Expression of phosphodiesterase 4D (PDE4D) is regulated by both the cyclic AMP-dependent protein kinase and mitogen-activated protein kinase signaling pathways. A potential mechanism allowing for the coordinated regulation of PDE4D activity and expression in cells.

Multiple families of cyclic nucleotide phosphodiesterases (PDE) have been described, and the regulated expression of these genes in cells is complex. Although cAMP is known to control the expression of certain PDE in cells, presumably reflecting a system of feedback on cAMP signaling, relatively little is known about the influence of non-cAMP signaling systems on PDE expression. In this study, we describe a novel mechanism by which activators of the protein kinase C (PKC)-Raf-MEK-ERK cascade regulate phosphodiesterase 4D (PDE4D) expression in vascular smooth muscle cells (VSMC) and assess the functional consequences of this effect. Whereas a prolonged elevation of cAMP in VSMC resulted in a protein kinase A (PKA)-dependent induction of expression of two PDE4D variants (PDE4D1 and PDE4D2), simultaneous activation of both the cAMP-PKA and PKC-Raf-MEK-ERK signaling cascades blunted this cAMP-mediated increase in PDE4D expression. By using biochemical, molecular biological, and pharmacological approaches, we demonstrate that this PDE4D-selective effect of activators of the PKC-Raf-MEK-ERK cascade was mediated through a mechanism involving altered PDE4D mRNA stability and markedly attenuated the cAMP-mediated desensitization that results from prolonged activation of the cAMP signaling system in cells. The data are presented in the context of activators of the PKC-Raf-MEK-ERK cascade having both short and long term effects on PDE4D activity and expression in cells that may influence cAMP signaling.

The cyclic nucleotides, cAMP and cGMP, control numerous physiological processes including intermediary metabolism and cellular proliferation, motility, or contractility. Cyclic nucleotide-mediated signaling is terminated either when the stimulus for generation of cAMP or cGMP is removed or when cyclic nucleotides are hydrolyzed. Hydrolysis of cAMP or cGMP is catalyzed by members of a multigene family of enzymes called cyclic nucleotide phosphodiesterases (PDE). 1 In mammals, 10 individual families of PDE have been described with each classified on the basis of molecular sequence, substrate selectivity, and inhibitor sensitivity (1)(2)(3)(4). Each PDE family has been shown to comprise several individual genes, each of which can be alternatively processed to yield several different mRNA. In most instances members of several PDE families are expressed in individual cells, with differential subcellular targeting, and selective regulation by phosphorylation by various protein kinases emerging as important factors in determining their individual roles (5)(6)(7)(8).
Due to the important role played by both cAMP and cGMP in the regulation of myocardial contractility and blood vessel relaxation, the PDE present in these tissues have received a considerable amount of attention as therapeutic targets (9,10). Members of at least four PDE families (PDE1, PDE3, PDE4, and PDE5) have been shown to contribute to termination of cyclic nucleotide signaling in vascular smooth muscle cells (VSMC) (10). Whereas enzymes of the PDE1 and PDE5 families hydrolyze cGMP, members of the PDE1, PDE3, and PDE4 families catalyze the breakdown of cAMP. A large body of evidence supports important roles for PDE3 and PDE4 in the regulation of cAMP-mediated effects in VSMC, and much of this evidence is consistent with the idea that these enzymes regulate separate, although overlapping, cAMP pools in cells (10 -26). Thus, a strong correlation exists between the increase in cAMP which occurs as a result of the pharmacological inhibition of PDE3 activity and inhibition of VSMC contractions (10,11,14,15). In contrast, although pharmacological inhibition of PDE4 raises VSMC cAMP levels, and PDE4 inhibitors inhibit VSMC proliferation (16,20) and migration (21) of these cells, these agents are generally poor relaxants (10,11,14). Interestingly, synergistic effects of simultaneous inhibition of PDE3 and PDE4 activities have been reported (11,14,18). In this context, dual-selectivity PDE3/PDE4 inhibitors may prove useful agents in cells expressing members of both these enzyme families (9,10,27).
The regulation of PDE4 activity and expression in cells is complex (7)(8). In both human and rat four distinct genes encode PDE4 enzymes (PDE4A, PDE4B, PDE4C, and PDE4D), with each, as a result of alternate splicing or the use of alternate promoters, giving rise to multiple variants (7)(8). Recently, a significant amount has been elucidated concerning the regulation of PDE4 activity and expression in cells. Indeed, it is now clear that selective PDE4 variants can be regulated by transcriptional and/or by post-translational mechanisms (7)(8).
Presumably reflecting a form of negative feedback for cAMP signaling, phosphorylation and activation of certain PDE4 variants by the cAMP-dependent protein kinase (PKA) have been reported (7, 8, 28 -32). Although a mitogen-activated protein kinase-mediated phosphorylation of a PDE4B variant had no functional consequences (34), activation of the mitogen-activated protein kinase cascade has recently been shown to affect PDE4D3 activity, perhaps in a cell-dependent manner (32)(33). While in HEK293 cells, ERK2-mediated phosphorylation led to a transient inhibition of PDE4D3 activity (33) and activation of the PKC-Raf-MEK-ERK cascade in rat aortic VSMC activated and translocated particulate PDE4D3 to the cytosol of these cells (32). In addition to this acute effect of cAMP elevation, more prolonged increases in cellular cAMP have been shown to bring about marked increases in PDE4 activity in several cell types (18,(35)(36)(37)(38)(39)(40)(41). Interestingly, although these treatments always result in elevated levels of PDE4 activity, the PDE4 gene responsible appears to be cell type-specific. For example, while incubation with forskolin decreased the expression level of PDE4A and increased that of PDE4D in Jurkat cells (40), a similar treatment of selected monocyte cell lines resulted in marked elevations in PDE4A expression and increases, or decreases, in PDE4D expression (38 -39). In addition, changes in PDE4B expression have also been reported following treatment with cAMP-elevating agents. Thus, treatment of rats with a prostacyclin analogue caused a marked increase, and decrease, in myocardial PDE4B and PDE4D, respectively (42). Recently, we have begun to investigate the role of cAMP in controlling PDE4 activity and expression in VSMC (18 -19, 21, 32). As a result of these studies, we recently reported that the two PDE4D variants expressed in rat aortic VSMC, PDE4D3 and PDE4D5, were activated following phosphorylation by protein kinase A (PKA) (32). Interestingly, a selective PKC-Raf-MEK-ERK cascade-mediated phosphorylation of the particulate fraction of PDE4D3 in these cells was shown to increase the efficiency with which this enzyme was phosphorylated by PKA and to result in the translocation of this particulate PDE4D3 (32). In the work described here, we have investigated the impact of the PKC-Raf-MEK-ERK cascade on the effects of prolonged increases in cAMP elevation on PDE4D activity and expression in VSMC. We show that a prolonged increase in cAMP results in a marked increase in the expression of two short variants of the PDE4D gene (PDE4D1 and PDE4D2) and that activation of the PKC-Raf-MEK-ERK cascade blunts this increase. Our data are discussed in the context of the PKC-Raf-MEK-ERK cascade having a biphasic effect on PDE4D activity in VSMC, increasing it in the short term, but decreasing it in the long term, and are consistent with a paradigm in which PDE4D activity and expression are subject to regulation by multiple and overlapping signaling systems.  ). Tris-HCl, benzamidine, EDTA, EGTA, dithiothreitol, phenylmethylsulfonyl fluoride, Triton X-100, and NaCl were from ICN Biomedicals (Costa Mesa, CA), and leupeptin, Affi-Gel 601, Dowex 50 (200 -400 mesh), alumina, and the columns supports were from Bio-Rad. The BCA protein assay and bovine serum albumin were from Pierce. All other chemicals were of reagent grade and purchased from Fisher. PDE4D-specific monoclonal antibody was provided by ICOS Corp. (Bothell, WA).

Materials
Cell Culture-Primary cultures of rat aortic VSMC were prepared as described previously (18). VSMC were routinely cultured in Dulbecco's modified Eagle's medium supplemented with 10% calf serum, 8 mM HEPES buffer, 100 units/ml penicillin, and 100 g/ml streptomycin at 37°C in a 95% air, 5% CO 2 humidified atmosphere. For all experiments, cells were used between passages 4 and 12.
Treatment of VSMC with Pharmacological Agents-Culture media were removed and replaced with fresh culture media supplemented with either (i) forskolin (0.1-100 M), (ii) 8-Br-cAMP (0.01-1 mM), (iii) PMA (0.1-100 nM), (iv) angiotensin II (0.1-1.0 M), or (v) vehicle (0.1% dimethyl sulfoxide (Me 2 SO)). At the end of the incubation period, treated cells were washed with HBSS (with Ca 2ϩ and Mg 2ϩ ) and harvested in lysis buffer containing 50 mM Tris-HCl (pH 7.4), 5 mM MgCl 2 , 5 mM benzamidine, 1 mM EDTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, 1 M leupeptin, and 1% Triton X-100. Cells were removed from the flask by scraping. Cellular debris and unlysed cells were removed by centrifugation at 1,000 ϫ g for 5 min at 4°C. The 1,000 ϫ g supernatant was transferred to microtubes and stored at 4°C until assayed for cAMP PDE activity (see below). For some experiments cytosolic and particulate fractions were prepared. In these instances, rat aortic VSMC were lysed as described above except that Triton X-100 was excluded from the lysis and the 1,000 ϫ g supernatant subjected to a further centrifugation step (132,000 ϫ g).
Assay of cAMP Phosphodiesterase Activity-cAMP phosphodiesterase activity was assayed as described previously (18). Briefly, reactions were carried out in a total volume of 100 l containing 5 mol of Tris-HCl (pH 7.4), 0.5 mol of MgCl 2 , 10 nmol of EGTA, and 0.1 nmol of [ 3 H]cAMP containing 55,000 -80,000 dpm for 30 min at 30°C. Reaction were terminated by addition of 50 l of 0.5 M ice-cold EDTA (pH 7.4) supplemented with 5Ј-[ 14 C]AMP (1,800 dpm). Samples were diluted with 0.3 ml of HEPES-NaCl buffer (0.1 M NaCl, 0.1 M HEPES (pH 8.5)) prior to purification of the product of the reaction, 5Ј-[ 3 H]AMP. 5Ј-[ 3 H]AMP and 5Ј-[ 14 C]AMP were recovered by chromatography using a polyacrylamide-boronate gel column (Affi-Gel 601, 1-ml bed volume), and the purified 5Ј-[ 3 H]AMP was quantified by liquid scintillation counting, corrected for recovery of 5Ј-[ 14 C]AMP, normalized to the total protein used in the assay, and the total activity expressed as pmol min Ϫ1 mg Ϫ1 of protein. Total protein concentration of each sample was determined using the BCA Protein Assay system from Pierce, according to the manufacturer's methodology using bovine serum albumin as the standard.
Measurements of VSMC cAMP-In experiments in which the impact of forskolin, PMA, AngII, or a combination of these agents on VSMC cAMP levels was determined, the method of Maurice et al. (43)  Reverse Transcription and Amplification by Polymerase Chain Reaction (RT-PCR)-RNA purification and RT-PCRs were carried out as described previously (19). Briefly, RNA was purified from cultured rat aortic VSMC (5-10 ϫ 10 6 cells), using a single step procedure, Trizol Reagent (Life Technologies, Inc). First strand cDNA was generated from 5 to 10 g of total RNA using oligo(dT) 18 (Cortec, Kingston, Ontario) to prime the reverse transcription (SuperScript Moloney mu-rine leukemia virus-reverse transcriptase). Amplification reactions were performed using Taq DNA polymerase with 1-10 l of the first strand reaction (total volume 100 l) and 20 pmol each of PDE4D splice variant-specific sense and antisense oligonucleotide primers. Conditions for PCR were 30 s at 95°C, 30 s at 58°C, and 1 min at 72°C for a number of cycles that were determined empirically to allow linear amplification of the relevant template (PDE4D3 (22-30 cycles), PDE4D1 (25-31 cycles), or PDE4D2 (27-32 cycles)). In some experiments, 1 Ci of [␣-32 P]dCTP was added to assist in quantitating amounts of PCR products generated. All PCRs were carried out with the same antisense primer (5Ј-CGCTAGTACAATGTCAATGGCC-3Ј) and one sense primer specific for the required PDE4D gene splice variant PDE4D3 (5-TCCTGGATATGTTTCGATGTGG-3Ј), PDE4D2 (5Ј-ACGG-CCTCCAACAAGTTCAAG-3Ј), and PDE4D1 (5Ј-AAGCGCTTAAGAAC-TGAGTCC). Amplification of RNA encoding glyceraldehyde-3phosphate dehydrogenase (GAPDH) (5Ј-AGCATCAAAGGTGGAAGAA-TG-3Ј; 5Ј-GGTTGCCATCAACGACCCCTT-3Ј) served as a control. When stability of PDE4D gene products was assessed, RNA was purified from forskolin-treated rat aortic VSMC at 10-min intervals following addition of actinomycin D (4 M) for a total of 280 min and processed as described above. PCR products were separated by electrophoresis on 1.5% agarose gels and visualized with ethidium bromide. Amounts of PDE4D RNA amplified were determined by scanning densitometry using Corel Photo-Paint 8.0 software as per manufacturer's recommendations and liquid scintillation counting of isolated bands.
Immunoblotting-VSMC cultures incubated with compounds of interest were homogenized in a buffer consisting of 20 mM Tris-HCl (pH 7.5), 1 mM MgCl 2 , 0.1 mM EGTA, 5 mM benzamidine, 1 g/ml aprotinin, and 1 g/ml leupeptin. These samples (5-20 g of total homogenate protein or of protein from isolated subcellular fractions) were subjected to SDS-PAGE. Following electrophoresis, proteins were transferred to nitrocellulose membranes (Bio-Rad), and the membranes were blocked by incubation with TBST (20 mM Tris (pH 7.5), 100 mM NaCl, 0.1% Tween 20) supplemented with 5% powdered Nonfat milk for 1 h. Blots were incubated with an appropriate dilution of PDE4D antibody for 1-2 h and rinsed three times with TBST. Rinsed blots were incubated with horseradish peroxidase-conjugated goat anti-mouse IgG for 1 h and then rinsed with TBST, and immunoreactivity was detected by chemiluminescence as per manufacturer's recommendations. When available, purified recombinant human PDE4D1 or PDE4D3 (ICOS Corp., Bothell, WA) was used to determine the specificity of the antiserum and for comparison of molecular masses.
Statistical Analysis-Data are presented as means Ϯ S.E. of at least three independent experiments. Within each experiment, values were means from three individual determinations for each experimental condition. Statistical differences between cAMP PDE activities or between [ 3 H]cAMP levels were determined using the Student's t test for either paired or unpaired samples, with p Ͻ 0.05 considered significant.

Incubation of Rat Aortic VSMC with Forskolin or 8-Br-cAMP
Increases PDE4 Activity-Consistent with our previous work (18 -19, 32), treatment of rat aortic VSMC with forskolin or 8-Br-cAMP caused a time-and concentration-dependent increase in total cAMP PDE activity, with marked increases in both PDE3 (18 -19) and PDE4 (Table I). By using a selective inhibitor of PDE3 (cilostamide, 1 M) or of PDE4 (Ro 20-1724, 10 M) to measure these activities, a 16-h incubation of VSMC with forskolin (10 M) caused increases of 210 Ϯ 22 and 108 Ϯ 17% in PDE3 (not shown) and PDE4 (Table I), respectively. Recently, we reported that both the cytosolic PDE3A as well as the particulate PDE3B were increased in rat aortic VSMC following prolonged increases in cAMP (19). Changes in cytosolic and particulate PDE4 activities were also noted in our experiments (Table I). Although some of the increase in PDE4 at short times could be due to PKA-mediated phosphorylation of PDE4D3 and PDE4D5 (32), since the increases using our treatment protocol were inhibited by addition of either actinomycin D or cycloheximide, a role for de novo mRNA and protein synthesis was established (Table II).
Forskolin or 8-Br-cAMP Induces the Expression of PDE4D1 and PDE4D2 in Rat Aortic VSMC-A strategy of selective immunoblotting with PDE4-selective antisera was used to determine the identity of the PDE4 variant being affected by forskolin, or 8-Br-cAMP, treatment of rat aortic VSMC. By using this approach, we determined that the major effect of prolonged increases in cAMP in VSMC was related to the induction of expression of two PDE4D variants not expressed in untreated control cells. Thus, while untreated control VSMC expressed both 95 Ϯ 2 and 105 Ϯ 3 kDa anti-PDE4D-immunoreactive proteins, which we have previously shown to be PDE4D3 and PDE4D5 (32), respectively, three novel PDE4Dimmunoreactive species were detected in VSMC lysates following prolonged incubations with either forskolin or 8-Br-cAMP (Fig. 1A). One of the anti-PDE4D-immunoreactive proteins detected in treated VSMC migrated slightly more slowly than PDE4D3 and was shown previously to be the PKA-phosphoryl-  (10 M). At the end of the incubation, cells were washed with HBSS and lysed in ice-cold lysis buffer (without Triton X-100), and 1,000 ϫ g supernatant, 132,000 ϫ g supernatant, and 132,000 ϫ g particulate fractions were obtained by differential centrifugation as described under "Experimental Procedures." Total cAMP PDE activity and PDE4 activity of each fraction were determined as described under "Experimental Procedures," using Ro 20 -1724 (    cAMP-PKA and PKC-Raf-MEK-ERK Cascades Regulate PDE4D Expression ated form of PDE4D3 (32). Based on several factors, we identify the 74 Ϯ 3-and 64 Ϯ 2-kDa anti-PDE4D immunoreactive proteins detected in treated VSMC as PDE4D1 and PDE4D2, respectively. Thus, consistent with the cAMP-dependent expression of PDE4D1 and PDE4D2 previously described by others (35)(36)(37)(38)(39)(40)(41) in different cell types, these proteins were only detected in VSMC homogenates following prolonged (t min Ն1 h) incubation with cAMP-elevating agents, and their appearance was completely blocked by inclusion of either actinomycin D or cycloheximide (Fig. 1B). A role for PKA in this effect was predicted based on the ability of a selective PKA inhibitor (H89, 10 M) to abolish the increases (Fig. 1C). In contrast to the marked changes in PDE4D1 and PDE4D2 observed following prolonged cAMP elevation, no change in the abundance of PDE4D3 or PDE4D5 was detected in treated cells (Fig. 1). Also consistent with our identification of the two novel anti-PDE4D immunoreactive proteins as PDE4D1 and PDE4D2, mRNA for these two PDE4D variants were detected when RT-PCR was conducted using RNA isolated from treated cells (Fig. 1D). Thus, although primer sets designed for the selective amplification of either PDE4D1, PDE4D2, or PDE4D3 allowed amplification of mRNA for PDE4D3 and PDE4D2 when RNA isolated from control cells was used, mRNAs for each PDE4D1, PDE4D2, and PDE4D3 variants were amplified from RNA isolated from treated rat aortic VSMC (Fig. 1D). Consistent with a more pronounced effect of cAMP on the expression of PDE4D1 and PDE4D2, relatively little change in PDE4D3 mRNA levels was seen following incubation with forskolin ( Fig.  1D). As controls, no change in the abundance of two internal standards, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Fig. 1D) or ␤-actin (not shown), were noted between control and treated VSMC at any time point and in any experiment.
PMA or AngII Selectively Inhibits the cAMP-dependent Increase in PDE4 Activity-We reported previously that incubation of rat aortic VSMC for short periods (5-30 min) with agents that activate PKC results in a PKC-Raf-MEK-ERKcascade-mediated and phosphorylation-dependent increase in particulate PDE4D3 activity (32). In the work described here, we report that incubation of these cells with these agents for longer periods (Ն2 h) results in a decrease in PDE4 activity and that this effect was much more marked in the presence of cAMP-elevating agents (Table III). Whereas incubation of VSMC with the inactive analogue of PMA, 4␣-phorbol 12,13didecanoate (1-100 nM), had no effect on either basal or forskolin-induced increases in VSMC cAMP PDE activities (not shown), addition of PMA caused a slight decrease in PDE4 activity and markedly attenuated the increase in cAMP PDE activity caused by cAMP-elevating agents. In the three experiments in which this effect was monitored, PMA caused a modest time-dependent decrease in basal cAMP PDE activity (10 Ϯ 3% at 16 h) and a marked attenuation of the forskolininduced increase in cAMP PDE activity (65 Ϯ 7% at 8 -16 h or 90 Ϯ 12% at 4 h). By using selective PDE3 (cilostamide) or PDE4 (Ro 20-1724) inhibitors, the effect of PMA or of AngII, which also activate PKC in these cells, was shown to be selective for PDE4 activity (Table III). Indeed, addition of PMA had no effect on the forskolin-or 8-Br-cAMP-induced increases in PDE3 activity (Table III) nor on the small increase in PDE5 (Me 2 SO) or forskolin (10 M) for 4 h. RT-PCR reactions using primers selective for PDE4D3, PDE4D2, PDE4D1, or GAPDH were carried out as described under "Experimental Procedures." Products of these reactions were separated on 1% agarose gels, stained with ethidium bromide, and visualized under UV light, excised, and counted by liquid scintillation. Anti-PDE4D immunoreactive species were visualized as in A. C, confluent rat aortic VSMC were incubated with forskolin (10 M) or forskolin (10 M) and H89 (10 M) for 2 h. Anti-PDE4D immunoreactive species were visualized as in A. D, confluent rat aortic VSMC were incubated with fresh culture medium supplemented with vehicle activity that occurs in response to incubation of VSMC with cAMP-elevating agents (not shown). The reduced effect of PMA at longer times is consistent with the increased importance of changes in PDE3 activity at these time points (18,19).
PMA or AngII Inhibits the cAMP-mediated Induction of PDE4D1 and PDE4D2 Expression-Since our data showed that a cAMP-dependent induction of PDE4D1 and PDE4D2 contributed to the forskolin-mediated increased PDE4 activity in VSMC and that PMA or AngII could blunt these increases, we determined the effects of PMA or of AngII on PDE4D1 and PDE4D2 expression in these cells. Thus, either PMA or AngII added simultaneously with cAMP-elevating agents markedly inhibited the cAMP-induced increase in PDE4D1 and PDE4D2 expression ( Fig. 2A). Since the phosphorylation-mediated short term activation of PDE4D3 caused by PMA or AngII was shown previously to involve the PKC-Raf-1-MEK-ERK cascade (32), experiments were undertaken to address the involvement of this pathway in the inhibition of PDE4D1 and PDE4D2 expression caused by these agents. Results consistent with a role for the PKC-Raf-MEK-ERK cascade in mediating the effects of PMA or AngII on rat aortic VSMC PDE4 expression were obtained using selective inhibitors. Since results of preliminary experiments with some of the agents used (Bis-1 and PD98059) indicated that their selectivity was reduced following prolonged incubations (Ն2 h) in cells (not shown), these experiments were carried out over the shortest incubation periods possible (Յ2 h). Addition of the PKC-selective inhibitor, Bis-1, completely inhibited the phosphorylation of PDE4D3 caused by either PMA or AngII (32) and reversed the inhibitory effect of these agents on the forskolin-induced increase in PDE4D1 and PDE4D2 expression (Fig. 2B). A role for MEK-ERK involvement was addressed using the MEK-specific inhibitor PD98059. Addition of PD98059 completely inhibited the effects of PMA and of AngII on PDE4D3 phosphorylation (Fig. 2C and Ref. 32) and reversed the inhibitory effects of PMA (Fig. 2C) or AngII (not shown) on PDE4D1 and PDE4D2 expression. Since in certain cells PMA had been shown to bring about cellular effects by ultimately increasing intracellular Ca 2ϩ concentrations, the effects of a Ca 2ϩ ionophore (ionomycin) were also tested in some experiments. In our experiments ionomycin (1-100 nM) had no effect on basal PDE4 activity nor on the increase in this activity caused by cAMP-elevating agents (not shown).

Effect of PMA on PDE4 Expression Alters cAMP Signaling in
VSMC-Since incubation of VSMC with PMA or AngII reduced basal PDE4 activity and since PDE4 activity is dominant in these cells, we hypothesized that activators of adenylyl cyclase should more effectively increase cAMP in PMA-pretreated cells. Consistent with this, incubation of PMA-pretreated cells with a ␤-adrenoceptor agonist, isoproterenol, resulted in a more marked increase in cAMP than was caused by this agent in untreated control cells (Fig. 3A). Indeed, in the three experiments in which this effect was assessed, PMA potentiated the increase in cAMP caused by isoproterenol by approximately 70%. In addition, since we had previously reported that a marked desensitization to isoproterenol accompanies a pro-  cAMP-PKA and PKC-Raf-MEK-ERK Cascades Regulate PDE4D Expression longed incubation of VSMC with forskolin or 8-Br-cAMP and that isobutylmethylxanthine reversed much of this desensitization (18,19), we proposed that the combined treatment of cells with forskolin and PMA should reduce the level of desensitization achieved, compared with that caused by forskolin alone. Again, our data were consistent. Thus, whereas VSMC treated with forskolin for 16 h were virtually completely desensitized to the cAMP-elevating effects of isoproterenol, co-incubation of cells with PMA and forskolin reversed this desensitization (Fig. 3A). Indeed, in the three experiments in which this was monitored, PMA restored the effectiveness of isoproterenol by 50 -75%, with a reduced effect at higher concentrations of isoproterenol. The role of PDE4 activity in this effect of PMA pretreatment was addressed using the PDE4-selective inhibi-tor Ro 20-1724 and the PDE3/PDE4 dual selectivity inhibitor, zardaverine. Thus, co-treatment of VSMC with forskolin and PMA virtually normalized the responses of cells to isoproterenol when compared with forskolin alone. Again, consistent with our proposed mechanism, PMA-treated cells responded better to isoproterenol, and addition of the PDE4 inhibitor to these cells did not result in as marked an increase in cAMP as was seen in untreated control cells. In all instances, zardaverine had effects virtually indistinguishable from isobutylmethylxanthine, attesting to the dominance of PDE3s and PDE4s in regulating cAMP in these cells.
Inhibition of cAMP-induced Increases in PDE4D Expression by PMA or Ang II Is Mediated by a Mechanism Involving PDE4D mRNA Stability-Neither PMA nor AngII had any effect on forskolin-or isoproterenol-induced increases in cAMP in untreated control cells (not shown). However, each PMA and AngII inhibited the induction of PDE4D1 and PDE4D2 expression by a direct activator of adenylyl cyclases (forskolin) and a direct activator of PKA (8-Br-cAMP). Based on these findings, PMA or AngII could, in principle, have altered the cAMP-dependent accumulation of PDE4D1 and PDE4D2 by affecting the expression or the stability of either the mRNA or the protein for these PDE4D variants. In an attempt to assess which of these potential mechanisms was at play in VSMC, we initially examined the influence of incubation of rat aortic VSMC with PMA on the rate of accumulation and clearance of PDE4D proteins in these cells. Our data from these studies was consistent with an effect of PMA on the rate of accumulation of PDE4D1 and PDE4D2 and inconsistent with an effect on the rate of clearance of these proteins from cells. Thus, addition of PMA to cultures in which expression of PDE4D1 and PDE4D2 had been induced by prior treatment with forskolin, followed by removal of the adenylyl cyclase activator, caused an accumulation of phosphorylated PDE4D3 (32) but had no effect on the amount of time required for PDE4D1 or PDE4D2 to return to basal levels (Fig. 4A). Indeed, as determined by densitometric analysis, the amount of time required for PDE4D1 to decrease below immunodetectable levels was about 2 h, whether or not PMA had been added. Although the amount of time required for PDE4D2 to return to control values was significantly longer (approximately 5 h), again PMA had no effect (Fig. 4A). The molecular basis for the different rates of PDE4D1 and PDE4D2 clearance in VSMC is not known. In contrast to the lack of effect of PMA on the rate of clearance of PDE4D1 and PDE4D2, PMA markedly prolonged the amount of time required for PDE4D1 and PDE4D2 to accumulate in forskolin-treated cells (Fig. 4B). Thus, while PDE4D1 was readily detected after 90 min of incubation with 10 M forskolin, simultaneous addition of PMA delayed the accumulation of PDE4D1 such that this protein was only detected after 150 min of treatment (Fig. 4B). Moreover, the level of PDE4D1 present in cells incubated with both forskolin and PMA was consistently lower than that obtained in cells incubated with forskolin alone (Fig. 4B). Data consistent with an effect of PMA on expression of PDE4D gene products were also reflected in the effects of this agent on the cAMP-induced changes in PDE4D mRNA levels. Thus, addition of PMA to cells inhibited the marked increases in PDE4D1 and PDE4D2 mRNA as well as the very modest increase in PDE4D3 mRNA (Fig. 5, A and B). Indeed, in the five experiments in which this was measured, PMA inhibited the forskolin-induced increase in PDE4D1, PDE4D2, and PDE4D3 mRNA by 75 Ϯ 10, 79 Ϯ 12, and 85 Ϯ 19%, respectively (Fig.  5A). Since PMA or AngII inhibited the cAMP-dependent, PKAmediated increase in PDE4 activity in rat aortic VSMC but had no effect on the cAMP-dependent, PKA-mediated increase in PDE3 activity in these cells (Table III), we hypothesized that cAMP-PKA and PKC-Raf-MEK-ERK Cascades Regulate PDE4D Expression these agents were acting by altering the stability of the PDE4D mRNA rather than through a transcriptionally based mechanism. Based on this we reasoned that the t1 ⁄2 of PDE4D mRNA should be shorter when determined in cells incubated with both forskolin and PMA, compared with that obtained in cells incubated with forskolin alone. By using an experimental approach based on measuring the decay of mRNA following inhibition of transcription with actinomycin D, our data are consistent with a PMA-mediated destabilizing effect on PDE4D mRNA (Fig. 6). Thus, in the four experiments in which this was measured, addition of 100 nM PMA reduced the t1 ⁄2 of PDE4D1 mRNA by about 60% (79 Ϯ 3 to 34 Ϯ 3 min) (Fig. 6, A and B). Similarly, in three separate experiments the t1 ⁄2 of mRNA encoding PDE4D2 and PDE4D3 was shortened by approximately 50% by addition of PMA to forskolin-treated cells. Thus, the t1 ⁄2 of PDE4D2 was reduced from 96.3 Ϯ 7 to 51 Ϯ 4 min following addition of PMA to the cells, while this agent reduced the t1 ⁄2 of PDE4D3 mRNA 73 Ϯ 10 to 39 Ϯ 9 min. PMA did not alter the rate of decay of GAPDH mRNA in this study (12% decay in 4 h).

DISCUSSION
In a previous report, we identified PDE4D as the major PDE4 variant expressed in cultured rat aortic VSMC, and we presented evidence for the expression of both PDE4D3 and PDE4D5 variants in these cells (32). In addition, we reported that incubation of rat aortic VSMC with cAMP-elevating agents, or with structural analogues of cAMP, caused a rapid PKA-mediated, phosphorylation-dependent activation of both these proteins. In this previous work (32) we also defined a novel role for the PKC-Raf-MEK-ERK cascade in the regulation of PDE4D activity whereby activation of this cascade led to a phosphorylation-dependent activation of particulate PDE4D3 that was additive with the effects of PKA. Interestingly, co-stimulation of both the cAMP-PKA and PKC-Raf-MEK-ERK cascades in VSMC led to the translocation of particulate PDE4D3 to the cytosol of these cells. Clearly, this form of regulation could potentially allow for the rapid regulation of PDE4D activity at selected regions within cells and impact on cellular functions requiring local PDE4D-mediated control of cAMP levels. In the studies detailed here, we have investigated the effect of prolonged elevations in cAMP on PDE4D expression in VSMC, and we have characterized the impact of activation of the PKC-Raf-MEK-ERK cascade on this cAMP-mediated regulation of PDE4D expression. These studies demonstrate that the regulation of PDE4D expression in cells is controlled by both these signaling cascades and that the coordinated regulation of PDE4D expression could significantly impact cAMP-mediated signaling.
Consistent with our previous work (18,19,21,32), and previous reports by others (35)(36)(37)(38)(39)(40)(41)(42) using non-vascular cell types, incubation of rat aortic VSMC with either forskolin or 8-Br-cAMP caused a time-and concentration-dependent increase in PDE4 activity. Whereas at lower levels of stimulation the increases in PDE4 were relatively modest, as much as 4-fold increases were achieved when cells were incubated for 16 h with 100 M forskolin. The effects of the PKA inhibitor H89, as well as those of actinomycin D or cycloheximide, demonstrated that the forskolin-or 8-Br-cAMP-mediated increases in PDE4 activity occurred as a result of activation of PKA and that new mRNA and protein synthesis were both necessary. At the end of these incubations, cells were lysed in ice-cold lysis buffer, and lysates were centrifuged at 1,000 ϫ g for 10 min at 4°C. The supernatant fractions were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and analyzed by immunoblot with a PDE4D-specific monoclonal antibody (61D10E). Anti-PDE4D immunoreactive species were visualized by ECL as per the supplier's recommendation using a horseradish peroxidase-conjugated goat anti-mouse IgG as described under "Experimental Procedures." Immunoblots were quantitated by scanning densitometry using Corel Photo-Paint 8.0, and the immunoreactive bands at 0 min were designated a relative score of 1.0. B, VSMC were incubated with fresh culture medium supplemented with forskolin (10 M) or forskolin (10 M) and PMA (100 nM) for the indicated period. At the end of the incubation, cells were processed and anti-PDE4D-immunoreactive proteins were visualized as described in A. Whereas up-regulation of PDE4 activity played a more important role in the cytosolic fraction, accounting for virtually all of the increase in total cAMP PDE activity, the relative role of increased PDE4 activity in the particulate fraction was less dominant, accounting for less than 50% of this increase. In recent work conducted in our laboratory (19), a role for increased expression of particulate PDE3B in rat aortic VSMC incubated with 8-Br-cAMP in the residual effects in the particulate fraction was elucidated.
By using both molecular biological (RT-PCR) as well as immunological methodologies (immunoblotting), our studies identify PDE4D1 and PDE4D2 as the major PDE4D variants increased following prolonged periods of incubation with cAMPelevating agents. Thus, low levels of mRNA encoding PDE4D2 and PDE4D3 could be amplified from untreated control cells, and proteins previously identified as PDE4D3 and PDE4D5 (32) were detected in homogenates of untreated VSMC using a PDE4D-specific antisera. Consistent with a cAMP-induced increase in VSMC expression of PDE4D1 and PDE4D2, a marked increase in the amount of PDE4D1 and PDE4D2 mRNA which could be amplified from treated cells was observed. In addition, whereas untreated control cells expressed PDE4D3 and PDE4D5 (32), immunoblots of treated cells allowed the detection of four new anti-PDE4D immunoreactive proteins. Whereas two of these were phosphorylated PDE4D3 and PDE4D5 (32), the two other PDE4D immunoreactive proteins present in treated cells are represented as PDE4D1 and PDE4D2, the two smallest known PDE4D splice variants (7,8). Consistent with these assignments are our findings that these PDE4D-immunoreactive proteins, which migrate at molecular weights consistent with PDE4D1 and PDE4D2 (7,8), are pres-ent only in treated VSMC. Although in some reports PDE4D1 and PDE4D2 have been shown to co-migrate with identical molecular weights (40,42), well resolved doublets of these PDE4D variants were shown to appear in mononuclear cells following prolonged incubation of these cells with a lipophilic analogue of cAMP (36,37). Our data are consistent with previous reports of the cAMP induction of the short variants of the PDE4D gene (35)(36)(37)(38)(39)(40)(41)(42) and show that the longer variants, PDE4D3 and PDE4D5, are not subject to cAMP regulation in rat aortic VSMC. The lack of an effect of cAMP on PDE4D3 expression is consistent with modest effects of cAMP on the expression of this variant in other cell types (40,41,45,46).
A novel finding of our work is that addition of the phorbol ester, PMA, or of the vasoactive agent, AngII, markedly attenuated the cAMP induction of expression of PDE4D1 and PDE4D2 in rat aortic VSMC. Although in previous reports PMA was shown to cause a PKC-dependent, rapid and transient induction of PDE1C in some cells (47,48), there have been no previous reports identifying a role for this agent in the regulated expression of any PDE4 gene, in any cell type. In addition to identifying a role for PMA, or AngII, in the expression of PDE4D1 and PDE4D2, our studies also identify the PKC-Raf-MEK-ERK cascade as potentially playing a signaling role in this phenomenon. Thus, the selective PKC inhibitor, Bis-1, or the MEK inhibitor, PD98059, both reversed the inhibitory effect of PMA on the cAMP-induced increase in PDE4D1 expression. Although we are extrapolating this mechanism to account for the effects of PMA, or AngII, on PDE4D2 expression, the time course of the induction of this splice variant (Ͼ2 h) and the reduced selectivity of both Bis-1 and PD98059 after incubation for periods greater than 2 h precluded our direct measurement of this effect. From the data presented in this work and that previously published (32) concerning the role of PMA, via the PKC-Raf-MEK-ERK cascade, on PDE4D3 activity and targeting, a paradigm emerges that involves both the cAMP-PKA and PKC-Raf-MEK-ERK signaling cascades in the regulation of PDE4D activity, targeting, as well as expression in cells. Some of the data presented in this report speak to the potential importance of these findings. First, our data demonstrate that following a prolonged treatment of cells with PMA, the isoproterenol-mediated increase in cAMP was potentiated significantly. A role of the PMA-induced reduction in PDE4 activity in this effect is demonstrated by the normalization of this effect with the PDE4-selective inhibitor Ro 20-1724. Second, since PMA inhibited the cAMP-mediated induction of PDE4D1/PDE4D2 expression, it partially reversed the cAMPdependent desensitization of these cells to further addition of adenylyl cyclase activating agents, such as isoproterenol. Since zardaverine normalized the effects of isoproterenol in these experiments, the residual desensitization not reversed by PMA was most likely due to PDE3 (18 -19). Although this paradigm would suggest that the activity of PDE4D variants expressed in cells, as well as the level of expression of PDE4D gene products in cells, could be subject to continued regulation by agents acting through cAMP-PKA and the PKC-Raf-MEK-ERK signaling systems, further work will be required to assess the relative impacts of these regulatory systems in vivo. In this context, given that the vasoactive hormone AngII mimicked all of the effects of PMA in vitro, we have recently begun to investigate the impact of prolonged treatments of rats with enalapril, an inhibitor of angiotensin-converting enzyme, or losartan, an AT 1 -receptor antagonist, on PDE4 activity in blood vessel smooth muscle cells. Our preliminary data from these studies indicate that inhibition of AngII signaling in vivo has FIG. 6. PMA reduces the stability of PDE4D1 mRNA in rat aortic VSMC incubated with forskolin. Confluent rat aortic VSMC were incubated with fresh culture medium supplemented with forskolin (10 M) or forskolin (10 M) and PMA (100 nM) for 4 h. Following these incubations, actinomycin D (4 M) was added and RNA isolated at 10-min intervals for 120 min. RT-PCRs used for PDE4D1-or GAPDHspecific primers were carried out as described under "Experimental Procedures" in the presence of 2 Ci of [␣-32 P]dCTP. PCR products for PDE4D1 or GAPDH were separated on 1% agarose gels, excised, and quantitated by liquid scintillation. A, a representative RT-PCR with RNA isolated following 50 min in the presence of actinomycin (4 M). B, semi-logarithmic plot of PDE4D1 mRNA decay, corrected for GAPDH, following inhibition of transcription with actinomycin D (4 M). PDE4D1 half-lives (t1 ⁄2 ) were calculated by linear regression of ln(M t /M 0 ) versus time, where M t is the PDE4D1 mRNA level at the time (t) and M 0 is the amount at time 0. marked effects on PDE4 activity and expression. 2 In addition to identifying that the PKC-Raf-MEK-ERK cascade may play a role in mediating the effects of PMA-or AngII-mediated inhibition of cAMP-induced increases in PDE4D gene expression, another novel finding of our work was the identification of a role for destabilization of PDE4D mRNA as a potential molecular basis for this effect. Although this effect could, in principal, have been related to alteration of the synthesis or the stability of either PDE4D mRNA or PDE4D proteins, our data indicate that a mechanism involving a destabilizing of PDE4D mRNA was most likely. Thus, although addition of PMA to cells in which PDE4D1 and PDE4D2 had been induced by a prior incubation with forskolin did not affect the rate at which these proteins were cleared from cells, a co-incubation of both agents markedly reduced the rate at which these proteins accumulated when compared with the effects of forskolin alone. These data are inconsistent with a role for increased clearance of PDE4D proteins in these cells but rather are consistent with a role for an effects mediated by a decreased rate of synthesis. Although a role for PMA in reducing the rate of transcription of the PDE4D gene was theoretically possible, our data and previously reported observations (44) were inconsistent with such a mechanism and, rather, indicated that a mechanism involving altered stability of PDE4D mRNA was more likely. Thus, although PMA reduced the cAMP-mediated increase in PDE4D expression in VSMC, addition of this agent had absolutely no effect on the cAMP-induced expression of PDE3B in these cells, a process known to also be regulated transcriptionally by cAMP in cells (6,18,19). Also, the observation that PMA reduced the level of PDE4D3 mRNA combined with the fact that expression of PDE4D3 is dependent on the activity of a different promoter than that which controls the expression of PDE4D1 and PDE4D2 is not consistent with a transcriptional mechanism. In this context, in previously reported studies in which the intronic PDE4D1/PDE4D2 promoter was used to drive expression of the luciferase gene, rather than suppress the activity of this promoter, PMA was shown to increase expression of luciferase and to potentiate the activating effects of cAMP-elevating agents (44). Experiments aimed at determining if addition of PMA to VSMC affected the stability of PDE4D mRNA were consistent with this mechanism being responsible for the effects of PMA on PDE4D1 and PDE4D2 expression. Thus, whereas incubation of VSMC with forskolin caused an increase in PDE4D1 and PDE4D2 mRNA, the t1 ⁄2 of this mRNA was significantly shortened when PMA had been included in the incubation. Indeed, addition of PMA reduced the t1 ⁄2 of each PDE4D1, PDE4D2, and PDE4D3 mRNA by approximately 50 -60%. Although these data are consistent with PMA and, by extension, AngII, in reducing cAMP-mediated expression of PDE4D by destabilizing PDE4D mRNA, the molecular basis of this effect will require further work. Since preliminary studies have indicated that the inhibitory effect of PMA on PDE4D mRNA was cycloheximide-sensitive, it will be necessary, in future experiments, to assess further the nature of this effect and characterize the protein(s) involved.
Taken together, our previous work (42) detailing a PKC-Raf-MEK-ERK cascade-mediated effect on PDE4D3 activity and targeting, combined with the observations reported in this paper strongly imply that PDE4D activity and expression are both subject to multiple and overlapping mechanisms of regulation in rat aortic VSMC and potentially other cells. Given that several vasoactive agents act either through cAMP or can activate the PKC-Raf-MEK-ERK cascades, it is likely that the events detailed in our studies will be relevant to the in vivo regulation of VSMC PDE4D, a hypothesis that we are currently testing. Although our work was limited to PDE4D, whether similar effects may occur for the three other PDE4 genes should be tested. Since the PDE4A gene also can encode short variants, the expression of which are sensitive to cAMP (7,8), a similar effect on expression of PDE4A may be anticipated.