Adenylyl cyclase 3 mediates prostaglandin E(2)-induced growth inhibition in arterial smooth muscle cells.

Arterial smooth muscle cell (SMC) proliferation contributes to a number of vascular pathologies. Prostaglandin E(2) (PGE(2)), produced by the endothelium and by SMCs themselves, acts as a potent SMC growth inhibitor. The growth-inhibitory effects of PGE(2) are mediated through activation of G-protein-coupled membrane receptors, activation of adenylyl cyclases (ACs), formation of cAMP, and subsequent inhibition of mitogenic signal transduction pathways in SMCs. Of the 10 different mammalian AC isoforms known today, seven isoforms (AC2-7 and AC9) are expressed in SMCs from various species. We show that, despite the presence of several different AC isoforms, the principal AC isoform activated by PGE(2) in human arterial SMCs is a calmodulin kinase II-inhibited AC with characteristics similar to those of AC3. AC3 is expressed in isolated human arterial SMCs and in intact aorta. We further show that arterial SMCs isolated from AC3-deficient mice are resistant to PGE(2)-induced growth inhibition. In summary, AC3 is the principal AC isoform activated by PGE(2) in arterial SMCs, and AC3 mediates the growth-inhibitory effects of PGE(2). Because AC3 activity is inhibited by intracellular calcium through calmodulin kinase II, AC3 may serve as an important integrator of growth-inhibitory signals that stimulate cAMP formation and growth factors that increase intracellular calcium.

vation of cyclin-dependent kinases (3). Synthesis of cAMP from ATP is catalyzed by adenylyl cyclases (ACs), which are, with some exception, transmembrane enzymes activated by receptors coupled to the stimulatory G-protein G s (4 -7). To date, at least 10 different isoforms of ACs (AC1 through AC10) have been cloned and identified in a wide array of vertebrate tissues. Most tissues express several AC isoforms, which exhibit remarkable diversities in their sensitivities toward signaling molecules such as different subunits of G-proteins, calcium/ calmodulin, protein kinases, and phosphatases (4 -7). Although stimulation through the ␣ subunit of G s is the principal mechanism whereby ACs are activated (4 -7), the activity of certain AC isoforms is also regulated by ␣ subunits of G i , G z , and G o (8), G-protein ␤␥ subunits (9), cAMP-dependent protein kinase phosphorylation (10), protein kinase C isoforms (11)(12)(13), changes in membrane potential (14), and calcium (15)(16)(17). Calcium regulates several AC isoforms directly or indirectly though other proteins. Increases in calcium though IP3 receptors can lead to protein kinase C activation, which in turn can activate AC1, AC2, AC3, AC5, and AC7 (7). Calcium binding to calmodulin can directly stimulate AC1 and AC8 (15)(16)(17), can activate phosphatase 2B-sensitive AC9 (18), and can inhibit AC1 (19) and AC3 (20,21) via activation of calmodulin kinase IV (CaM KIV) and CaM KII phosphorylation, respectively.
The expression of AC isoforms in mammalian tissues is also diverse. Some isoforms exhibit extremely broad patterns of expression, such as AC2 and AC9, whereas expression of other isoforms appears to be tissue-specific, such as the neurospecific expression of AC1 (22). Although AC3 was originally thought to be expressed only by the olfactory neuroepithelium, it is now known to be expressed in multiple tissues (23). Clearly, the expression of multiple AC isoforms in a cell provides an intricate system for cross-talk and fine tuning of signals increasing cAMP formation.
We show here that normal human and murine arterial SMCs express AC3 and that the principal AC isoform activated by PGE 2 in these cells is a calcium-inhibited AC with pharmacological characteristics of AC3. Furthermore, arterial SMCs isolated from AC3-deficient mice are resistant to PGE 2 -mediated growth inhibition. Thus, AC3 mediates the growth-inhibitory effects of PGE 2 in arterial SMCs.

Reagents-PGE 2 ,
8-bromo-cAMP, 3-isobutyl-1-methylxanthine (IBMX), and forskolin were obtained from Biomol (Plymouth Meeting, PA) and were dissolved in ethanol, distilled water, and Me 2 SO, respectively. The calcium ionophore A23187 was from Calbiochem-Novabiochem Corp. Human recombinant platelet-derived growth factor-BB (PDGF-BB) and a polyclonal anti-G s ␣ antibody generated against the peptide RMHLRQYELL of bovine G s ␣ were from Upstate Biotechnology (Lake Placid, NY). Anti-AC3 antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA), and a monoclonal anti-CaM KII antibody was from Transduction Laboratories (Lexington, KY). [2, H]Adenine (20 -40 Ci/mmol) was obtained from PerkinElmer Life Sciences. High glucose DMEM and calcium-free DMEM were from Life Technologies, Inc.
AC3-deficient Mice-The AC3 gene has recently been disrupted in mice in the laboratory of D. R. Storm (24). A colony of AC3ϩ/Ϫ mice on a C57/BL6 ϫ 129 (50:50) background was maintained and used to generate age-and sex-matched wild-type (AC3ϩ/ϩ), AC3ϩ/Ϫ, and AC3Ϫ/Ϫ mice for this study. Disruption of the AC3 locus in AC3ϩ/Ϫ and AC3Ϫ/Ϫ mice and SMCs was confirmed by polymerase chain reaction analysis of tail biopsies or cultured arterial SMCs. The two primers (5Ј-CTGGTGAAGTGGCTTGACCT-3Ј) and (5Ј-GTTATGAAGA-AGGAGAAGACA-3Ј) that hybridize to sequences within the deleted region of the AC3 locus were used to identify the presence of the wildtype allele. The mutant allele was revealed by a forward primer (5Ј-C-CTGTGCTCTAGTAGCTTTACGG-3Ј) that hybridizes to the reverse complement of the 5Ј region of the neomycin cassette and a second primer (5Ј-CTGTGAAGTAGGTTCCTACCTG-3Ј) that hybridizes 230 base pairs downstream of the forward primer as described by Wong et al. (24).
Tissues and Cells-Human fetal aortas were obtained from the Central Laboratory for Human Embryology at the University of Washington. Eleven aortas with a gestational age of 74 -145 days (mean value of 102 days) were obtained. A segment of each aorta was fixed in methyl Carnoy's fixative for immunohistochemical detection of AC3.
Human newborn (2-day-old to 3-month-old) thoracic aortas were obtained from infants following accidental death, death from sudden infant death syndrome, or death from congenital defects. Human SMCs were isolated by the explant method and cultured and characterized as described previously (25). All experiments were performed in DMEM with 1% human plasma-derived serum (PDS).
Mouse SMCs were isolated from the thoracic aorta of wild-type (AC3ϩ/ϩ), AC3ϩ/Ϫ, and AC3Ϫ/Ϫ mice. The mice were killed by carbon dioxide, and the thoracic aorta was immediately dissected and cleaned from extraneous tissue, blood, and fat. The aorta was then transferred to a 35-mm dish with 4 ml of enzyme solution (2 mg/ml bovine serum albumin, 1 mg/ml collagenase type 1 (Worthington Biochemical Corp., Lakewood, NJ), 0.375 mg/ml soybean trypsin inhibitor (Worthington Biochemical Corp.), and 0.125 mg/ml elastase type III (Sigma) in DMEM for 30 min at 37°C. Incubation of the aorta in this medium allows separation of the smooth muscle layer and the adventitia and removes endothelial cells. The adventitia was carefully separated from the smooth muscle layer using watchmaker forceps, and the smooth muscle layer was then minced and incubated in the above medium for 2 h at 37°C on a shaker. After incubation, the cells were centrifuged, washed with DMEM, 10% fetal bovine serum, and plated onto 25-cm 2 tissue culture flasks. Cells were characterized as SMCs by morphologic criteria and by expression of smooth muscle ␣-actin. One animal was used to generate one SMC strain, and the SMCs were used between passages 2 and 6.
Measurements of AC Activity-AC activity was measured in intact cells as described previously (20). In short, human and mouse SMCs in six-well plates (150,000 cells/well) were incubated in DMEM and 1% human PDS for 2 days. The cells were labeled with [2, H]adenine (2 Ci/ml) for 18 h and then preincubated with 1 mM IBMX for 30 min to inhibit cyclic nucleotide phosphodiesterases (PDEs) that hydrolyze cAMP. Following stimulation of the cells with PGE 2 , forskolin, A23187, and KN-62 for the indicated periods of time, cellular proteins were precipitated with ice-cold 5% trichloroacetic acid containing 1 M cAMP. [ 3 H]cAMP was isolated by a sequential Dowex-alumina chromatography method (26). AC activity was calculated as the percentage of Analysis of AC3 by Immunoprecipitation and Western Blot-For detection of AC3 protein in human and mouse SMCs, the cells were plated onto 100-mm dishes and maintained until the cultures were nearly confluent. The SMCs were then incubated in DMEM containing 1% PDS for 2 days. For immunoprecipitation of AC3, SMCs were harvested in 1 ml of 4°C immunoprecipitation buffer (150 mM NaCl, 10 mM Tris, pH 7.4, 1% Nonidet P-40, 0.1% SDS, 0.5% sodium deoxycholate, 1 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, 1 mM dithiothreitol), sheared with a 23-gauge needle, and centrifuged at 12,000 ϫ g for 10 min at 4°C. Supernatants were subjected to immunoprecipitation with 2 g of anti-AC3 rabbit polyclonal antibody plus 20 l of protein A-agarose (Pierce) at 4°C overnight on a rotator. As a control, a 10ϫ excess (by weight) of antigen peptide (Santa Cruz Biotechnology) was included in the immunoprecipitation. Immunoprecipitates were washed three times with immunoprecipitation buffer, 10 l of 5ϫ SDS sample buffer was added, and samples were heated at 95°C for 5 min. Samples were subjected to SDS-polyacrylamide gel electrophoresis according to Laemmli (27) on 7.5% polyacrylamide gels, transferred to polyvinylidene difluoride membranes (Bio-Rad), and Western blotted with 0.1 g/ml anti-AC3 followed by horseradish peroxidase-conjugated goat anti-rabbit IgG (Life Technologies, Inc.). For Western blot analysis of AC3 in membrane fractions, the cells were cultured as described above and harvested by scraping in a buffer containing 50 mM HEPES (pH 7.4), 50 mM NaCl, 1 mM MgCl 2 , 2 mM EDTA, 10 mM pyrophosphate, 10 mM NaF, 500 M Na 3 VO 4 , 1 mM dithiothreitol, 1 mM benzamidine, 1 mM leupeptin, 1 mM pepstatin, and 1 mM aprotinin. The samples were sheared with a 23-gauge needle and centrifuged for 10 min at 10,000 ϫ g at 4°C in a microcentrifuge. The supernatant was then centrifuged for 60 min at 100,000 ϫ g at 4°C, which resulted in a pellet rich in plasma membranes and a soluble fraction. The pellets were resuspended in the buffer above with the addition of 1% Triton X-100. Protein concentrations in the pellet fractions were quantitated by either the BCA ® protein assay (Pierce) or the Bio-Rad protein assay according to Bradford (Bio-Rad). Western blots were developed by enhanced chemiluminescence (ECL) according to the manufacturer's instructions (Amersham Pharmacia Biotech). Enhanced chemiluminescence films were digitized using a scanner, and images were processed using Photoshop 5.0 and Powerpoint.
Immunohistochemistry-Expression of AC3 in human fetal aorta was studied by using immunohistochemistry. Tissues were fixed in methyl Carnoy's fixative, embedded in paraffin, and cut into 5-m sections. Tissue sections were deparaffinized, hydrated, and subjected to ABC immunohistochemistry (ABC Elite kit, Vector Laboratories, Burlingame, CA) using 1 g/ml anti-AC3. A 20-fold molar excess of peptide antigen was included as a negative control. The reaction was carried out using nickel-enhanced diaminobenzidine (Fast DAB, Sigma) and terminated with distilled water. Sections were counterstained with methyl green, dehydrated, cleared in xylene, and mounted with Permount (Fisher Scientific). Microscopy was carried out using a Sony DK5000 3CCD digital camera coupled to a Nikon Eclipse E800 microscope. Images were acquired via Photoshop. Omission of the primary antibody resulted in no staining, and the specific staining was abolished by preincubation with the peptide antigen.
DNA Synthesis and Proliferation of Cultured SMCs-For measurement of DNA synthesis, SMCs (50,000 cells/well) were plated in 24-well trays and grown in DMEM, 10% fetal bovine serum. The medium was changed to DMEM, 1% PDS for 2 days when the cell cultures were nearly confluent. The cells were then stimulated as indicated in the presence or absence of 1 nM PDGF-BB for 18 h and subsequently pulsed with [ 3 H]thymidine (1 Ci/ml) during an additional 2-h incubation as described previously (28). The radioactivity incorporated into DNA was normalized to the amount of cellular protein.
Cell proliferation was also measured by determining cell number. Cells (30,000 or 50,000 cells/well) were incubated in the presence of the indicated agents for 4 or 6 days. The cells were trypsinized, fixed in Holley's fixative (3.7% formaldehyde, 86 mM NaCl, 106 mM Na 2 SO 4 ), and counted using a cell counter (Coulter Corp., Hialeah, FL).

PGE 2 Activates AC3 in Human
SMCs-To investigate the characteristics of the AC isoform(s) activated by PGE 2 in SMCs, normal human aortic SMCs were stimulated with 2 M PGE 2 . Forskolin, which activates all known AC isoforms except AC9, was used as a control. PGE 2 -induced stimulation of AC activity was observed at 1 min (80% increase over basal) and was maximal at 30 min (data not shown). Therefore, subsequent experiments were performed using a 30-min stimulation with PGE 2 . As shown in Fig. 1A, PGE 2 induced an approximate 20-fold increase in AC activity in human SMCs. This stimulation was blocked by 65% by an increase in intracellular calcium levels induced by the calcium ionophore A23187. The results shown in Fig. 1B demonstrate that A23187 did not inhibit PGE 2 -induced AC activation in the absence of extracellular calcium. An extracellular calcium concentration of 0.5-10 mM was required for the inhibitory effect of A23187 (Fig. 1B). Thus, the effect of A23187 was dependent on increases in intracellular calcium levels. The inhibition of PGE 2 -induced AC activation by the calcium ionophore was completely reversed by co-incubation with the CaM KII inhibitor KN-62 (Fig. 1A). KN-62 alone had no effect on AC activity (data not shown). AC3 is the only AC isoform known to be inhibited by CaM KII, indicating that the main isoform activated by PGE 2 in human SMCs is AC3. Expression of CaM KII in human SMCs was verified using Western blot analysis (data not shown). When four different experiments were summarized, the A23187-and CaM KII-sensitive AC3 component of PGE 2 -stimulated AC activation was 62.4 Ϯ 3.8% (data not shown). Forskolin (50 M), on the other hand, induced an approximate 80-fold stimulation of AC activity in these cells, and this effect was not significantly inhibited by increasing intracellular calcium levels (Fig. 1A). This finding is consistent with results that show that several AC isoforms that are not inhibited by calcium are expressed in human SMCs. Together, the results show that AC3 is the principal AC isoform activated by PGE 2 in human arterial SMCs.
Human and Murine Arterial SMCs Express AC3-To investigate whether AC3 is indeed expressed in arterial SMCs, we used Western blot and immunohistochemical analyses. To verify that AC3 is expressed in human arterial SMCs in vivo, we used human fetal aortas for immunohistochemical detection of AC3 (Fig. 2, A and B). These studies demonstrated a clear expression of AC3 in smooth muscle of fetal human aorta sections at gestational days 74 and 84 (data not shown) with much more labeling demonstrated at day 127 ( Fig. 2A). Labeling was completely blocked by including AC3 peptide antigen in the reaction (Fig. 2B). AC3 was also expressed in cultured human arterial SMCs as shown by immunoprecipitation and subsequent Western blot analysis. The AC3 antibody was found to precipitate a band of ϳ170 kDa, corresponding to the glycosylated form of AC3 (Fig. 2C). This band was efficiently blocked by the AC3 antigen used to generate the antibody (Fig.  2C). Thus, AC3 is expressed in isolated human arterial SMCs in culture and in vivo.
Next we took advantage of an AC3-deficient mouse that was recently developed in the laboratory of D. R. Storm (24) to investigate the role of AC3 in PGE 2 -induced signaling. SMCs were isolated from the thoracic aortas of wild-type (AC3ϩ/ϩ), heterozygous mouse (AC3ϩ/Ϫ), and AC3 knockout (AC3Ϫ/Ϫ) mice and were characterized as SMCs by the expression of smooth muscle ␣-actin. Expression of AC3 in membrane fractions from these cells was studied by Western blot analysis. SMCs from wild-type mice showed two prominent bands at ϳ170 and 180 kDa corresponding to glycosylated forms of AC3 (Fig. 2D). These bands were present in the membrane fraction and were absent from the soluble fraction (data not shown). SMCs from AC3Ϫ/Ϫ mice, on the other hand, were devoid of AC3 expression (Fig. 2D). AC3ϩ/Ϫ mice showed an ϳ50% reduction of AC3 protein expression (24). To investigate if there is compensatory up-regulation of expression of G s ␣ in SMCs from AC3Ϫ/Ϫ mice, Western blot analyses were performed. These studies showed that the expression of G s ␣ was similar in SMCs from wild-type and AC3Ϫ/Ϫ mice (data not shown). PGE 2 Preferentially Activates AC3 in Murine Arterial SMCs-As shown in Fig. 3A, PGE 2 induced a 5-fold stimulation of AC activity in murine aortic SMCs. This activation was completely inhibited in the presence of A23187 and was nearly normalized by the CaM KII inhibitor KN-62, indicating that in murine aortic SMCs, like in human aortic SMCs, AC3 is the principal AC isoform activated by PGE 2 . Forskolin (10 M), on the other hand, induced a stimulation of AC activity about 50-fold higher than that of PGE 2 (data not shown), indicating that forskolin-sensitive AC isoforms other than AC3 were expressed by these cells. SMCs were also isolated from AC3ϩ/Ϫ C, immunoprecipitation and Western blot analysis of AC3 expression in cultured human SMCs. Cell lysates were subjected to immunoprecipitation and immunoblotted using an anti-AC3 antibody. Lane 1, immunoprecipitation in the presence of 2 g/ml of anti-AC3; lane 2, immunoprecipitation in the presence of 2 g/ml anti-AC3 plus a 10ϫ excess (by weight) of antigen peptide. The arrow shows AC3 running at ϳ170 kDa. The numbers at the left indicate molecular weight standards. D, Western blot analysis of AC3 expression in cultured murine SMCs. Membrane fractions were prepared from aortic SMCs isolated from wild-type (AC3ϩ/ϩ) and AC3Ϫ/Ϫ mice. Proteins (50 g/lane) were separated on 10% SDS gels, transferred to Immobilon membranes, and probed with the anti-AC3 antibody (1:1000 dilution). The arrows show two glycosylated forms of AC3 at 170 and 180 kDa in the membrane fraction from SMCs isolated from wild-type (AC3ϩ/ϩ) mice. AC3 expression was abolished in SMCs from AC3Ϫ/Ϫ mice. mice that retain ϳ50% of the AC3 expression level found in wild-type littermates (24). We used AC3ϩ/Ϫ animals rather than AC3Ϫ/Ϫ animals in several of our experiments because of the low number of adult AC3Ϫ/Ϫ mice available. Aortic SMCs from AC3ϩ/Ϫ mice showed only half of the PGE 2 -induced AC activation found in SMCs from wild-type mice (Fig. 3A). As in SMCs from wild-type mice, the PGE 2 -induced AC activity in SMCs from AC3ϩ/Ϫ mice was completely blocked by increasing intracellular calcium concentrations and normalized by co-incubation with the CaM KII inhibitor KN-62 (Fig. 3A).
We next investigated whether AC isoforms other than AC3 could be activated by high concentrations of PGE 2 . For these experiments, SMCs isolated from AC3ϩ/Ϫ mice and wild-type littermates were stimulated with concentrations of PGE 2 up to 50 M (Fig. 3B). The highest concentrations were well above receptor saturating concentrations. We showed that increasing concentrations of PGE 2 could not compensate for the 50% reduction in AC activation seen in SMCs from AC3ϩ/Ϫ mice, indicating that other AC isoforms did not efficiently couple to PGE 2 receptors (EP2 and/or EP4 receptors) even at supraphysiological concentrations of PGE 2 (Fig. 3B). Thus, AC3 is the principal AC isoform activated by PGE 2 in murine aortic SMCs.
AC3 Mediates PGE 2 -induced Inhibition of SMC Proliferation-The role of AC3 in PGE 2 -mediated growth inhibition of SMCs was investigated next. PGE 2 (10 M) induced an ϳ40% inhibition of basal DNA synthesis and proliferation measured as cell number in human SMCs and inhibited PDGF-BB-induced proliferation to a similar extent (Table I). Although the AC activity studies showed that AC3 played a major role in PGE 2 -induced AC activation, the role of AC3 in growth inhibition cannot be readily studied in human SMCs. Instead we used SMCs isolated from AC3ϩ/Ϫ mice, AC3Ϫ/Ϫ mice, and wild-type littermates to address the role of AC3 in SMC proliferation. The results in Fig. 4A show that 10 M PGE 2 caused an approximate 65% inhibition of DNA synthesis in SMCs from wild-type mice. The ability of PGE 2 to suppress DNA synthesis was reduced in SMCs from AC3ϩ/Ϫ and AC3Ϫ/Ϫ animals and only reached a 35% inhibition in both cases (Fig. 4A). Thus, although tissues from AC3ϩ/Ϫ mice retained about half of their AC3 expression, this expression level did not appear sufficient to efficiently mediate the growth-inhibitory effects of PGE 2 receptors (Fig. 4, A and C). It is likely that a reduced expression of AC3, which is normally a protein present at low levels, results in inefficient cAMP formation and subsequent inhibition of DNA synthesis and cell replication following exposure of AC3ϩ/Ϫ SMCs to PGE 2 . This concept is supported by the reduced ability of PGE 2 to stimulate cAMP accumulation in AC3ϩ/Ϫ SMCs (Fig. 3).
Changes in DNA synthesis were reflected by changes in the number of cells. Whereas PGE 2 (10 M) gave an approximate 70% inhibition of PDGF-BB-induced proliferation, SMCs from AC3Ϫ/Ϫ mice were completely resistant to the growth-inhibitory effects of PGE 2 (Fig. 4B). Similar results were obtained when the SMCs were stimulated to proliferate by 10% fetal bovine serum (data not shown). Dose-response curves show that the concentration of PGE 2 required to mediate half-maximal inhibition (IC 50 ) of basal and PDGF-BB-stimulated DNA synthesis was in the range of 5 nM in SMCs from wild-type mice (Fig. 4C). This value is similar to the K d values for PGE 2 binding to the EP2 (K d Ϸ5 nM) and EP4 (K d Ϸ1 nM) receptor subtypes (29). PGE 2 resulted in only a 25-40% inhibition of basal and PDGF-BB-stimulated DNA synthesis in SMCs from AC3Ϫ/Ϫ and AC3ϩ/Ϫ mice (Fig. 4C). Similar results were obtained when the SMCs were incubated in the presence of 100 M IBMX, indicating that the inability of PGE 2 to induce growth inhibition in SMCs from AC3-deficient mice was not because of an increased PDE activity in these cells (data not shown). Furthermore, 10 M forskolin induced an 85% inhibition of DNA synthesis in SMCs from AC3Ϫ/Ϫ mice, showing that the proliferation of these cells was inhibited by increased cAMP levels and that other AC isoforms were capable of inducing SMC growth arrest (data not shown).

AC3 Mediates the Growth-inhibitory Effects of PGE 2 in
SMCs-At least 10 isoforms of AC (AC1 through AC10) have been cloned to date, and they are expressed in a tissue-selective manner. However, most tissues express several AC isoforms. Previous studies on arterial smooth muscle from different species have demonstrated expression of multiple AC isoforms. AC2, AC4, AC5/6, and AC7, but not AC3, have been found in bovine pulmonary artery SMCs (30,31). Rat aortic SMCs have been shown to express AC3 and possibly AC8 but not the neuronal AC1 (32). Human SMCs also express several different AC isoforms. In addition to AC3, AC4, AC5/6, AC9, and possi- A, murine SMCs were isolated from the thoracic aorta of wild-type (WT) (AC3ϩ/ϩ) and age-and sex-matched AC3ϩ/Ϫ littermates by enzymatic digestion of the smooth muscle layer. The cells were characterized as SMCs by positive staining for smooth muscle ␣-actin and by morphological criteria. SMCs in six-well plates were preincubated with IBMX and KN-62 and then stimulated for 30 min with PGE 2 and/or A23187 as described in the legend of Fig. 1. AC activity was measured in intact cells according to Wayman et al. (20) after chromatographic separation of cAMP, ATP, ADP, and AMP according to Salomon et al. (26) and was expressed as the percentage of cAMP of total ATP ϩ ADP ϩ AMP. B, SMCs from wild-type and AC3ϩ/Ϫ mice were stimulated with the indicated concentrations of PGE 2 for 30 min. The results are expressed as mean ϩ S.E. of triplicate samples. The experiments were repeated three times with similar results.

inhibits human arterial SMC proliferation
Human SMCs were plated (50,000 cells/well) in 12-well trays. The next day, the medium was changed to DMEM, 1% human PDS, and the cells were preincubated in the presence or absence of 10 M PGE 2 or vehicle (ethanol). After 30 min, 1 nM PDGF-BB or vehicle (10 mM acetic acid, 0.25% bovine serum albumin) was added. New PGE 2 was added daily for 3 days. The cells were trypsinized and counted on day 4. To measure DNA synthesis, the cells were preincubated with 10 M PGE 2 for 30 min and incubated in the presence of PDGF-BB (1 nM) for 18  bly AC2 are expressed. 2 Expression of AC2, AC3, and AC9 in human SMCs is consistent with a recent study on human uterine smooth muscle (33). Thus, it is clear that SMCs express several AC isoforms, the specific biological functions of which are still largely unknown.
It is becoming evident that the properties of different AC isoforms can determine the intracellular response to extracellular stimulation of G-protein-coupled receptors. It has been suggested that ACs act as coincidence detectors. For example, AC3 activity is inhibited by increases in calcium levels in intact cells. This inhibition is due to phosphorylation of Ser-1076 in AC3 by CaM KII, which is activated by calcium/calmodulin (34). AC3 activity is also inhibited by regulator-of-G-proteinsignaling-2, RGS2 (35). Because RGS2 expression can be stimulated by growth-promoting factors in arterial SMCs (36), this provides another mechanism of cross-talk between growth inhibitors and growth factors. Other AC isoforms are stimulated by increases in calcium or regulated by other protein kinases and/or phosphatases (7). Previous studies have shown that specific AC isoforms may have specific biological functions. Accordingly, in NIH3T3 cells, overexpression of AC6 does not affect proliferation, whereas overexpression of AC2 leads to inhibition of cell cycle progression and inhibition of the extracellular signal-regulated kinase pathway (37). Furthermore, AC2 and several other AC isoforms are up-regulated during growth arrest and differentiation of P19 cells (38,39).
Our results show that AC3 is the principal AC isoform activated by PGE 2 in human and murine aortic SMCs despite the presence of other AC isoforms. It is possible that AC3 has a greater sensitivity to G s -activated receptors than other ACs (20) and therefore is the preferred AC isoform activated by all FIG. 5. Increases in levels of calcium may efficiently turn off the cAMP signal in proliferating arterial SMCs. A, agents that elevate levels of cAMP (e.g. PGE 2 ) potently inhibit proliferation of arterial SMCs by antagonizing a number of mitogenic signal transduction pathways (3). In the present report, we show that AC3, an AC isoform that is inhibited in the presence of calcium/calmodulin by CaM KII, mediates the growth-inhibitory effects of PGE 2 in SMCs. B, growth factors, such as PDGF-BB, induce increased levels of intracellular calcium in human SMCs (58). The increased calcium is likely to inhibit AC3 activity by CaM KII-mediated phosphorylation of AC3. Previously, we have shown that a calcium/calmodulin-stimulated cAMP/cGMP PDE (PDE1C) is induced in proliferating SMCs (59). Thus, in arterial SMCs, increases in intracellular calcium induced by growth-promoting factors lead to a simultaneous inhibition of cAMP formation and an induction of cAMP degradation. Together AC3 and PDE1C provide a powerful system to turn off inhibitory cAMP signaling in proliferating arterial SMCs.
FIG. 4. The growth-inhibiting effects of PGE 2 are reduced in SMCs from AC3؉/؊ and AC3؊/؊ mice. Arterial SMCs were isolated from wild-type (WT) (AC3ϩ/ϩ), AC3ϩ/Ϫ, and AC3Ϫ/Ϫ mice. The SMCs were plated (50,000 cells/well in A and C and 30,000 cells/well in B; passages 3-5) in 24-well trays for 3 days followed by a 48-h incubation in the presence of 1% human PDS. In A, the cells were stimulated with or without 10 M PGE 2 for 18 h. The results are the mean ϩ S.E. of experiments performed on SMCs isolated from five wild-type mice, four AC3ϩ/Ϫ mice, and two AC3Ϫ/Ϫ mice, each analyzed in triplicates. SMCs from each mouse were analyzed at least twice with similar results. Thymidine incorporation into DNA in cells incubated in the absence of PGE 2 was set to 100% and was in the range of 200 cpm/g of protein in SMCs from wild-type, AC3ϩ/Ϫ, and AC3Ϫ/Ϫ mice. B, aortic SMCs from wild-type and AC3Ϫ/Ϫ mice were stimulated with or without 1 nM PDGF-BB in the presence or absence of 10 M PGE 2 . The media were replaced by fresh medium and agonists after 3 days, and the cells were trypsinized, fixed, and counted using a Coulter counter on day 6. PDGF-BB increased SMC number from 36,347 Ϯ 851 to 558,686 Ϯ 19,696 cells/well in wild-type SMCs and from 42,883 Ϯ 4,079 to 74,533 Ϯ 9,974 cells/well in AC3Ϫ/Ϫ SMCs. The number of cells stimulated with PDGF-BB in the absence of PGE 2 was set to 100%. The results are mean Ϯ S.E. of triplicate samples. In C, wild-type (WT), AC3ϩ/Ϫ, and AC3Ϫ/Ϫ SMCs were stimulated with the indicated concentrations of PGE 2 in the absence (open circles) or presence (solid squares) of 1 nM PDGF-BB for 18 h. Thymidine was incorporated into DNA during a subsequent 2-h incubation in the presence of 1 Ci/ml [ 3 H]thymidine and was measured as trichloroacetic acid-insoluble radioactivity. The radioactivity of each well was normalized to the amount of cellular protein. The results are expressed as the percentage of thymidine incorporation in cells incubated in the absence of PGE 2 (100%). Thymidine incorporation in wild-type SMCs was 191 Ϯ 7 cpm/g of protein in the absence of PDGF-BB and 413 Ϯ 17 cpm/g of protein in the presence of PDGF-BB. The corresponding values for AC3Ϫ/Ϫ SMCs were 277 Ϯ 28 cpm/g of protein and 431 Ϯ 6 cpm/g of protein, respectively. AC3ϩ/Ϫ SMCs showed similar thymidine incorporation values. The results are mean Ϯ S.E. of triplicate samples. G s receptor agonists. It is also possible that the extracellular and/or intracellular conditions favor activation of AC3. For example, a high cAMP-dependent protein kinase activity is likely to inhibit AC5 and AC6 activities (7). Consistent with AC3 being the principal AC isoform activated by PGE 2 , we further show that AC3 mediates the growth-inhibitory effects of PGE 2 in SMCs.
Regulation of cAMP Levels by Calcium in Human Arterial SMCs-PGE 2 is a major prostanoid secreted by endothelial cells and SMCs (40). The growth-inhibitory actions of PGE 2 are mediated by cAMP. Cyclic AMP inhibits proliferation of SMCs in culture (41)(42)(43)(44) and reduces formation of neointimal lesions after arterial injury in vivo (42,45,46). Previous studies have shown that cAMP, most likely through activation of cAMP-dependent protein kinase, inhibits several mitogenic signal transduction pathways in SMCs. Thus, elevation of cAMP levels results in inhibition of PDGF-induced activation of the extracellular signal-regulated kinase pathway (47)(48)(49), inhibition of S6K1, and inhibition of growth factor-induced phosphorylation of PHAS-1, a translation initiation factor 4E-binding protein that regulates translation initiation (50). In rat SMCs, S6K1 activity and proliferation are inhibited by forskolin at concentrations that do not result in inhibition of the extracellular signal-regulated kinase pathway. It is thus possible that S6K1 or the upstream phosphatidylinositol 3-kinase (PI3K) is especially sensitive to the inhibitory action of cAMP (50). Interestingly cAMP was recently found to inhibit the lipid kinase activity of PI3K in COS cells transfected with the catalytic subunit of PI3K (51). It is not known whether cAMP has several targets in the PI3K pathway in SMCs and whether inhibition of the PI3K pathway indeed mediates the growth-inhibitory actions of cAMP.
Why is AC3 the major AC isoform selected by nature to mediate the growth-inhibitory signaling of PGE 2 ? AC3 activity is inhibited in intact cells by low concentrations of calcium/ calmodulin through the phosphorylation of AC3 by CaM KII (34). Calcium is required for proliferation of many cell types, including SMCs (52,53), and calcium channel blockers can reduce SMC proliferation and accumulation in vitro and in vivo (54,55). Furthermore, many growth factors lead to an increased intracellular calcium level in SMCs (56 -58). Thus, AC3 with its reciprocal regulation by calcium and cAMP-elevating agents provides an excellent switch that the cell can utilize to promote growth versus growth arrest. Interestingly proliferating human SMCs induce a strong expression of a calcium/calmodulin-stimulated cAMP/cGMP PDE (PDE1C) that is absent in quiescent SMCs (59). Thus, in human SMCs, increases in intracellular calcium levels may simultaneously turn off a cAMP-generating enzyme (AC3) and turn on a cAMPdegrading enzyme (PDE1C). Together, AC3 and PDE1C provide an elegant system to efficiently reduce cAMP levels and the growth-inhibitory action of cAMP in SMCs exposed to growth-promoting agents that increase intracellular calcium levels (Fig. 5).