Activation of p70 S6 kinase and erk-encoded mitogen-activated protein kinases is resistant to high cyclic nucleotide levels in Swiss 3T3 fibroblasts.

Treatment of Swiss mouse 3T3 fibroblasts with certain cyclic nucleotide phosphodiesterase inhibitors (theophylline, SQ 20,006, and MY-5445) prevents the activation of the Mr 70,000 S6 kinase (p70S6k) induced by a variety of external stimuli. Concentrations giving half-maximal inhibition were 800, 50, and 25 μM, respectively. Western blot analysis and immunocomplex kinase assays showed that these compounds inhibit the phosphorylation and activation of p70S6k without affecting the erk-encoded mitogen-activated protein (MAP) kinases or the rsk-encoded S6 kinase (p90rsk). A distinct collection of cAMP and cGMP agonists and analogues did not suppress p70S6k activation, indicating that 1) high intracellular cyclic nucleotide concentrations do not antagonize the p70S6k pathway and 2) phosphodiesterase inhibitors block p70S6k activation by a mechanism that is independent of cAMP or cGMP production. The effect of theophylline and SQ 20,006, but not MY-5445, on p70S6k signaling may be due in part to the inhibition of a phosphatidylinositol 3-kinase that acts upstream of p70S6k. Finally, in contrast to many other cell types, cAMP and cGMP were also found to have no inhibitory effect on the MAP kinase/p90rsk signaling pathway in Swiss 3T3 fibroblasts.

Treatment of Swiss mouse 3T3 fibroblasts with certain cyclic nucleotide phosphodiesterase inhibitors (theophylline, SQ 20,006, and MY-5445) prevents the activation of the M r 70,000 S6 kinase (p70 S6k ) induced by a variety of external stimuli. Concentrations giving halfmaximal inhibition were 800, 50, and 25 M, respectively. Western blot analysis and immunocomplex kinase assays showed that these compounds inhibit the phosphorylation and activation of p70 S6k without affecting the erk-encoded mitogen-activated protein (MAP) kinases or the rsk-encoded S6 kinase (p90 rsk ). A distinct collection of cAMP and cGMP agonists and analogues did not suppress p70 S6k activation, indicating that 1) high intracellular cyclic nucleotide concentrations do not antagonize the p70 S6k pathway and 2) phosphodiesterase inhibitors block p70 S6k activation by a mechanism that is independent of cAMP or cGMP production. The effect of theophylline and SQ 20,006, but not MY-5445, on p70 S6k signaling may be due in part to the inhibition of a phosphatidylinositol 3-kinase that acts upstream of p70 S6k . Finally, in contrast to many other cell types, cAMP and cGMP were also found to have no inhibitory effect on the MAP kinase/p90 rsk signaling pathway in Swiss 3T3 fibroblasts.
Addition of mitogens to quiescent mammalian cells induces a signaling cascade that leads to the multiple phosphorylation of 40 S ribosomal protein S6 (1). S6 phosphorylation is thought to increase the rate of synthesis of certain proteins which are required for efficient G 1 progression and whose mRNAs contain a polypyrimidine tract at the 5Ј end (2). Two families of mitogen-stimulated S6 kinases have been identified: the M r 70,000 S6 kinases (p70 S6k ) 1 (1,3,4) and the M r 90,000 ribosomal S6 kinase (p90 rsk ) (5). Enzymes in both families are activated by phosphorylation of serine/threonine residues (6 -8). p90 rsk is activated by mitogen-activated protein (MAP) kinases (9) and participates in a signaling network that includes ras, raf-1, and Mek1 (10). By contrast, p70 S6k lies on a distinct pathway that does not appear to include MAP kinases (11).
Injection of antibodies that neutralize p70 S6k activity (12) and use of the immunosuppressant rapamycin, which blocks the activation of the enzyme (13,14), has suggested that p70 S6k function during the G 1 phase of the cell cycle is important for proliferation in some cell types. Three serines and one threonine clustered at the carboxyl terminus of p70 S6k become phosphorylated in response to mitogen treatment (15). However, recent data have shown that deletion of the carboxyl terminus (16) or mutation of the four mitogen-induced phosphorylation sites to acidic residues 2 yields p70 S6k molecules which can still be activated by mitogens. Therefore, the contribution of these sites to enzyme activation remains unclear. In addition to the four mitogen-responsive phosphorylation sites, p70 S6k contains other phosphates that turn over very slowly and that appear to be essential for enzyme activity. Rapamycin induces the dephosphorylation of these unmapped sites and therefore prevents the activation of p70 S6k (17).
Little is known about the signaling components that function upstream of p70 S6k . One approach to identify participants in the p70 S6k pathway has been to study the mechanism of action of inhibitors of the pathway. For example, Kunz and co-workers (18) showed that rapamycin suppresses the growth of Saccharomyces cerevisiae by interacting with two gene products encoded by TOR1 and TOR2. Homologous proteins were subsequently found in higher eucaryotes (19). These proteins show significant homology to the catalytic subunit of mammalian phosphatidylinositol (PtdIns) 3-kinase (20), which plays an important role in mitogenesis and other cellular responses (21,22). It was subsequently shown that wortmannin and other specific inhibitors of mammalian PtdIns 3-kinase also prevent p70 S6k activation induced by a variety of agents (23)(24)(25). Together, these results suggested that PtdIns 3-kinase or a related enzyme might be involved in the activation of p70 S6k . This conclusion is supported by the recent observation that expression of a constitutively active PtdIns 3-kinase leads to activation of p70 S6k and phosphorylation of a novel site within the kinase catalytic domain (26).
A second possible class of inhibitors of the p70 S6k pathway was suggested by work of Thomas and co-workers (27), who showed that pretreatment of Swiss mouse 3T3 fibroblasts with theophylline or SQ 20,006 blocked the serum-induced phospho-* This work was supported in part by a grant from the Austrian Industrial Research Promotion Fund. 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  rylation of S6. We demonstrate here that these two compounds block S6 phosphorylation by selectively inhibiting the activation of p70 S6k . Theophylline and SQ 20,006 are best known as nonspecific cyclic nucleotide phosphodiesterase inhibitors (28,29) that might raise the intracellular concentration of cAMP and cGMP, leading to the activation of cAMP-and cGMP-dependent protein kinases (PKA and PKG, respectively). It has recently been shown that cAMP antagonizes p70 S6k activation in T cells (30) and the MAP kinase/p90 rsk pathway in a number of other cell types (31)(32)(33)(34)(35). However, we show that inhibition of p70 S6k activation in Swiss 3T3 fibroblasts by theophylline and SQ 20,006 is independent of increased cyclic nucleotide concentrations or PKA activation. Finally, we find that cyclic nucleotides do not negatively regulate either the p70 S6k or the MAP kinase/p90 rsk pathways in this cell type.
Wild-type S49 mouse lymphoma cells (subclone 24.3.2) were grown as described previously (38). To make extracts, 1-2 ϫ 10 7 cells were centrifuged at 500 ϫ g for 5 min, washed twice with cold phosphatebuffered saline (PBS; 137 mM NaCl, 3 mM KCl and 12 mM P i , pH 7.4), and resuspended in 400 l of extraction buffer. The cells were homogenized with 10 strokes in a Potter-Elvehjem homogenizer, and the homogenates were centrifuged at 4°C. Supernatants were retained.
BAC-1 macrophages were grown to confluence in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and colonystimulating factor-1. Cells were starved for colony-stimulating factor-1 for 27 h before use. Cells extract supernatants were prepared as described above for the fibroblasts.
For S6 kinase immunocomplex assays, extract supernatants were diluted into immunoprecipitation buffer (50 mM Tris, 1% Triton X-100, 50 mM NaCl, 20 mM NaF, 1 mM benzamidine, 5 mM EGTA, 10 mM PP i , 30 mM pNPP, 200 M vanadate, 1 g/ml leupeptin, 1 g/ml pepstatin, 0.1% BSA, 0.1 mM DTT, and 0.1 mM PMSF, pH 7.2) and the solutions were incubated with antibodies to p70 S6k for 2 h at 4°C. Protein Aagarose beads (Sigma) which had been preincubated with immunoprecipitation buffer were added, and the samples were incubated for 1 h at 4°C. The beads were washed twice with immunoprecipitation buffer and twice with S6 kinase assay buffer without DTT. S6 kinase assays were then performed as described above.
For MAP kinase immunocomplex assays, extract supernatants in immunoprecipitation buffer were incubated with antibody 122 against p42 MAPK as described above. After washing with immunoprecipitation buffer, the protein A-agarose beads were washed twice with MAP kinase assay buffer (30 mM Tris, pH 8, 20 mM MgCl 2 , 2 mM MnCl 2 , 0.1% Triton X-100, and 0.1 mM DTT). MAP kinase assays were initiated by adding 15 l of MAP kinase assay buffer containing 10 M ATP, 2 M PKI, 10 g of myelin basic protein, and 0.33 l of [␥ 32 P]ATP. After 30 min at 37°C the reactions were stopped by adding 10 l of SDS sample buffer and heating at 95°C. The reactions were subjected to electrophoresis on SDS-20% polyacrylamide gels and autoradiography.
PKA Assay-Cells were washed twice with cold PBS and homogenized with 20 strokes in 500 l of buffer containing 10 mM potassium phosphate, 100 mM KCl, 20 mM NaF, 0.5 mM theophylline, and 0.1 mM PMSF, pH 6.8. Homogenates were centrifuged at 4°C for 10 min at 8000 ϫ g, and supernatants were diluted 1:21 in PKA assay buffer (S6 kinase assay buffer plus 50 mM P i , pH 7). Protein kinase activity was determined in the presence or absence of 10 M 8-Br-cAMP or 4 M PKI. Diluted cell extract (10 l) was added to 10 l PKA assay buffer containing 100 M ATP, 20 mM pNPP, 50 M Kemptide and 0.1 l [␥-32 P]ATP. Following incubation for 10 min at 30°C, the reactions were stopped by addition of 10 l of 10 mg/ml BSA and 10 l of 7% trichloroacetic acid. After centrifugation the supernatants were pipetted onto Whatman P-81 paper. The papers were washed four times with cold 75 mM phosphoric acid, rinsed with ethanol, dried, and counted in a scintillation counter. PKA activity ratios were calculated as described previously (39).
cAMP Assay-Cells were washed twice with cold PBS and lysed in 0.1 M HCl for 30 min, and the lysates were centrifuged at 4°C for 10 min at 8000 ϫ g. cAMP was purified from the supernatants and radioimmunoassays were performed as described in the instruction manual of the RIANEN cAMP RIA Kit.
Polyclonal Antibodies to p70 S6k -A cDNA fragment encoding amino acids 258 -469 of rat p70 S6k (3) was inserted into the BamHI and HindIII sites of the vector pETH-2a (40) and expressed in Escherichia coli. The recombinant protein, which contained 12 additional amino acids at its amino terminus including 6 histidines and 2 additional amino acids at the carboxyl terminus, was purified on Ni 2ϩ -nitrilotriacetic acid-agarose (Quiagen) under denaturing conditions (41). The protein was further purified on SDS-polyacrylamide gels and used to raise antisera in rabbits. Purified recombinant protein was also coupled to CNBr-activated Sepharose 4B (Pharmacia) and used to affinity purify the antibodies (42). The purified antibodies showed no cross-reaction with p90 rsk on Western blots or in immunocomplex kinase assays. 3 Immunoblots-Proteins (8 g) in extract supernatants were resolved on SDS-15% polyacrylamide gels and electrophoretically transferred onto nitrocellulose. Membranes were blocked in 3% BSA for 1 h and incubated for 2-15 h in PBS, 0.5% Tween 20 containing purified antibody to p70 S6k (1:1000 dilution), MAP kinase R2 antibody (recognizes p42 MAPK and p44 MAPK ; 1:1000), or p90 rsk antibody (1:1000). Membranes were washed several times and further incubated with anti-rabbit horseradish peroxidase-linked whole antibodies from donkey (1:5000; Amersham Corp.) for 1 h. Specific signals were detected with the ECL kit (Amersham).
PtdIns 3-kinase Assays-The catalytic (p110␣) and regulatory (p85␤) subunits of PtdIns 3-kinase were expressed in insect cells using the baculovirus system and partially purified by ion exchange and heparin chromatography (43). Lipid kinase assays were performed using PtdIns 4,5-bisphosphate as substrate as described elsewhere (43).
Anion Exchange Chromatography-Cells treated as described in the text were washed twice with cold buffer A (25 mM Tris, 40 mM pNPP, 2 mM EGTA, 1 mM DTT, and 1 mM benzamidine, pH 7.5) containing 0.1 mM PMSF, scraped into 900 l/plate of buffer A containing PMSF and 0.1% Triton X-100 and homogenized with five strokes in a Potter-Elvehjem homogenizer. The homogenate was centrifuged at 4°C for 10 min at 8000 ϫ g, and the supernatant was applied to a 0.5-ml Mono Q column (Pharmacia Biotech Inc.) at a flow rate of 0.5 ml/min. After the column was washed with 2 ml of buffer A plus 0.1% Triton X-100, bound material was eluted with a 10-ml linear gradient from 0 to 0.5 M NaCl in buffer A plus 0.1% Triton X-100. Fractions (0.5 ml) were collected and assayed for S6 kinase activity.

RESULTS
Inhibition of S6 Kinase Activation by Theophylline and SQ 20,006 -It was shown earlier that high concentrations of the nonspecific phosphodiesterase inhibitors theophylline and SQ 20,006 block the serum-induced phosphorylation of S6 in intact fibroblasts (27). This effect could be mediated by inhibition of an S6 kinase or activation of an S6 phosphatase. To test the first possibility, quiescent fibroblasts were pretreated with or without theophylline or SQ 20,006, and then EGF was added to stimulate S6 kinase. In control cells S6 kinase was maximally activated after 10 min of EGF treatment and then the activity slowly decreased (Fig. 1A). Pretreatment with theophylline or SQ 20,006 lowered the basal level of S6 kinase activity in unstimulated cells and greatly reduced the EGF-induced activation of the enzyme (Fig. 1A). Similar results were obtained with 2 mM IBMX, another nonspecific phosphodiesterase inhibitor that is structurally related to theophylline. 3 Dose-response curves showed that the concentrations of theophylline and SQ 20,006 that inhibit S6 kinase activation 50% (IC 50 ) were 800 M and 50 M, respectively (Fig. 2). Thus, theophylline and SQ 20,006 block S6 phosphorylation in vivo by causing the inhibition of an S6 kinase.
Selective Inhibition of p70 S6k Activation by Theophylline and SQ 20,006 -Mouse fibroblasts contain both p90 rsk and p70 S6k (44). The majority of S6 kinase activity measured in cell extracts using 40 S ribosomal subunits as a substrate is contributed by p70 S6k . To determine which of these kinases is sensitive to theophylline and SQ 20,006, p90 rsk and p70 S6k in fibroblast extracts were assayed after separating the two enzymes on an anion exchange column (Fig. 3A). The identity of p90 rsk and p70 S6k was confirmed by immunocomplex kinase assays using antibodies specific for each enzyme. 3 Untreated quiescent cells showed a low level of S6 kinase activity, all of which appeared in a peak corresponding to p70 S6k (Fig. 3A,   15). Treatment with SQ 20,006 almost completely abolished this basal activity. When control cells were exposed to EGF for a short time to maximally activate p90 rsk (see Fig. 6C) a small peak of activity corresponding to this enzyme appeared in fraction 7; p70 S6k was activated to a much larger extent (Fig.  3A, open circles). Pretreatment of cells with SQ 20,006 almost completely blocked the EGF-induced activation of p70 S6k but did not inhibit p90 rsk (Fig. 3A, closed circles). Virtually identical results were obtained with theophylline. 3 Immunocomplex kinase assays of p70 S6k also showed that this enzyme could no longer be activated by EGF in cells pretreated with theophylline or SQ 20,006 (Fig. 3B, upper panel, lanes 1-4). Thus, the effect of SQ 20,006 and theophylline seen in extracts (Fig. 1A) is due to the selective inhibition of p70 S6k .
All evidence obtained so far indicates that p70 S6k is activated by phosphorylation (7,15,17,26). Therefore, theophylline and SQ 20,006 might reduce p70 S6k activity either by affecting its phosphorylation or by stimulating the degradation of the enzyme. To distinguish between these two possibilities, we examined the protein levels and phosphorylation state of p70 S6k on immunoblots. p70 S6k migrates differently in SDS-polyacrylamide gels depending on its phosphorylation state, with the highly phosphorylated and active form migrating more slowly than the dephosphorylated, inactive enzyme (7). In resting cells the kinase was present mainly as hypophosphorylated species and kinase activity in immunoprecipitates was low (Fig. 3B, lane 1). After EGF treatment p70 S6k activity increased and only the most highly phosphorylated forms were detected on the immunoblot (Fig. 3B, lane 2). When EGF was added to cells pretreated with theophylline or SQ 20,006 most of the p70 S6k molecules remained hypophosphorylated and the kinase activity did not increase (Fig. 3B, lanes 3 and 4). These data demonstrate that theophylline and SQ 20,006 either prevent the EGF-induced phosphorylation of p70 S6k or stimulate its dephosphorylation. In addition, no change in p70 S6k protein levels was seen after treatment with theophylline or SQ 20,006 (Fig. 3B, lower panel), indicating that these agents do not induce the degradation of the enzyme.
p70 S6k Activation Is Resistant to High Levels of cAMP and PKA Activity-Since theophylline and SQ 20,006 act as cyclic nucleotide phosphodiesterase inhibitors in vitro (28,29), the negative regulation of p70 S6k by these two compounds in intact cells might be mediated by cAMP and PKA. However, cAMP and PKA did not inhibit p70 S6k activity directly in an in vitro assay. 3 To determine whether high cAMP levels or PKA activity antagonize the p70 S6k pathway in vivo, several specific cAMP agonists and analogues were tested for the ability to block S6 kinase activation. The compounds tested included PGE 1 , a hormone that activates adenylyl cyclase through a specific receptor, and 8-Br-cAMP and S p -8-Br-cAMPS, which are cell-permeant, hydrolysis-resistant cAMP analogues. To our surprise, none of these compounds was an effective inhibitor of the EGF-induced activation of p70 S6k (Fig. 1B). Immunocomplex kinase assays and Western analysis confirmed that the EGF-induced phosphorylation and activation of p70 S6k still occurred in cells pretreated with 8-Br-cAMP, S p -8-Br-cAMPS or PGE 1 (Fig. 3B, lanes 5-7). PKA assays and cAMP radioimmunoassays indicated that, as expected, PGE 1 , 8-Br-cAMP, and S p -8-Br-cAMPS activated PKA, and PGE 1 increased cAMP levels (Table I). However, despite the known ability of theophylline and SQ 20,006 to inhibit phosphodiesterases in vitro (28,29), we did not detect a significant activation of PKA or an increase in cAMP levels after treatment of cells with these two compounds (Table I).
In addition to EGF, p70 S6k is activated by a wide variety of agonists that act by different mechanisms. These include pep-tide growth factors that function through receptor-tyrosine kinases (insulin and PDGF) or receptors coupled to guanosine nucleotide-binding proteins (bombesin), the Ca 2ϩ ionophore A23187, the protein kinase C activator PMA and cycloheximide. To determine whether 8-Br-cAMP inhibits p70 S6k activation induced by any of these other compounds, cells were pretreated with or without 8-Br-cAMP and then exposed to one of these p70 S6k agonists. Cell extracts were prepared at those times when p70 S6k was maximally stimulated. Kinase activity was increased to distinct levels and pretreatment with 8-Br-cAMP had little or no effect on the activation of p70 S6k induced by any of the compounds tested (Table II). By contrast, theophylline strongly inhibited the activation of p70 S6k by every agonist tested (Table II). Three conclusions can be made from these results. First, cAMP and PKA do not antagonize the p70 S6k activation pathway in fibroblasts. Second, theophylline and SQ 20,006 block p70 S6k signaling by a mechanism that is independent of cAMP or PKA. And third, the target of theophylline seems to be a common regulatory element in p70 S6k signaling pathways induced by different agonists.
Effect of cAMP on p70 S6k Activation in Other Cell Types-In contrast to our results in fibroblasts ( Fig. 3B and Table II), cAMP has been reported to block the activation of p70 S6k in T cells (30). T cells are unlike Swiss fibroblasts in that they undergo growth arrest in response to high intracellular cAMP concentrations (45). We therefore determined if an inhibitory effect of cAMP on p70 S6k also occurs in two additional cell lines whose proliferation is sensitive to cAMP (38,46). Cycling wildtype S49 mouse lymphoma cells (38) exhibited a relatively high level of S6 kinase activity that was not significantly reduced in the presence of 8-Br-cAMP (Fig. 4). Similar results were obtained with cycling Swiss fibroblasts (Fig. 4). In addition, the activation of S6 kinase in BAC-1 macrophages (46) induced by colony-stimulating factor-1 was not decreased but rather augmented by 8-Br-cAMP (Fig. 4). Chromatography of cell extract supernatants on a Mono Q column (see Fig. 3A) confirmed that the major S6 kinase activity in lymphocytes and macrophages is contributed by p70 S6k . 3 Thus, p70 S6k can be activated in the presence of high intracellular cAMP concentrations in some cell types whose proliferation is arrested by cAMP.
Activation of p70 S6k Is Resistant to High cGMP Levels-Theophylline and SQ 20,006 also inhibit cGMP-specific phosphodiesterases (28,29) and by this mechanism may increase cGMP levels and activate PKG. We therefore tested specific cGMP agonists for the ability to inhibit p70 S6k activation. MY-5445, an inhibitor of cGMP-specific phosphodiesterases, completely blocked the activation of p70 S6k by EGF (Fig. 1C) with an IC 50 of 25 M (Fig. 2). By contrast, SNAP, an activator of guanylyl cyclase, and 8-Br-cGMP showed little effect on p70 S6k activity (Fig. 1C). Therefore, no correlation between increased cGMP levels and suppression of the p70 S6k activation pathway was evident. Immunocomplex kinase assays and Western analysis confirmed that inhibition of the EGF-induced activation of p70 S6k by MY-5445 was not mimicked by 8-Br-cGMP (Fig. 3B,  lanes 8 and 9). In addition, 8-Br-cGMP did not reduce the stimulation of p70 S6k in response to other agonists (Table II).
These results indicate that cGMP does not negatively regulate p70 S6k . Theophylline and SQ 20,006 Inhibit PtdIns 3-Kinase in Vitro-A variety of evidence suggests that a PtdIns 3-kinase can be involved in the activation of p70 S6k (23)(24)(25)(26). We therefore tested whether the three inhibitors of p70 S6k activation identified here have an effect on PtdIns 3-kinase activity in vitro. Dose-response experiments showed that theophylline and SQ 20,006 significantly inhibited lipid kinase activity in vitro, with IC 50 values of 810 M and 183 M, respectively (Fig.  5). MY-5445 did not inhibit PtdIns 3-kinase activity at concentrations up to 1.25 mM (Fig. 5). Thus, one in vivo target of theophylline and SQ 20,006 might be a PtdIns 3-kinase that functions upstream of p70 S6k .
No Inhibition of MAP Kinase Activation by Cyclic Nucleotide Reagents-It has been reported that cAMP antagonizes the activation of MAP kinases in a number of different cell types (31)(32)(33)(34). We therefore tested whether cAMP has a similar effect in Swiss fibroblasts. To determine the conditions under which MAP kinases become fully activated, a time course of kinase activation was visualized on Western blots probed with antibodies that recognize two erk-encoded MAP kinases referred to as p42 MAPK and p44 MAPK . The two MAP kinases were fully activated 2.5 min after addition of EGF, as indicated by their shift to a slower mobility due to phosphorylation (Fig. 6A). The kinases remained maximally active for up to 5 min and then dephosphorylated species appeared after 10 min (Fig. 6A). Subsequent experiments were done using extracts from fibroblasts stimulated for 5 min with EGF.
To directly examine the effect of cyclic nucleotide reagents on MAP kinase activity, the erk-2-encoded p42 MAPK was precipitated with a specific polyclonal antibody and assayed in immunocomplexes. These assays showed that p42 MAPK was strongly activated by EGF (Fig. 6B, upper panel, lanes 1 and 2). This response was still fully intact in cells pretreated with cAMP analogues or PGE 1 (Fig. 6B, upper panel, lanes 5-7). p42 MAPK was also activated normally in cells pretreated with 8-Br-cGMP (Fig. 6B, upper panel, lane 8). In contrast to the results obtained with p70 S6k (Figs. 1 and 3B), pretreatment of cells with theophylline, SQ 20,006, or MY-5445 did not block the activation of p42 MAPK by EGF (Fig. 6B, upper panel, lanes 3, 4,  and 9). Thus, these compounds do not disrupt all EGF receptormediated responses. We also tested whether MAP kinase activation induced by insulin or PDGF is affected by cAMP. Although these two growth factors were relatively poor activators of p42 MAPK at these concentrations, it was evident that addition of 8-Br-cAMP had no inhibitory effect (Fig. 6B, upper  panel, lanes 10 -13). Examination of MAP kinases in these cell extracts on Western blots indicated that the activation of p44 MAPK , like p42 MAPK , was also resistant to high cyclic nucleotide levels and phosphodiesterase inhibitors (Fig. 6B, lower  panel). In addition, the growth factor-induced mobility shift of p90 rsk was not affected by these reagents (Fig. 6C); these results confirm and extend the finding in Fig. 3A. Thus, in Swiss 3T3 fibroblasts activation of the erk-encoded MAP kinases and p90 rsk is not antagonized by cAMP or cGMP. DISCUSSION It was reported earlier that treatment of Swiss mouse 3T3 fibroblasts with the nonselective phosphodiesterase inhibitors theophylline and SQ 20,006 blocks the mitogen-induced phosphorylation of ribosomal protein S6 (27). We show here that these two compounds mediate this effect by inhibiting the activation of p70 S6k (Fig. 3). Theophylline and SQ 20,006 did not act as general kinase inhibitors or disrupt all receptormediated responses, as shown by the fact that activation of  MAP kinases and p90 rsk was not inhibited (Figs. 3A and 6). Thus, like rapamycin (13,14) and wortmannin (23)(24)(25), theophylline and SQ 20,006 are selective inhibitors of the p70 S6k pathway as opposed to the MAP kinase/p90 rsk pathway. Also like rapamycin and wortmannin, theophylline and SQ 20,006 did not inhibit p70 S6k directly 3 , but rather caused a reduction in the phosphate content of the protein (Fig. 3B). This could be due to activation of a negative regulator of the p70 S6k pathway or inhibition of a positive regulator. Inhibition by theophylline was rapidly reversed by washing out the inhibitor even in the presence of cycloheximide, 3 indicating that all of the components required for p70 S6k activation are still present in theophylline-treated cells.
Theophylline and SQ 20,006 inhibit cyclic nucleotide phosphodiesterases in vitro (28,29). Therefore, one hypothesis to explain the mechanism of inhibition of p70 S6k would be that the compounds increase intracellular cAMP levels, and as a result PKA is activated and phosphorylates a regulatory protein in the p70 S6k pathway. However, several lines of evidence indicate that cAMP/PKA do not negatively regulate p70 S6k in these cells. First, theophylline and SQ 20,006 were not found to be effective cAMP agonists under conditions in which p70 S6k was inhibited (Table I). This result is consistent with other reports that theophylline has little or no effect on basal cAMP levels in mouse fibroblasts (47). Second, increasing the intracellular cAMP concentration by addition of cell-permeant cAMP analogues or PGE 1 did not block the stimulation of p70 S6k induced by EGF (Fig. 1B) or other agonists (Table II), even though PKA was strongly activated (Table I). These results are consistent with the observation made earlier that PGE 1 does not inhibit serum-induced S6 phosphorylation in vivo (27). Finally, work by others has shown that cAMP/PKA promotes the proliferation of Swiss mouse 3T3 fibroblasts (48); therefore, this pathway would not be expected to interfere with the activation of p70 S6k , which is also required for efficient cell cycle progression in this cell type (12,13).
In contrast to our results, Monfar et al. (30) recently reported that cAMP inhibits the interleukin-2-mediated activation of p70 S6k in CTLL-20 T cells, in part by antagonizing the activity of PtdIns 3-kinase. This discrepancy could be due to cell type or agonist specificity. Unlike Swiss fibroblasts, T cells arrest in the G 1 phase of the cell cycle in response to high cAMP concentrations (45) or rapamycin treatment (49). Since rapamycin inhibits interleukin-2-mediated p70 S6k activation (14), it seems plausible that the p70 S6k pathway could also be the target for a cAMP-dependent cell cycle block in T cells. On the other hand, Monfar et al. (30) used forskolin/IBMX as a cAMP agonist, raising the possibility that p70 S6k activation in CTLL-20 T cells might be sensitive to forskolin/IBMX and not to cAMP per se. Our finding that 8-Br-cAMP has no effect on p70 S6k activity in a T cell lymphoma line (Fig. 4) also suggests that the sensitivity of T cell p70 S6k to IBMX/forskolin might be independent of cAMP. Macrophages also arrest in the G 1 phase of the cell cycle in response to high cAMP levels or exposure to rapamycin (46). We found that 8-Br-cAMP enhances rather than inhibits the activation of p70 S6k by colony-stimulating factor-1 in these cells (Fig. 4).
Since theophylline and SQ 20,006 also inhibit cGMP-specific phosphodiesterases in vitro (28,29), we explored the possibility that cGMP and PKG might be involved in the inhibition of p70 S6k activation. Although MY-5445, a selective inhibitor of cGMP-specific phosphodiesterases, strongly blocked the activation of p70 S6k , 8-Br-cGMP and SNAP, a guanylyl cyclase activator, had no effect ( Fig. 1C and Table II). These results strongly suggest that inhibition of p70 S6k by MY-5445 is not a consequence of increased cGMP production or PKG activity.
Having ruled out the possibility that cAMP/PKA or cGMP play a role in theophylline-induced p70 S6k inhibition, we searched for alternative mechanisms that might mediate this response. PtdIns 3-kinase, which has been proposed to act as a positive upstream regulator of p70 S6k (23)(24)(25)(26), seemed to be a likely target. Indeed, we show here that theophylline and SQ 20,006, but not MY-5445, also inhibit PtdIns 3-kinase activity in vitro (Fig. 5). However, inhibition of PtdIns 3-kinase cannot be the only reason why theophylline blocks the p70 S6k pathway. We showed earlier that wortmannin, a specific inhibitor of PtdIns 3-kinase, is a poor inhibitor of bombesin-and PMAinduced p70 S6k activation (25). We concluded from those studies that some pathways leading to p70 S6k activation are independent of PtdIns 3-kinase (25). By contrast, theophylline strongly inhibited the p70 S6k response to all of the agonists listed in Table II, including bombesin and PMA. These data suggest that theophylline might act on another target in addition to PtdIns 3-kinase. This target appears to be a regulatory element that functions either independently of all activation pathways (such as a p70 S6k phosphatase) or in every pathway (such as a p70 S6k kinase).
Theophylline has several other known cellular effects that could play a role in its ability to inhibit the activation of p70 S6k . Theophylline mobilizes intracellular Ca 2ϩ by opening ryanodine receptor Ca 2ϩ channels (50) and is a potent antagonist of adenosine receptors (29,51). The possible role of Ca 2ϩ or adenosine receptors in the activation of p70 S6k remains to be tested. More importantly, many purine analogues and derivatives such as theophylline are competitive inhibitors of kinases due to a structural similarity to ATP (52). This property may account for the ability of theophylline to inhibit PtdIns 3-kinase activity in vitro (Fig. 5). Therefore, theophylline may inhibit the p70 S6k pathway by targeting an upstream kinase that is required for a response to all agonists tested so far. Since no such upstream component has been identified yet, this hypothesis cannot be tested directly. SQ 20,006 and MY-5445 bear some structural resemblance to theophylline and may act in a similar manner. Identification or design of specific and potent inhibitors of p70 S6k activation, based on the compounds we identified here, could lead to the isolation of upstream activators that function in this pathway.
We show here that the activation of p42 MAPK /p44 MAPK and p90 rsk in Swiss fibroblasts is also resistant to high cAMP and cGMP levels (Fig. 6). MAP kinase activation in PC12 cells is similarly resistant to high cAMP levels (31). By contrast, negative regulation of the MAP kinase/p90 rsk cascade by cAMP has been observed in Xenopus oocytes, Rat1 cells, smooth muscle cells, CHO cells, and adipocytes stimulated with a variety of agonists (32)(33)(34)(35). In some of these cell types high cAMP levels also antagonize cell proliferation (33,34) or meiotic cell division (35). Several different mechanisms may contribute to inhibition of MAP kinase activation, including a reduced activation of Raf-1 due to phosphorylation by PKA (34).
Our results, together with those from other laboratories, indicate that cyclic nucleotide-dependent signaling may interact in a variety of ways with the p70 S6k and MAP kinase/p90 rsk pathways. These interactions may be important for generating cell type-specific responses to identical physiological conditions. Identification of regulatory components in the p70 S6k pathway will allow us to further elucidate the mechanisms of cell type-specific control.