 min. PKA activity was measured as described in Fig. 1  pathways have been described previously in other cell types, we report here the complete conversion of a growth factor signal (i.e. PDGF) to cAMP in primary cultures of normal diploid cells, specifically human arterial SMC. In addition, our studies demonstrate that at least three independent signals, i.e. calcium, PKC, and MAP kinase activity are necessary for this event to occur. Several observations support these concepts. Prostaglandins, such as PGE 2 are produced from arachidonyl precursors and potently stimulate cAMP formation and PKA activity in human and other smooth muscle cells (22,38,42). In the experiments described here, the PDGF-dependent formation of cAMP required prostaglandin synthesis, as it was completely inhibited by the cyclooxygenase inhibitor indomethacin. Our results demonstrate that both arachidonic acid release and PGE 2 formation were stimulated by PDGF and that the rapid increase in PGE 2 synthesis could account for the formation of cAMP. Addition of PGE 2 was sufficient to trigger PKA activation in hSMC, and the time course of PGE 2 formation was consistent with the activation of PKA elicited by PDGF. In support of this model, Rozengurt et al. (18) reported that addition of PDGF to Swiss 3T3 cells stimulated a slow, sustained formation of cAMP that was prevented by the addition of indomethacin. However, in Swiss 3T3 cells, addition of phosphodiesterase inhibitors was required to observe this effect. In contrast, in hSMC, PDGF stimulates a potent increase in cAMP and PKA activity in the absence of phosphodiesterase inhibitors. Interestingly, cyclic AMP is a mitogen for Swiss 3T3 cells (43), whereas this nucleotide inhibits the proliferation of smooth muscle cells (21).
We examined the possibility that, in addition to effects of PDGF on prostaglandin metabolism, alternative mechanisms could facilitate the coupling of growth factor signals to cAMP synthesis in hSMC. We were unable to find a direct effect of PDGF on adenylyl cyclase activity in membranes obtained from hSMC or on the phosphorylation of the ␣-subunit of the Gprotein G s , 3 in contrast to results obtained in epidermal growth factor (EGF)-stimulated cells (44 -46). Instead, our results suggest that increased prostaglandin metabolism through the activation of cPLA 2 can account for the majority of PDGF-stimulated cAMP synthesis observed in hSMC.
Both PDGF-BB and PDGF-AA potently stimulated MAPK activity, cPLA 2 phosphorylation, arachidonic acid release, and cAMP synthesis in this study. IGF-I did not influence any of these events, despite the fact that in hSMC the number of IGF-I receptors is equivalent to the number of PDGF-␣ receptors, and that the IGF-I receptor is coupled to phosphatidylinositol turnover, calcium mobilization, and chemotaxis (32). In the present study, the inability of IGF-I to elevate cAMP correlates with an absence of effect of this growth factor on MAPK activity and cPLA 2 phosphorylation in hSMC. These experiments confirm the findings of others (13)(14)(15), demonstrating that MAPK phosphorylation is an essential signal in the activation of cPLA 2 . Recently, Sa et al. (47) reported that, in endothelial cells, basic fibroblast growth factor stimulates a MAPKdependent activation of cPLA 2 , supporting the role for MAPK in cPLA 2 regulation in other cell types.
In addition to MAP kinase activation, PKC activity and intracellular calcium mobilization were critical for the activation of PKA by PDGF in SMC. Since both cPLA 2 (13-15) and adenylate cyclase (48 -50) can be regulated by calcium and PKC in other cell types, we investigated some of the mechanisms responsible for this calcium and PKC dependence. PMA did not significantly stimulate arachidonic acid release, PGE 2 release, or cAMP synthesis, although PMA increased both MAPK and PKC activities as expected. Thus, activation of MAPK or PKC alone is insufficient for stimulation of cAMP synthesis in hSMC. PKC down-regulation blocked both PDGFinduced and PGE 2 -induced PKA activation, without affecting the ability of forskolin to activate the adenylate cyclase or the ability of PDGF to stimulate PGE 2 release. Together, these results suggest that PKC is required for PGE 2 receptor signaling to PKA activation. Depletion of calcium from intracellular stores blocked PDGF-induced arachidonic acid release, PGE 2 release, and the subsequent PKA activation in hSMC without inhibiting MAPK activation. Further, no effect on PGE 2 -stimulated PKA activation was seen, suggesting that inhibition of PDGF-induced PKA activation by intracellular calcium depletion is due to inhibition of cPLA 2 activation. In addition, transient increases in cAMP synthesis were observed in hSMC when intracellular calcium levels were increased dramatically by a short stimulation with thapsigargin or by sphingosine-1phosphate (51), for example. This effect was independent of MAPK activation, and the concentrations of intracellular cal- cium required for direct cAMP stimulation were higher than those obtained following stimulation with PDGF or IGF-I (51). 4 Possibly, high intracellular concentrations of calcium may directly stimulate the adenylate cyclase types I and III (reviewed in Ref. 35).
Previously, we reported that PKA can inhibit PDGF-stimulated MAPK signaling in hSMC (22). Therefore, we examined whether the increase in cAMP and PKA activity in response to PDGF could inhibit the MAP kinase cascade in a "feedback" manner. Such a negative feedback mechanism on MAPK (Erk2) activity has recently been demonstrated by Pyne and co-workers following bradykinin-induced cAMP accumulation (52). We were unable to find an effect of the PDGF-stimulated PKA activity on the time course of MAPKK or MAPK activation in response to PDGF, nor was the PDGF-stimulated increase in PKA activity sufficient to limit the magnitude of MAPK activation by this growth factor. At this time, we can only speculate that the PDGF-stimulated increase in PKA activity occurs too transiently to sufficiently impede the activation of MAPK in hSMC. Alternatively, the PDGF-stimulated increase in cAMP and PKA activity may be involved in PKA-mediated transcription or cytoskeletal remodeling events such as actin filament reorganization which is known to occur in response to PDGF (51).