Cyclic AMP Activates the Mitogen-activated Protein Kinase Cascade in PC12 Cells*

Mitogen-activated protein ( M A P ) kinases are acti- vated in response to a large variety of extracellular signals, including growth factors, hormones, and neuro- transmitters, which activate distinct intracellular signaling pathways. Their activation by the CAMP-de- pendent pathway, however, has not been reported. In rat pheochromocytoma PC12 cells, we demonstrate here a stimulation of the MAP kinase isozyme extracellular sig-nal-regulated kinase 1 (ERK1) following elevation of intracellular CAMP after exposure of the cells to isobutyl- methylxanthine, cholera toxin, forskolin, or CAMP- analogues. CAMP acted synergistically with phorbol ester, an activator of protein kinase C, in the stimulation of ERK1. In accordance with this observation, the peptide neurotransmitter pituitary adenylate cyclase-acti- vating polypeptide 38 (PACAP38), which stimulates CAMP production as well as phosphatidylinositol break-down in PC12 cells, was an efficient activator of ERK1. In combination with various growth factors, CAMP acted in a more than additive manner on ERKl activity. of intracellular CAMP increased in vivo szP-labeling of ERK1, suggesting that CAMP stimulated ERKl by activating MAP kinase kinase, an immediate upstream activator of ERKl in the MAP kinase and a CAMP analogue were to the activity of MAP kinase

Mitogen-activated protein ( M A P ) kinases are activated in response to a large variety of extracellular signals, including growth factors, hormones, and neurotransmitters, which activate distinct intracellular signaling pathways. Their activation by the CAMP-dependent pathway, however, has not been reported. In rat pheochromocytoma PC12 cells, we demonstrate here a stimulation of the MAP kinase isozyme extracellular signal-regulated kinase 1 (ERK1) following elevation of intracellular CAMP after exposure of the cells to isobutylmethylxanthine, cholera toxin, forskolin, or CAMPanalogues. CAMP acted synergistically with phorbol ester, an activator of protein kinase C, in the stimulation of ERK1. In accordance with this observation, the peptide neurotransmitter pituitary adenylate cyclase-activating polypeptide 38 (PACAP38), which stimulates CAMP production as well as phosphatidylinositol breakdown in PC12 cells, was an efficient activator of ERK1. In combination with various growth factors, CAMP acted in a more than additive manner on ERKl activity. Elevation of intracellular CAMP increased in vivo szPlabeling of ERK1, suggesting that CAMP stimulated ERKl by activating MAP kinase kinase, an immediate upstream activator of ERKl in the MAP kinase cascade. Supporting this view, forskolin and a CAMP analogue were found to increase the activity of MAP kinase kinase in PC12 cells, alone as well as in combination with phor-bo1 ester. PACAP38 also stimulated in uiuo 32P-labeling of ERKl and MAP kinase kinase activity. Finally, cAMP or PACAP38 increased by %fold nerve growth factorstimulated neurite formation in PC12 cells, which may be correlated with the potentiating effect of these agents on nerve growth factor-stimulated ERKl activity.
In mammalian cells the ability to activate the mitogen-activated protein (MAP)' kinase cascade is a feature common to * This work was supported by funds from the Institut National de la Sant6 et de la Recherche Mkdicale, Universit6 de Nice-Sophia-Antipolis, Association Pour la Recherche Contre le Cancer, ARC Grant 6760, Ligue Nationale Franeaise Contre le Cancer, FBd6ration des Comit6s DBpartementaux, and the ComitB DBpartemental du Var, France. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. EGF, epidermal growth factor; ERK1, extracellular signal-regulated kinase 1; IBMX, isobutylmethylxanthine; IGF-I, insulin-like growth factor-I; MAP kinase, mitogen-activated protein kinase; MAPKK, mitogen-activated protein kinase kinase; MBP, myelin basic protein; NGF, nerve growth factor; PACAP38, pituitary adenylate cyclase-activating many extracellular stimuli including growth factors, hormones, and neurotransmitters (1,2). The various stimuli which can activate the MAP kinase cascade employ distinct initial signaling pathways. Some stimuli activate receptor tyrosine kinases ( 3 4 , or non-receptor tyrosine kinases (6)(7)(8). Other stimuli activate G protein-coupled receptors generating the second messengers diacylglycerol or calcium, or activating ion-channels (9-12). Stimuli generating the prominent second messenger CAMP, however, have not been reported to activate the MAP kinase cascade.
In the case of the receptor tyrosine kinases, a sequence of events leading to MAP kinase activation is emerging. Activated receptor tyrosine kinases increase GTP-binding to c-Ras, through guanine nucleotide exchange factors and adaptor proteins, leading to its activation (13,14). Activated c-Ras may activate c-Raf kinase by direct binding (15)(16)(17). c-Raf can phosphorylate and activate MAP kinase kinase (18-211, which in turn activates MAP kinase (22,231. How other signaling pathways couple to the cascade is less understood. In some cell types, certain G protein-stimulated pathways and protein kinase C-dependent pathways may also activate the cascade through c-Ras/c-Raf (12,24,25), while in other cell types, G protein-stimulated pathways may be predominantly Ras independent and employ MAP kinase kinase kinases distinct from c-Raf (26-29).
The MAP kinase cascade is likely to serve specific functions in different cell types. In rat pheochromocytoma PC12 cells, nerve growth factor (NGF) induces a very robust and sustained activation of the MAP kinase cascade (30-33). Since NGF induces neuronal differentiation in PC12 cells, the cascade has been proposed to mediate neurotrophic signals leading to a neuronal phenotype or other neurotrophic responses in this cell line (30)(31)(32)(33). In agreement with this model, the transfection of PC12 cells with constitutively active, oncogenic Ras or Raf mimicks NGF action by inducing activation of the MAP kinase cascade and neuronal differentiation of PC12 cells, respectively (24,25,34,35).
Elevation of intracellular CAMP, alone or in combination with NGF, has also been reported to promote neuronal differentiation of PC12 cells (36-39). A recently identified neuropeptide, pituitary adenylate cyclase-activating polypeptide 38 (PACAP381, which generates CAMP and stimulates phosphatidyl inositol breakdown, has been shown to promote neurite outgrowth in PC12 cells (40). These observations prompted us to investigate the possibility that CAMP might stimulate the MAP kinase cascade in these cells.
We demonstrate that elevation of the intracellular cAMP stimulates the MAP kinase isozyme, extracellular signal-regulated kinase 1 (ERKl), as well as MAP kinase kinase in PC12 cells. cAMP acts synergistically with phorbol ester, and thus polypeptide, 38-amino acid form: PMA, phorbol 12-myristate 13-acetate; PAGE, polyacrylamide gel electrophoresis.

6207
protein kinase C. This interaction may serve as a model for signal integration at the level of the MAP kinase cascade. Such an interaction may be used by different ligands that stimulate CAMP generation or phosphatidylinositol breakdown, respectively, or ligands which stimulate both pathways, like PACAP38. Finally, CAMP increased ERKl activation as well as neurite outgrowth in response to NGF in PC12 cells.
Most importantly, our observations describe a novel regulation of the MAP kinase cascade, ie. stimulation by intracellular CAMP. In addition, these data may provide new insight into the neurotrophic effects of CAMP (and physiological ligands raising intracellular CAMP) previously described in PC12 cells, but also in sympathetic and sensory neurons and neuroblastoma cells where neurotrophic effects of CAMP have been reported (41,42). EXPERIMENTAL PROCEDURES carboxyl-terminal of ERKl(356-367) (TAFCFQPGAPEAP), synthesized Materials-The keyhole limpet hemocyanin-coupled 12-amino-acid by Neosystem (Strasbourg, France) was used to generate rabbit polyclonal antisera (43). For the generation of polyclonal rabbit antisera to MAP kinase kinase, the 17-amino-acid amino-terminal (PKKKPTPIQL-NPAPDGS (44) was used as an antigen. Sodium orthovanadate, BSA, leupeptin, phenylmethylsulfonyl fluoride, MBP from bovine brain, protein A-Sepharose, Triton X-lOO,8-(4-chlorophenylthio~-cAMP, forskolin, cholera toxin, N6,2'-O-dibutyryl CAMP, iP,2'-O-dibutyryl-cGMP, 3-isobutyl-1-methyl-xanthine (IBMX), and phorbol 12-myristate 13-acetate (PMA) were purchased from Sigma. The final experimental concentration of dimethyl sulfoxide, used as carrier for PMA and forskolin, was 0.04% which had no effect on ERKl activity in itself. Ovine PACAP38 (identical to rat PACAP38) from Peninsula Laboratories, Inc, was in part a generous gift from Prof. J. Fahrenkrug, Bispebjerg Hospital, Copenhagen, Denmark. Mouse 7 s NGF was purified and generously provided to us by Dr P. Kitabgi (Nice-Sophia Antipolis, France).
Immunopurification of ERKl or MAPKKActivated in Intact Cells-Following the serum starvation period, PC12 cells were incubated for various times with peptides or agonists. Medium was aspirated and the cell monolayers were solubilized for 15 min in solubilization buffer containing 1% (v/v) Triton X-100, 50 mM HEPES, pH 7.5, 150 mM NaCl, 10 mM Na,P,O,, 2 mM Na3V04, 100 mM NaF, 100 unitdml aprotinin, 20 p~ leupeptin, and 0.2 mg/ml phenylmethylsulfonyl fluoride. The solubilized cell extracts were clarified by centrifugation at 18,000 x g for 15 min and then incubated for 2 h with antibodies to ERKl or MAPKK preadsorbed on protein A-Sepharose beads. Following the immunoprecipitation period pellets were washed three times with solubilization buffer.
ERKl Assay-Pellets with immunoprecipitated ERKl were washed two times with HNTG buffer (50 mM HEPES, 150 mM NaC1, 10% (v/v) glycerol, 0.1% (v/v) Triton X-100 with 0.2 mM Na,VOJ, dried, and resuspended in 50 pl of HNTG buffer supplemented with 0.2 mM Na3V04, 100 unitdml aprotinin, 20 p~ leupeptin, and 0.2 mg/ml phenylmethylsulfonyl fluoride. Phosphorylation of MBP was started by addition of 150 pg/ml MBP, 10 m~ magnesium acetate, 1 mM dithiothreitol, and [y-32P]ATP (5 p~, 33 Ci/mmol) added as a 6-fold concentrated mixture in a volume of 10 pl. The phosphorylation reaction was allowed to proceed for 15 min a t room temperature (linear assay condition) and was stopped by spotting Whatman P-81 filter papers which were then dropped into 0.1% (v/v) orthophosphoric acid. The papers were washed overnight with several changes in this solution, rinsed once in ethanol, air dried, and radioactivity was determined by Cerenkov-counting. The reaction blank which received identical treatment was a mixture containing all of the reagents but with the omission of cell lysate during immunoprecipitation was subtracted from a11 values.
MAP Kinase Kinase Assay-MAPKK activity was measured in a reconstitution assay as the ability of immunopurified MAPKK, stimulated in intact cells, to activate added recombinant rat ERKl (19,20).
ERKl activity was then measured using MBP as a substrate. Briefly, pellets with immunoprecipitated MAPKK were washed two times in 50 m~ HEPES, pH 7.4, dried, and resuspended in 50 pl of 50 mM HEPES, pH 7.4, containing bacterially expressed, recombinant ERKl and 0.2 m~ NaaV04, 100 unitdm1 aprotinin, 20 VM leupeptin, and 0.2 mg/ml phenylmethylsulfonyl fluoride. The phosphorylation cascade was started by the addition of [Y-~~PIATP (50 p~, 50 Ci/mmol), 150 pglml MBP, 15 mM magnesium chloride, 1 mM EGTA added as a 6-fold concentrated mixture in a volume of 10 pl. The phosphorylation reaction was allowed to proceed for 10 min a t room temperature and was stopped by spotting Whatman P-81 filter papers which were then dropped into 0.1% (v/v) orthophosphoric acid. The papers were washed overnight with several shifts in this solution, rinsed once in ethanol, air dried, and radioactivity was determined by Cerenkov-counting. The reaction blank, which received identical treatment, was a mixture containing all of the reagents except the cell lysate during immunoprecipitation, and was subtracted from all values. Control experiments showed that, when MBP or recombinant ERKl were omitted from the phosphorylation reaction with immunopurified MAPKK from unstimulated as well as stimulated cells, radioactivity associated with the papers dropped by approximately 90%. The remaining 10% may derive from (auto-) phosphorylation of the recombinant ERKl or [-p3"?PIATP adsorbed to MBP, respectively. The data were not corrected for these minor contributions. 32P-Labeling of PC12 Cells-Cells were washed twice in serum-and phosphate-free culture medium with 0.2% BSA and incubated for 3.5 h in this medium containing [32Plorthophosphate (0.5 mCi/ml). At the end of the labeling period, cells were stimulated as indicated. Following stimulation the monolayers were washed once in ice-cold phosphatebuffered saline and solubilized for 15 min in the solubilization buffer. The cell extracts were clarified by centrifugation at 18,000 x g for 15 min and incubated for 2 h with antibodies to ERKl preadsorbed to protein A-Sepharose beads. Following the incubation period, the beads were washed six times in solubilization buffer. Pellets were dried, resuspended in Laemmli buffer (3% SDS), and boiled for 5 min. Dissolved proteins were submitted to SDS-PAGE under reducing conditions on a 10% acrylamide resolving gel followed by autoradiography of the dried gel.
Neurite Assay-A single cell suspension of PC12 cells was produced by trituration of cells through a syringe needle following trypsination.
Cells were seeded in 12-well dishes at lo4 cells/cm2 and cultured for 1 day before they were shifted to medium containing 0.5% serum and 0.2% BSA and stimulated as indicated. After 24, 48, and 72 h of treatment, all cells in randomly selected fields were analyzed for neurite outgrowth using phase contrast microscopy. Neurites were defined as processes exceeding the length of one cell soma diameter and were generally tipped with a growth cone. 150-250 cells were scored per well and all treatments were performed in duplicates.

CAMP Stimulates ERKl in PC12 Cells-We first measured
ERKl activity in PC12 cells treated with a variety of agents which share the ability to raise the intracellular level of CAMP or being themselves a CAMP analogue. After stimulation of ERKl in PC12 cells, the cell monolayers were solubilized. ERKl was immunopurified from the cell lysates, and its activity was measured in vitro using myelin basic protein as a substrate.
The various agents, shown in Fig. 1, were found to increase the activity of ERKl from 2-to 12-fold above the level seen in unstimulated cells (Fig. 1). Maximal activation occurred rapidly, within 5-25 min depending on the agent, except for cholera toxin, which may be due to the fact that it raises intracellular CAMP only slowly. The agents increase CAMP in widely differing ways. IBMX increases CAMP levels by inhibiting its degradation by CAMP phosphodiesterase (46). Cholera toxin activates the G protein Gs through ADP-ribosylation of its a subunit, which subsequently stimulates adenylyl cyclase (47).
Forskolin increases intracellular CAMP by direct binding and activation of adenylyl cyclase (48). Dibutyryl-CAMP (dbt-cAMP) permeates the cell membrane and is metabolized in the cell to generate the active CAMP analogue (491, while chlorophenylthio-CAMP (CPT-CAMP) is a membrane permeant CAMP analogue (50). The second messenger cGMP is structurally related to CAMP and elicits some cellular responses equipotently to CAMP (51). However, cGMP, when added as dibutyryl-cGMP, had no effect on ERKl activity, indicating that the CAMP effect is not due to cross-reaction with a cGMP-dependent signaling pathway (data not shown). The stimulatory effect of CAMP on ERKl is not limited to the PC12 cell subclone used in this study, since we observed similar effects of CAMP in two other PC12 subclones (data not shown).
Importantly, the five agents used in Fig. 1 increase intracellular CAMP each in their own distinct way. Their sole common denominator is the ability to elevate intracellular CAMP. We therefore conclude that CAMP can activate ERKl in PC12 cells.

CAMP Acts Synergistically with Phorbol Ester to Stimulate
ERKl in PC12 Cells-Although CAMP stimulates ERK1, it has a rather modest effect compared with certain other activators in PC12 cells, e.g. phorbol ester (PMA), an activator of protein kinase C (52). CAMP, however, was found to greatly amplify the stimulatory effect of PMAon ERKl in PC12 cells. Fig. 2.4 shows the time course of ERKl activation by CPT-CAMP and PMA, both used at maximally stimulating concentrations. At each time point tested, the effect of PMA was potentiated by a factor of 3 to 4 by the presence of CPT-CAMP.
IBMX, cholera toxin, forskolin, and dibutyryl-CAMP (but not dibutyryl-cGMP, not shown) were also found to potentiate the effect of PMA on ERK1, indicating that the potentiation was indeed due to an increase in intracellular CAMP (Fig. 2 B ) .
ERK2, another MAP kinase isozyme expressed in PC12 cells (5), appeared to be regulated by cAMP in much the same way as ERK1. In experiments performed like the ones shown in Fig.  2A, but instead using antibodies directed against the COOH terminus of ERK2, we found that CAMP by itself stimulated ERK2 activity only slightly but potentiated severalfold the effect of PMA (data not shown).
Finally, we tested a neuropeptide, PACAP38, which stimulates CAMP synthesis as well as phosphatidyl inositol breakdown in PC12 cells (40, 53). PACAP38 was found to be an efficient activator of ERK1, showing a rapid, but transient, activity peak, reached within 5 min (lower curves in Fig. 8).
Half-maximal and maximal stimulation of ERKl by PACAP38 were observed a t approximately 25 and 200 nM, respectively 1 " " ' 1 1 1 " 1 1 1 1 " '  CAMP. Serum-starved PC12 cells were incubated with the various agents at the following concentrations and for the following periods of time: IBMX, 1 mM, 10 min; cholera toxin, 500 ng/ml, 90 min; forskolin (FOR j, 10 PM, 15 min; dbt-cAMP, 1 m~, 20 min. PMA was added at 2 p~ for the same length of time as the agent with which its interaction was investigated, except in the experiment with cholera toxin, where PMA was added for the final 10 min. At the end of the incubation period cells were solubilized. MBP phosphotransferase activity of immunoprecipitated ERKl was measured and expressed in percent of ERKl activity stimulated by PMA alone. Data are means of triplicate wells k S.D. of one representative experiment performed three times or more with similar results. (data not shown). These values are consistent with the binding affinity of the PACAP type 1 receptor and the biological potency of PACAP38 in PC12 cells and chromaffin cells (40, [53][54][55] suggesting receptor specificity and physiological relevance of ERKl stimulation by PACAP38. PACAP38 stimulation of ERKl was inhibited by 40% following protein kinase C downregulation, through overnight preincubation with PMA (data not shown), suggesting protein kinase C involvement in the PACAP38 response. Forskolin acted non-additively with PACAP38 on ERKl activation (data not shown), indicating CAMP involvement in the PACAP38 response. Taken together, the observations with PACAP38 suggest that the synergistic stimulation of ERKl by CAMP and activators of protein kinase C, shown in Fig. 2, may occur in response to a physiological stimulus in PC12 cells. ERKl Activation in PC12 Cells-In PC12 cells, peptide growth factors like NGF or epidermal growth factor (EGF) are among the strongest activators of ERKl (30-33) capable of increasing ERKl activity more than 100-fold in our assay system. In contrast, we found that insulin-like growth factor I (IGF-I) by itself has a weak effect in PC12 cells. However, a strong synergistic activation of ERKl was observed by CPT-CAMP or forskolin in the presence of a maximal concentration of IGF-I (Fig. 3). In combination with the strongly stimulating growth factors, CPT-cAMP or forskolin were also found to potentiate ERKl activation in a more than additive manner, as shown with NGF, used at a maximally stimulating concentration (see Fig. 9). In combination with EGF, however, a potentiation by CAMP was most clearly observed, when EGF was used at a half-maximal concentration (data not shown).

CAMP Acts Synergistically with Receptor Orosine Kinases on
These observations indicate that in PC12 cells CAMP also potentiates activation of ERKl by growth factors, which initiate their signaling through tyrosine kinase receptors.

CAMP Increases ERKl Phosphorylation in Intact PC12
Cells-The activity of ERKl is known to be stimulated upon phosphorylation. We therefore investigated whether elevation of intracellular CAMP increased the level of ERKl phosphorylation. Cells were metabolically labeled with [32Plorthophosphate and exposed to various agonists. Thereafter, cells were solubilized, and cell extracts were subjected to immunoprecipitation with antibodies to ERK1. The immunoprecipitates were analyzed by SDS-PAGE followed by autoradiography.
We found that both CPT-CAMP and forskolin increased [32P]phosphate incorporation in a 44-kDa protein (Fig. 4, lefc  and right panels, respectively). This protein has previously been identified as ERKl(33,43,56). PMA also increased ERKl phosphorylation. However, when PMA was added in combination with CPT-CAMP or forskolin, the level of ERKl phosphorylation was markedly higher than with either agent alone. Phosphorylation of ERKl in response to NGF was also increased by CPT-CAMP. Finally, PACAP38 increased ERKl phosphorylation to a high level after 5 min of stimulation, which had returned to near-basal levels within 45 min.
The labeling of an additional phosphoprotein, having a M, of approximately 90,000, was increased by stimulation with CPT-CAMP and PMA, NGF, or PACAP38 at 5 min. We have previously shown that this protein corresponds to the pp90rsk protein kinase, which coimmunoprecipitates with ERKl (33,56) and which is a presumed physiological substrate of ERKl (4,57).
In summary, the increase in the phosphorylation of ERKl in response to CAMP (alone or in combination with other stimuli) paralleled the stimulation of its kinase activity observed in %use Cascade in PC12 Cells Figs. 1 and 2. These data suggest that CAMP activates ERKl by increasing its phosphorylation.
CAMP Activates MAP Kinase Kinase in PC12 Cells-MAPKK functions immediately upstream of ERKl in the MAP kinase cascade, stimulating its activity by phosphorylating it on threonine and tyrosine residues (23). To investigate whether cAMP stimulation of ERKl phosphorylation and activation could be traced upstream in the MAP kinase cascade, we investigated the effect of CPT-CAMP and forskolin on the activity of MAPKK. MAPKK activity was measured in a reconstitution assay as the ability of W K K , immunopurified from intact cells, to activate bacterially expressed, recombinant ERK1. The activity of ERKl was measured using MBP as a substrate.
In Fig. 5 the results of three experiments are shown (panels A-C, respectively). Data were not pooled in this series of experiments, since the variation between individual experiments would obscure the cross-talk between CAMP-and PMA-stimulated pathways. We found that CPT-CAMP or forskolin increased the activity of W K K from 75 to 250% above its level in unstimulated PC12 cells (Fig. 5, A X ) . PMA increased MAPKK activity to approximately the same extent as cAMP at the time point tested. The effects of CPT-CAMP or forskolin in combination with PMA were additive at the least or significantly synergistic, increasing MAPKK activity from 400 to 1100% above basal levels. Finally, we tested the effect of the physiological ligands on MAPKK activity. PACAP38 stimulated MAPKK with a time course paralleling that observed for ERKl activation and phosphorylation (Fig. 6). For comparison, NGF was found to increase W K K activity from 1000 to 2000% over basal levels depending on the experiment (data not shown).
In summary, the pattern of MAPKK activation by the various stimuli generally reflected the pattern observed for ERKl activation (and phosphorylation) in PC12 cells. The stimulation as expressed in percent over basal was approximately 10 times lower in the MAPKK assay as compared with the ERKl assay. This may reflect the fact that the reconstitution assay did not mirror the exact extent of stimulation of W K K activity that may occur in the living cell. It could also be due t o the fact that W K K is one step upstream of ERKl in a potential signal amplification cascade.
In conclusion, elevation of intracellular CAMP stimulates MAPKK in PC12 cells. CAMP may therefore act, at least in part, at the level of MAPKK to increase ERKl phosphorylation and activity. CAMP can therefore be said to stimulate the W kinase cascade in PC12 cells.

CAMP and PACAP38 Potentiate NGF-stimulated Neurite Outgrowth with a Concomitant Potentiation of NGF-stimulated
ERKl Activation-Finally, we investigated ERKl activation by CAMP or PACAP38 in relation to the neuronal differentiation of PC12 cells. NGF is known to promote the differentiation of PC12 cells into a sympathetic neuron-like phenotype which is associated with the progressive outgrowth of neuronal processes as originally described (45) and illustrated in Fig. 7. Likewise, it has been reported that CAMP promotes stable neurite outgrowth in some PC12 subclones. In other PC12 subclones, however, cAMP has only a modest and transient neurite promoting effect by itself, but it can potentiate the effect of NGF on long term, stable differentiation. The PC12 cells used in this study belong to the latter category as evidenced by the potentiating effect of CPT-CAMP or forskolin on NGF-stimulated differentiation (Fig. 7). Likewise, PACAP38 had only very little neurite-promoting effect by itself in our PC12 cells, but it effectively potentiated neurite formation in response to NGF. After prolonged exposure to NGF (1-2 weeks), in the presence or absence of CPT-CAMP, forskolin or PACAP38, most of the PC12 cells appeared to have neurites, and the cultures dis- played a similar, dense network of neurites. The effect of CAMP or PACAP38 may thus be characterized as an acceleration of the NGF-induced neuronal differentiation process. As PACAP38 has the potential to stimulate also protein kinase C, we tested the effect of PMA in combination with forskolin or NGF (Fig. 7). PMA led to only a small potentiation of NGFinduced neurite outgrowth during the first 24 h, but thereafter did not potentiate the neuronal differentiation of PC12 cells. Prolonged exposure of cells to PMA, however, is known to downregulate protein kinase C, which may explain the lack of any detectable effects in this long term experiment.
NGF is known to induce a robust and prolonged activation of ERKl in PC12 cells as previously described (30-33) and shown in Fig. 8. Interestingly, PACAP38, while having a transient effect itself, increased the sustained activation of ERKl by NGF in a more than additive manner. This potentiation was seen with a half-maximally as well as a maximally stimulating concentration of NGF (Fig. 8, A and B, respectively). Forskolin, CPT-CAMP, and PMA also increased NGF-stimulated ERKl activity during the sustained secondary phase of the NGFresponse as measured after 1 h of stimulation (Fig. 9). At this In summary, these data suggest that there may be a causal relationship between the potentiating effect of cAMP (and PACAP38) on NGF-induced ERKl activation and NGF-induced neurite outgrowth.

DISCUSSION
In the present study we describe the activation and phosphorylation of ERKl in response to several agents known to increase the intracellular level of CAMP. Since these agents act at distinct levels in the pathway of cAMP generation or degradation we conclude that the second messenger CAMP can stimulate ERKl in PC12 cells. In mammalian cells, the vast majority of CAMP effects can be attributed to activation of the CAMP-dependent protein kinase. This may also be the case for cAMP stimulation of ERKl in PC12 cells. However, protein kinase A-independent CAMP responses have been described. For instance, the inhibitory effect of CAMP on glucose transport is thought to be mediated by direct binding of the nucleotide to the glut 4 transporter molecule (58). Further, cardiac pace- maker ion channels (591, ion channels in olfactory receptor cilia (511, and potassiudcalcium-specific Drosophila eug channels (60) have been reported t o be activated directly by CAMP. PC12 cells are neuroendocrine cells in which several biological responses to extracellular stimuli are elicited via ion channel activation. A mechanism for ERKl regulation involving the activation of ion channels by cAMP may therefore be a possibility.
cAMP increased the in vivo phosphorylation of ERKl as well as the activity of MAPKK in PC12 cells. It therefore appears that CAMP stimulates ERKl through activation of the MAP kinase cascade operating at the level of MAPKK or upstream of it. The synergistic interaction observed between CAMP and PMA on ERKl was reflected in MAPKK activation, suggesting that this interaction also occurred upstream in the MAP kinase cascade, again being at least at the level of MAPKK. The MAP kinase cascade seems to utilize several activators upstream of MAPKK including Raf-1 kinase, MAPKK kinases, and possibly other functional equivalents, perhaps in a cell type-specific or pathway-specific manner (27,29). We found a very pronounced electrophoretic mobility shift of Raf-1 kinase in response to NGF, which is indicative of its activation, as described previously in PC12 cells (25,611. We observed no such mobility shift of Raf-1 in response to CPT-CAMP or forskolin (data not shown). While we cannot exclude that CAMP is activating Raf-1 without causing a concomitant mobility shift, our data would indicate that CAMP activates MAPKK through a MAPKK kinase, or a functional equivalent, which is distinct from Raf-1 kinase. We are currently investigating whether the other members of the Raf-family, A-Raf and B-Raf, could be involved.
Our observations may have several biological implications. An interesting feature of the CAMP stimulation of ERKl is its synergistic interaction with PMA, which presumably reflects synergy with the protein kinase C-dependent pathway. This finding opens the possibility that the MAP kinase cascade may integrate and amplify signals originating from receptors employing CAMP and receptors using phosphatidyl inositol breakdown products as second messengers, respectively. By extension, receptors activating both pathways, may equally well employ this mechanism. With PACAP38 we found a time course and efficiency of activatiodphosphorylation of ERK1, MAPKK, and pp90rsk, which were overall similar to that observed with CAMP in combination with PMA. It is reasonable to believe that PACAP38 employs the synergistic interaction between CAMP and phosphatidyl inositol-derived second messengers to activate the MAP kinase cascade in PC12 cells. With the exception of the early events, which are CAMP synthesis, phosphatidylinositol breakdown, and increases in the intracellular calcium, nothing is known about PACAP38 signal generation. The present findings, showing the activatiodphosphorylation of MAPKK, ERKl as well as pp90rsk, make these kinases putative transducers of PACAP38-stimulated signaling. These observations may be relevant for the understanding of the endocrine function of chromafin cells, as recent evidence strongly suggests that in the adrenal medulla PACAP38 functions as a peptide neurotransmitter to induce catecholamine secretion from the chromaffin cells (55,62). Rapid and transient activation of MAP kinase (like with PACAP38) has been observed in chromaffin cells in response to analogues of acetylcholine, the major physiological secretagogue in the adrenal medulla. MAP kinase activation has therefore been proposed to be part of the transductional mechanism in the stimulus/secretion coupling in chromaffin cells (10). The same could be true for PACAP38. Further, the activity of tyrosine hydroxylase, the rate-limiting enzyme in catecholamine biosynthesis, is stimulated upon phosphorylation, which occurs in response to secretagogues, and which serves to replenish catecholamine stores following secretion (63,64). In uiuo, phosphorylation occurs on serine residues within three distinct amino acid sequences, which are phosphorylation sites for protein kinase A, protein kinase C, calciudcalmodulin-dependent protein kinase, and ERK (65)(66)(67). The present demonstration that PACAP38 stimulates ERKl illustrates that PACAP38 possesses a potential unique among secretagogues to regulate tyrosine hydroxylase activity by all four kinases.
In PC12 cells the capacity to activate the MAP kinase cascade is shared by neurotrophic factors and oncogenes, and its activation is thought to play a role in neuronal differentiation (30)(31)(32)(33). Stimulation of the MAP kinase cascade by PACAP38 or cAMP may therefore alternatively (or additionally) be viewed in the light of the neurotrophic effects of PACAP38 or CAMP on PC12 cells and possibly other neural cell types. In particular, the potentiation by CAMP of NGF-induced MAP kinase activation may help to explain the synergistic interaction between cAMP and NGF on neurite outgrowth in PC12 cells reported previously (37, 38).
Finally, our data illustrate that at least in PC12 cells the MAP kinase cascade provides meeting points for cross-talk between the different signaling pathways initiated by receptor tyrosine kinases, protein kinase C, and CAMP.