Pituitary Adenylyl Cyclase-activating Polypeptide 38 Reduces Astroglial Proliferation by Inhibiting the GTPase RhoA*

Pituitary adenylyl cyclase-activating polypeptide 38 (PACAP38) plays an important role in the proliferation and differentiation of neural cells. In the present study, we have investigated how PACAP38 inhibits the proliferation of cultured neocortical astroglial cells. When applied to synchronized cells during the G 1 phase of the cell cycle, PACAP38 diminished the subsequent nuclear uptake of bromodeoxyuridine. When applied for 2 days, it reduced the cell number. PACAP38 did not exert its antiproliferative effect by activating protein kinase A. It also did not reduce the activity of mitogen-activated protein kinases essential for G 1 phase progression. In- stead, PACAP38 acted on a member of the Rho family of small GTPases. It reduced the activity of RhoA as was shown with a Rhotekin pull-down assay. The decrease in endogenous RhoA activity induced by treatment of the cells with C3 exotoxin or by expression of dominant negative RhoA also reduced the nuclear uptake of bromodeoxyuridine. In contrast, expression of constitutively active RhoA prevented the effect of PACAP38. Our data show a novel signal transduction pathway by which the neuropeptide influences cell proliferation.

Pituitary adenylyl cyclase-activating polypeptide 38 (PACAP38) plays an important role in the proliferation and differentiation of neural cells. In the present study, we have investigated how PACAP38 inhibits the proliferation of cultured neocortical astroglial cells. When applied to synchronized cells during the G 1 phase of the cell cycle, PACAP38 diminished the subsequent nuclear uptake of bromodeoxyuridine. When applied for 2 days, it reduced the cell number. PACAP38 did not exert its antiproliferative effect by activating protein kinase A. It also did not reduce the activity of mitogen-activated protein kinases essential for G 1 phase progression. Instead, PACAP38 acted on a member of the Rho family of small GTPases. It reduced the activity of RhoA as was shown with a Rhotekin pull-down assay. The decrease in endogenous RhoA activity induced by treatment of the cells with C3 exotoxin or by expression of dominant negative RhoA also reduced the nuclear uptake of bromodeoxyuridine. In contrast, expression of constitutively active RhoA prevented the effect of PACAP38. Our data show a novel signal transduction pathway by which the neuropeptide influences cell proliferation.
Pituitary adenylyl cyclase-activating polypeptide 38 (PACAP38) 1 and vasoactive intestinal polypeptide (VIP) are neuropeptides of the secretin-glucagon family (1). Whereas both peptides stimulate the G protein-coupled receptors VPAC1 and VPAC2 at subnanomolar concentrations, only PACAP38 stimulates PAC1 receptors with a potency in the same range (2,3). During the embryonic and postnatal period, the neuropeptides and the receptors are expressed in cells of the rodent neocortex and its growth zones. There is strong evidence for the involvement of these peptides in neural cell proliferation, phenotypic determination, differentiation, and survival (4 -12).
Activation of PAC1 receptors by PACAP38 can change the proliferation of neural cells. However, inhibitory as well as stimulatory effects mediated by protein kinase A (PKA) have been observed in cultured neocortical or cerebellar neurons and spinal cord glial cells (3,4,(12)(13)(14)(15)(16). These contradictory effects may be related to the diversity in receptor isoforms and their signal transduction pathways. The 7 isoforms reported differ in their third intracellular loop and can couple to adenylyl cyclase as well as phospholipases C and D (12,17).
In the cellular proliferation cycle, cyclins and cyclin-dependent kinases (CDKs) determine the progression through the G 1 phase into the S phase. Thus, the association of cyclin D with CDK4/6 and of cyclin E with CDK2 is important for G 1 progression and G 1 /S transition, respectively. The complexes induce the phosphorylation of the retinoblastoma gene product that allows the transcription of E2F-regulated genes (18). Upon activation, extracellular signal-related kinases 1 and 2 (ERK1 and ERK2) increase the activity of cyclin D1. However, only the sustained activation of ERKs and cyclin D1 leads to G 1 progression (19,20). In primary cells, such a sustained activation is only observed if growth factors are applied to cells attached to the matrix (21)(22)(23)(24)(25). The small GTPases of the Rho family RhoA, Rac, and Cdc42 organize the actin cytoskeleton and thereby cell attachment (26 -28). RhoA has been shown to be necessary for sustained activation of ERKs and cyclin D1 (29). In addition, RhoA supports G 1 /S transition by preventing the action of the CDK inhibitors p21 Cip1 and p27 Kip1 , which inactivate the cyclin E-CDK2 complex (30 -35).
In preliminary experiments, we have observed that PACAP38 strongly inhibited the S phase-dependent nuclear uptake of bromodeoxyuridine (BrdUrd) in cultured astroglial cells from the rat neocortex. This apparent antiproliferative effect was in contrast to our previous observation of a PACAP38-induced activation of ERKs (36). In the present study, we have examined the antiproliferative effect of PACAP38 in cultured astroglial cells. The antiproliferative effect of PACAP38 was shown to be because of inactivation of the small GTPase RhoA that was essential for the proliferation of cultured astroglial cells.

EXPERIMENTAL PROCEDURES
Primary Cultures of Dissociated Neocortical Astroglial Cells-Cultures were set up as described previously (37). Briefly, neocortices of newborn Wistar rats were dissected, chopped, and dissociated by trituration. Cell debris was removed by 3 centrifugations (150 ϫ g, 5 min). The last cell pellet was suspended in Dulbecco's modified essential medium supplemented with 10% endotoxin-free fetal calf serum (FCS; Biochrom KG, Berlin, Germany). Viable cells were seeded in six-well plates at a density of 2.5 ϫ 10 6 cells per well (35 mm diameter, BD Biosciences) and incubated with the FCS-containing Dulbecco's modified essential medium. After 4 days, the culture medium was renewed. Experiments were performed 6 to 7 days after seeding, when the cells were pre-confluent. At this time, more than 96% were astroglial cells of type I, as indicated by the presence of glial fibrillary acidic protein and the absence of A2B5 antigen (37).
Nuclear Uptake of BrdUrd-The nucleotide BrdUrd is taken up during the S phase of the cell cycle. Unless indicated otherwise, experiments were performed in astroglial cells synchronized by FCS withdrawal from the incubation medium for 24 h. To start again the G 1 phase progression, FCS (final concentration 10%) was added for 16 h. PACAP38 and other agents to be tested were present during this period. BrdUrd (10 M; Roche Diagnostics GmbH, Mannheim, Germany) was added to the medium for the last 60 min of the incubation. Cells were fixed with methanol (Ϫ20°C), washed with phosphate-buffered saline for 5 min, and then incubated with 2 N HCl (10 min, 37°C). After the medium had been neutralized with borate buffer (0.1 M, pH 8.5), the cells were washed and incubated for 3 h with a mouse anti-BrdUrd antibody (Roche Diagnostics). The immune complex was detected with a Cy3-labeled secondary goat anti-mouse antibody (Dianova, Hamburg, Germany).
To estimate the total cell number, fixed cells were stained with 4Ј,6-diamidine-2Ј-phenylindole dihydrochloride (DAPI; Roche Diagnostics) for 5 min. Apoptotic cells were stained with propidium iodide (Sigma, Taufkirchen, Germany) added to the incubation medium at a concentration of 5 g/ml for 20 min prior to fixation.
MAP Kinase Assays-ERKs were determined in their activity with a non-radioactive procedure (New England BioLabs, Schwalbach, Germany). After incubation with PACAP38, the astroglial cells were lysed in buffer. After sonication and centrifugation protein concentrations were determined. Aliquots containing 300 g of protein were used to immunoprecipitate active ERKs with an immobilized monoclonal anti-phospho-ERK antibody (overnight at 4°C). To estimate the activity of the precipitated ERKs, the washed precipitate was incubated for 30 min with an Elk-1 fusion protein at 30°C in the presence of ATP. Reactions were stopped by addition of 3ϫ Laemmli buffer. After electrophoresis in a 12.5% SDS-PAGE, the proteins were transferred to a nitrocellulose membrane. Unspecific binding was blocked with 5% nonfat dry milk in Tris-buffered saline containing 0.05% Tween 20 (TBS-T). Elk-1 proteins phosphorylated during the incubation were detected with a phospho-Elk-1 (Ser-383) antibody overnight at 4°C. After washing with TBS-T, the blot was incubated for 1 h with an anti-rabbit antibody coupled to horseradish peroxidase (1:2000). Proteins were visualized using the Phototype®-horseradish peroxidase Western blot Detection Kit according to the manufacturer's instructions.
Infection of Astroglial Cells with a Tetracycline responsive Adenovirus System-The adenoviral Tet-On system used here (BD Adeno-X Tet-On Expression System 2, Clontech) is based on the infection of target cells with two different viruses: the recombinant adenovirus containing the gene-of-interest and the Tet-On virus. The latter produces a protein that binds to the Tet-responsive expression cassette in the recombinant adenovirus thereby initiating transcription. As the protein can only bind in the presence of doxycycline, the system allows the graded expression of the gene of interest by application of doxycycline (39,40).
The coding regions of dominant negative RhoA(T19N) (dnRhoA), constitutively active RhoA(G14V) (caRhoA) or PKI were cloned into the tetracycline-responsive pAdeno-X vector according to the manufacturer's instructions. The vector was modified by introducing an enhanced green fluorescent protein (EGFP) and an internal ribosomal entry site into the EcoRI and XhoI sites. Following propagation in HEK 293 cells, the viral lysate was purified by filtration (pore size 0.45 m) and viral titers were determined using the Adeno-X Rapid Titer Kit (Clontech). In general, infectious units in the range of 3-7 ϫ 10 9 /ml were obtained. All constructs were verified by sequence analysis. Six to 7 days after seeding, neocortical astroglial cells were infected for 24 h with recombinant adenovirus and Tet-On virus (in a ratio of 1:1; multiplicity of infection ϳ10). To induce expression of EGFP and the protein of interest, 1 M doxycycline was added to the incubation medium. 36 h after addition of doxycycline the actual experiments were performed. In all experiments, we obtained an infection rate Ն90%, measured by counting EGFP positive cells compared with DAPI positive cells.
Transient Transfection of Astroglial Cells-Neocortical astroglial cells were transfected with a pEGFPC1 vector (Clontech) that expressed an enhanced green fluorescent protein/ERF fusion protein (pEGFP-ERF) (41). The vector was a gift of Dr. Mavrothalassitis (IMBB-Forth, University of Crete, Crete, Greece). Astroglial cells were used after 7 days in vitro. Transient transfection was performed with a mixture of DOTAP and cholesterol (1:1) as a DNA carrier; the lipid: DNA ratio was 8:1. Three g of pEGFP-ERF plasmid were added per 35-mm well. Cells were incubated at 37°C and pH 7.3. After 6 h, the transfection medium was removed and fresh Dulbecco's modified essential medium supplemented with 10% FCS was added. After 64 h, the cells were used for the experiments. Control experiments were performed with cells transfected with pEGFPC1 without ERF. In these cells, the intracellular localization of EGFP fluorescence was not changed by treatment with serum-free medium or PACAP38 (data not shown).
Expression of GST-C21 Protein-The fusion protein GST-C21 contains the Rho-binding region of the Rho effector Rhotekin. The pGEXC21 plasmid encodes the N-terminal 90 amino acids of Rhotekin (42). It was a gift of Prof. Collard (Amsterdam, The Netherlands). The GST fusion protein was expressed in Escherichia coli BL21 cells grown at 37°C. Expression was induced by adding 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside (final concentration) at A 600 1.0. One hour after induction, the cells were collected and lysed by sonication in lysis buffer (50 mM Tris-HCl, pH 8.0, 2 mM MgCl 2 , 2.0 mM dithiothreitol, 10% glycerol, and 1 mM phenylmethylsulfonyl fluoride). The lysate was centrifuged at 10,000 ϫ g, and the supernatant was used for purification of the GST-C21 protein by affinity purification using glutathione-Sepharose beads (Amersham Biosciences). Loaded beads were washed twice with GST-fishing buffer (50 mM Tris/HCl, pH 8.0, 100 mM NaCl, 2 mM MgCl 2 , 1% Nonidet P-40, 10% glycerol) at 4°C.
GST-C21 Pull-down Experiments-After drug treatment, cells were washed twice with phosphate-buffered saline. After addition of 250 l of ice-cold GST-Fish lysis buffer (50 mM Tris, pH 7.4, 100 mM NaCl, 2 mM MgCl 2 , 10% glycerol, 1% (v/v) Nonidet P-40, and 25 g/ml aprotinin), the cells were harvested. The detergent-soluble supernatant was recovered after centrifugation for 15 min at 14,000 ϫ g and 4°C. GTP-Rho was precipitated for 1 h with 20 l of GST-C21 fusion protein at 4°C. The complex was washed three times with ice-cold phosphate-buffered saline, resuspended, and boiled with Laemmli buffer. Bound RhoA protein was detected by Western blotting using anti-RhoA antibodies (Santa Cruz Biotechnology, Santa Cruz, CA).
Bacterial Toxins-The C2II binding component from Clostridium botulinum C2 toxin and the C2IN-C3 fusion toxin C3 toxin (C3FT) were gifts from Dr. H. Barth (51). C2II binds to the cell membrane and mediates the entry into the cell of the C3FT, which then selectively inactivates the small GTPase RhoA. CNFy toxin from Yersinia pseudotuberculosis was prepared as described previously (43).
Data Collection and Statistical Evaluation-Experiments were repeated at least once. For experiments on BrdUrd uptake, at least 3 wells of a 6-well plate were used per group. BrdUrd as well as DAPI positive cells were counted in an area of 0.1 mm 2 with use of a Zeiss Axiophot microscope. Five or more areas were counted per well. Statistical comparisons were made by Kruskal Wallis' test followed by Mann-Whitney U test. The mean value of the control areas was used to express all groups as percent of controls. Mean Ϯ S.E. (n) are given; (n) signifies the number of areas counted.

PACAP38 Reduces the Nuclear Uptake of BrdUrd in Neocor-
tical Astroglial Cells-In the presence of FCS (10%) in the incubation medium, astroglial cells passed through the G 1 phase and entered the S phase, as indicated by nuclear BrdUrd immunostaining. When either PACAP38 or VIP was present in the incubation medium during the 16 h of incubation, the peptides decreased the nuclear uptake of BrdUrd in a concentration-dependent manner, with EC 50 values of ϳ0.2 and 50 nM, respectively (Fig. 1A). The lower potency of VIP indicated that it did not act via VPAC1 and VPAC2 receptors. To test whether PAC1 receptors were involved, we used the PAC1 receptor-specific antagonist PACAP-(6 -38), which blocks adenylyl cyclase activation by PACAP38 at a concentration of 1.5 nM (44). When used at a concentration of 100 nM, PACAP-(6 -38) had no effect of its own on nuclear BrdUrd uptake. Despite the high concentration, it did not block the inhibitory effect of 0.5 nM PACAP38, indicating that PAC1 receptors coupled to adenylyl cyclase were not involved (Table I).
To test the long-term effects of PACAP38, we added the neuropeptide to free-cycling cells for 12, 24, or 48 h. Compared with controls, PACAP38 (10 nM) reduced the subsequent nuclear immunostaining of BrdUrd by Ն72% independent of the duration of its presence (Fig. 1B). Because PACAP38 was active over 2 days, we tested whether it reduced the total cell number. When applied for 48 h, PACAP38 reduced the total cell number by 18.5% (Fig. 1C). To exclude that this decrease was because of PACAP38-induced cell death, we used propidium iodide (5 M; added for 20 min) to stain dead cells. Treatment with 10 nM PACAP38 for 16 h did not significantly increase the number of propidium iodide positive cells (controls 1.7 Ϯ 0.33%; PACAP38 2.8 Ϯ 0.52% (p Ͼ 0.05; n ϭ 15). We concluded that PACAP38 indeed exerted an antiproliferative effect.
PACAP38 Inhibits Growth Factor-induced BrdUrd Uptake-The EGF and bFGF enhance proliferation in the absence of fetal serum constituents. So does tetradecanoylphorbol acetate (TPA), which activates protein kinase C. We examined whether PACAP38 changed the effects of EGF (5 nM), bFGF (100 nM), or TPA (100 nM) on the nuclear uptake of BrdUrd. Glial cells were incubated in FCS-free incubation medium for 24 h in the absence or presence of the growth factors. Compared with controls, TPA, EGF, and bFGF enhanced the number of BrdUrd positive cells by 106, 62, and 65%, respectively. PACAP38 (10 nM) reduced the effects of the 3 different stimuli by ϳ64% (Fig.  1D), suggesting that it did not act on the receptors of the growth factors but on a signal transduction pathway shared by the 3 stimuli.
PACAP38 Can Reduce Proliferation Independent of PKA-In neuroblasts, PACAP38 reduces proliferation by activating adenylyl cyclase and thus PKA (4). To study whether PACAP38 decreased astroglial proliferation via this pathway, we used the PKA inhibitor protein PKI (45), which was expressed in glial cells using an adenoviral mediated expression system. After infection with either Ad.EGFP or Ad.PKI/EGFP, 91.5 Ϯ 1.3 and 93.6 Ϯ 2%, respectively, of the glial cells showed green fluorescence. Compared with untreated cells, expression of EGFP alone did not change nuclear BrdUrd uptake in synchronized FCS-stimulated astroglial cells ( Fig. 2A). Also the inhibitory effect of PACAP38 (10 nM) was not changed by EGFP expression ( Fig. 2A). Compared with cells expressing only EGFP, those expressing PKI/EGFP showed an unchanged nuclear uptake of BrdUrd, indicating that active PKA was not necessary for the FCS-induced proliferation ( Fig. 2A). Expression of PKI/ EGFP did not prevent but only slightly reduced (by 13.3%; p ϭ 0.086) the inhibitory effect of PACAP38 on nuclear BrdUrd uptake ( Fig. 2A). To confirm that PKI expression blocked PKA activity, we studied the effect of the adenylyl cyclase activator forskolin on proliferation. In EGFP expressing cells, forskolin (10 M) reduced the nuclear uptake of BrdUrd by 76.8%. This effect was indeed no longer present in cells expressing PKI (Fig. 2B). These findings suggested that PACAP38 reduced glial proliferation mainly by an action independent of PKA.
To test this hypothesis, we simultaneously applied FSK and PACAP38. The strong reduction in FCS-induced BrdUrd uptake caused by forskolin was indeed further enhanced by additional application of PACAP38 (Fig. 2C). Finally, we used the pharmacological PKA inhibitor H89. It neither changed FCSstimulated nuclear BrdUrd uptake nor its reduction caused by PACAP38 (Fig. 2D). Taken together, these data indicated that PACAP38 exerted an antiproliferative effect, which was independent of PKA, although active PKA was able to reduce glial proliferation.
PACAP38 Does Not Reduce Proliferation by an Action on ERKs-Kinases of the MAP kinase family, i.e. ERK1 and ERK2, are essential for G 1 progression and entry of the S phase. First, we confirmed the role of the ERKs in astroglial proliferation. We used U0126 to inhibit the dual-specificity kinases MEK1/2, which prevents the activation of the ERKs (46). U0126 (10 M) indeed reduced the FCS-stimulated nuclear BrdUrd uptake of synchronized cells by 83.5% (Fig. 3A). When used alone, PACAP38 (10 nM) reduced the BrdUrd uptake by 66.2%. Together with U0126, however, PACAP38 reduced the nuclear BrdUrd uptake by 93.6%. This significant additional effect suggested that U0126 and PACAP38 acted via different mechanisms. Next, we examined whether PACAP38 (10 nM) reduced the activity of the ERKs using an immunoprecipitation assay. We added PACAP38 (10 nM) for up to 4 h to the FCS-containing incubation medium of astroglial cells. It slightly reduced ERK activity only after 1 h, but had no effect at later times (Fig. 3B). In a third approach, we investigated the activity of ETS2 repressor factor (ERF), which is phosphorylated by ERK2 (41,(47)(48)(49). When localized in the nucleus, ERF can block access of the transcription factor ETS2 to its binding site. Upon phosphorylation at multiple sites by ERK2 within the nucleus, ERF translocates to the cytosol and thus allows the transcription necessary for the initiation of the proliferative cycle (41,47,49). Astroglial cells were transiently transfected with a plasmid coding for an EGFP/ERF fusion protein. Two days later, EGFP immunofluorescence showed the cellular localization of the transcription factor. In cells incubated with FCS-containing medium, ERF was mainly located in the cytosol. Only a few cells showed a nuclear localization of ERF (Fig. 3C). One hour after serum deprivation, we found ERF mainly in the cell nucleus (Fig. 3C), confirming previous findings (41). When astroglial cells in FCS-containing medium were treated for 1 h with PACAP38 (10 nM), most cells still showed a cytosolic localization of ERF (Fig. 3C). Apparently, PACAP38 did not decrease the activity of ERK2 so that ERF re-entered the nucleus. Taken together, our results provided no evidence that PACAP38 reduced ERK activity.
The Small GTPase RhoA Is Involved in the Antiproliferative Effect of PACAP38 -The small GTPase RhoA supports the G 1 phase progression via 2 mechanisms. It can contribute to the sustained increase in ERKs activity necessary for cyclin D/CDK4/6 activation as well as prevent the inhibitory effects of CKIs p21 Cip1 and p27 Kip1 on the cyclin E-CDK2 complex (29,31,33,35). Because morphological observations indicated that PACAP38 may affect RhoA activity, we examined the role of RhoA.
In FCS-containing medium, proliferating astroglial cells are widely spread and show numerous stress fibers, which depend on RhoA activity. PACAP38 had a pronounced effect on the morphology of most glial cells. Within 1 h, the polygonal cells started to loose their stress fibers. After 4 h the stress fibers were completely destroyed (Fig. 4), suggesting that RhoA was inactive. C3 toxin of Clostridium limosum, which ADP-ribosylates and inactivates RhoA but not Rac1 or Cdc42 (for review, see Aktories et al. (50)) also destroyed the stress fibers in most cells (Fig. 4), when used together with a carrier system that facilitates permeation of the cell membrane (51). In addition, C3 toxin (200 ng/ml) diminished nuclear BrdUrd uptake by 63% (Table II). In contrast to treatments with forskolin or U0126, PACAP38 had no additional effect, when applied together with C3 toxin (Table II). This finding suggested that PACAP38 induced proliferation by an action on RhoA.
To study whether RhoA supported nuclear BrdUrd uptake by organizing the actin cytoskeleton, we destroyed the latter with cytochalasin D, i.e. without disturbing Rho activity (1 M; Fig.  4). Cytochalasin D decreased by 84.1% the nuclear uptake of BrdUrd (Table II). Simultaneous application of PACAP38, however, further decreased BrdUrd uptake. This finding suggested that PACAP38 did not reduce proliferation by destroying the actin cytoskeleton.

PACAP38, RhoA, and Astroglial Cell Cycle
To further study whether active RhoA was necessary for astroglial BrdUrd uptake, we expressed dominant negative RhoA in glial cells using an adenovirus vector. dnRhoA scavenges the guanine-nucleotide exchange factors (GEFs) necessary for the activation of endogenous RhoA (26). In cells infected with Ad.EGFP or Ad.dnRhoA, 91.7 Ϯ 1.4 and 94.7 Ϯ 0.5%, respectively, of the astroglial cells were infected as visualized by the co-expressed EGFP protein. In such cells, stress fibers were no longer observed, indicating that endogenous Rho was indeed inactive (Fig. 5A). dnRhoA reduced the nuclear uptake of BrdUrd by 51.5%, i.e. to an extent similar to the effect of PACAP38 alone. When the neuropeptide was added to such cells, it had no additional effect (Fig. 5D). Taken together, these results indicated that inactivation of RhoA and PACAP38 had similar and non-additive effects on BrdUrd uptake.
PACAP38 Reduces BrdUrd Uptake by Acting on RhoA-To examine whether PACAP38 reduced the activity of RhoA, we used an affinity precipitation assay. The C21 domain of the RhoA effector Rhotekin was used to precipitate the active GTPase, because it interacts only with the GTP bound, active form of RhoA. Compared with controls, treatment with PACAP38 (10 nM) strongly reduced the amount of RhoA bound to the Rhotekin beads (Fig. 5B), whereas expression of dnRhoA abolished the binding of RhoA to Rhotekin (Fig. 5C).
To find out whether enhanced activity of RhoA prevented the reduction in nuclear BrdUrd uptake caused by PACAP38, we   FIG. 2. Studies on  PACAP38, RhoA, and Astroglial Cell Cycle infected glial cells with caRhoA. Like V12Ras V14RhoA no longer has endogenous GTPase activity (52,53). Once it binds GTP after its synthesis, it remains active. Of the cells infected with Ad.EGFP or Ad.caRhoA, 91.5 Ϯ 2.1 and 94.2 Ϯ 0.2%, respectively, expressed EGFP. Determination of GTP binding by caRhoA with the Rhotekin affinity precipitation assay indeed showed that PACAP38 did not reduce GTP binding (Fig.  6A). Moreover, in cells expressing caRhoA PACAP38 was no longer able to reduce nuclear uptake of BrdUrd (Fig. 6B). Also stress fibers produced by caRhoA were not destroyed by PACAP38 (data not shown).
PACAP38 Reduces RhoA Activity in Astroglial Cells Treated with CNFy-CNFy is a cytotoxic necrotizing factor from Y. pseudotuberculosis. It deamidates glutamine 63 of RhoA and thus inhibits the GTP hydrolyzing activity of the latter (54). When used alone, CNFy strongly enhanced GTP binding of RhoA as measured by the Rhotekin affinity precipitation assay (Fig. 6C). However, PACAP38 applied simultaneously was still able to diminish GTP binding, suggesting that it decreased the activation of RhoA by a GEF. The PKA inhibitor H89 did not prevent this inhibitory effect indicating that PKA was not involved in this action (Fig. 6C). The persistent inhibitory effect of PACAP38 on RhoA activity was also observed in a morphological study. PACAP38 destroyed stress fibers in cells treated with CNFy (Fig. 6D). Because CNFy alone reduced nuclear BrdUrd in FCS-stimulated synchronized cells, the effect of PACAP38 could not be evaluated under these conditions (data not shown).

DISCUSSION
There is increasing evidence that the neuropeptide PACAP38 acts as a growth factor that regulates the proliferation and differentiation of neuronal and glial cells in the neocortex and other brain areas (see Introduction). In the present study, PACAP38 reduced the number of cultured neocortical astroglial cells that incorporated BrdUrd during the S phase of the proliferative cycle. This effect was independent of the cAMP/PKA signaling pathway. Also an action on the ERKs of the MAP kinase family could not be shown. Instead, we obtained evidence that PACAP38 inhibited the small GTPase RhoA that supports progression through the G 1 phase into the S phase.
Whereas PACAP38 reduced nuclear BrdUrd uptake at subnanomolar concentrations, VIP was 250 times less potent, suggesting that PAC1 but not VPAC1 or VPAC2 receptors were involved, because the latter are stimulated by similar, low nanomolar concentrations of PACAP38 and VIP (2, 3). PAC1 receptors are indeed expressed at low density by neocortical astroglial cells (11,(55)(56)(57). PAC1 receptor mRNA was also found in our cells with use of reverse transcriptase-PCR (data not shown). However, our further results did not confirm an exclusive role of PAC1 receptors. Thus, PAC1 receptors coupled to adenylyl cyclase or phospholipase A are stimulated only by micromolar concentrations of VIP (2, 3), whereas in our cultures VIP was active at a concentration of 10 nM. Moreover, the competitive antagonist PACAP (6 -38), which blocks PAC1 receptor-mediated cAMP production at low nanomolar concentrations (44), did not prevent the inhibitory effect of PACAP38 on nuclear uptake of BrdUrd. By stimulating PAC1 or VPAC1/2 receptors, nanomolar concentrations of PACAP38 or VIP can also activate phospholipase D via an ARF-dependent mechanism, which is blocked by brefeldin A (58). In preliminary experiments, however, brefeldin A abolished astroglial proliferation (data not shown). Because this result confirmed previous findings (59), the involvement of ARF was not further tested. Thus, the second messenger involved in the antiproliferative effect of the neuropeptides remained unclear.
In neuronal precursors, the short isoform of PAC1 receptors has been suggested to mediate the antimitogenic effects of PACAP via production of cAMP and activation of PKA (12,16). In astroglial cells, however, the contribution of PKA was rather  Fig. 1A. a, indicates significant difference (p Ͻ 0.05) to control group; b, indicates significant difference (p Ͻ 0.05) to groups treated with U0126 or PACAP38 alone. Panel B, cells in serum-containing incubation medium were treated with PACAP38 (10 nM) for 1, 2, and 4 h, respectively; controls (Co) remained untreated for 2 h. Cells were harvested and 300 g of proteins were immunoprecipitated with an immobilized phospho-p44/42 MAP kinase antibody. The immunoprecipitates were assayed for MAP kinase activity with use of GST-Elk-1. The positions of different phosphorylation states of GST-Elk-1 are indicated as Phospho-Elk-1. The experiment was repeated twice. Panel C, PACAP38 (10 nM) does not reduce the intracellular localization of ERF. Confocal laser micrographs show localization of ERF immunoreactivity in the cytosol of control cells and in the nucleus of astroglial cells serum deprived for 1 h. A PACAP38-treated cell is also shown. Nuclear and cytosolic distribution of ERF is quantified in the lower panel. Cells that contained ERF predominantly in the nucleus (open columns) or the cytosol (filled columns) are shown as percent of the total number; cells that contained similar densities of ERF in the nucleus and the cytosol were also counted but are not shown. Mean Ϯ S.E. (n ϭ 4). a, shows significant difference (p Ͻ 0.05) to respective controls in serum containing medium.
small to the antiproliferative effect of PACAP38. When expressed in the astroglial cells using an adenovirus system, PKI, a peptide inhibitor of PKA (45), did not have a significant effect on the inhibition of the nuclear uptake of BrdUrd caused by PACAP38. Also the pharmacological PKA inhibitor H89 was without effect. In contrast, PKI prevented the strong inhibition of the nuclear BrdUrd uptake caused by the adenylyl cyclase stimulator forskolin. This finding confirmed that PKI was expressed by the adenoviral system and efficiently blocked the activity of PKA. It also showed that active PKA had an antiproliferative effect in astroglial cells, as it has in neuroblasts (4,12). However, PACAP38 receptors do not seem to be efficiently coupled to this signal transduction pathway.
ERKs mediate the G 0 /G 1 phase transition and the G 1 phase progression upon activation by growth factors such as EGF and bFGF (60,61). When we prevented the activation of the ERKs with the specific MEK1/2 inhibitor U0126, the nuclear uptake of BrdUrd in the astroglial cultures was reduced by 83.5%, indicating that ERKs were indeed essential for astroglial proliferation. However, our additional findings disagreed with the hypothesis that PACAP38 reduced astroglial proliferation by inactivating the ERKs. Thus, PACAP38 further decreased by Ͼ50% the nuclear uptake of BrdUrd already reduced by U0126, suggesting an additional and independent effect. Moreover, we obtained no evidence that PACAP38 reduced the activity of the ERKs. When tested with an immunoprecipitation assay, PACAP38 did not significantly reduce the activity of the ERKs in serum-treated cells. According to previous findings, PACAP38 may even activate these kinases in neurons and in serum-deprived astroglial cells (36,62). Finally, we studied the transcriptional repressor ERF, which translocates from the nucleus to the cytosol, when ERKs are active (41,49). PACAP38 did not affect its cytosolic location, confirming that the neuropeptide did not decrease the ERK activity.
The morphological effects of PACAP38 suggested that the peptide acted on the small GTPase RhoA, which affects G 1 progression not only by organizing cell attachment but also by inhibiting p27 Kip1 (18). Three lines of evidence were in agreement with the hypothesis that PACAP38 exerted its antiproliferative effect in astroglial cells by reducing the activity of RhoA. First, PACAP38 reduced the BrdUrd uptake to the same extent as C3 fusion toxin, which selectively inactivates RhoA but not the GTPases Rac1 or Cdc42 (50). PACAP38 applied together with the C3 fusion toxin caused no further reduction in nuclear BrdUrd uptake. Also dnRhoA expressed in the astroglial cells decreased the nuclear uptake of BrdUrd. This finding was in agreement with the inhibitory effect of dnRhoA on platelet-derived growth factor-induced proliferation in IIC9 cells (34,63). In dnRhoA expressing cells, PACAP38 did not further inhibit BrdUrd uptake, indicating a common mechanism of action. It is noteworthy that the expression of dnRhoA caused the disappearance of stress fibers without changing  astroglial morphology. This finding indicates that inactivation of endogenous RhoA reduced proliferation independent of changes in cell morphology. Second, PACAP38 strongly reduced the cellular content of active RhoA extracted with the Rhotekin pull-down assay. The destruction of stress fibers caused by PACAP38 in the astroglial cells provides additional evidence for the inactivation of RhoA. Finally, caRhoA expressed in the astroglial cells prevented the inhibitory effect of PACAP38 on BrdUrd uptake. In primary cultures, the sustained activation of ERKs and cyclin D1 necessary for cell cycle progression is only possible if the cells attach to the matrix (21)(22)(23)(24)(25). Cell attachment is mediated by the actin cytoskeleton that is organized by the Rho GTPases RhoA, Rac, and Cdc42. Although PACAP38 caused pronounced changes in the astroglial morphology, our results did not show that the peptide reduced nuclear BrdUrd uptake via this mechanism. Thus, destruction of the actin cytoskeleton by cytochalasin D also reduced by 84.1% the nuclear uptake of BrdUrd, but PACAP38 applied together with cytochalasin D even further reduced BrdUrd uptake.
RhoA has been shown to decrease the content in p27 Kip1 (31). It does so by activating the cyclin E-CDK2 complex, which then induces the degradation of p27 Kip1 (34). In this context, it is noteworthy that p27 Kip1 , but not p21 Cip1 , is strongly expressed in rat neocortical tissue during the late pre-and early postnatal period (64), when astroglial cells proliferate. However, a reduction in cyclin D1 protein content has been also observed that was induced by RhoA inactivation with C3 toxin (65).
By using the toxin CNFy, we obtained some information how PACAP38 inactivated RhoA. The bacterial toxin deamidates glutamine 63 of RhoA and thus inhibits the latters GTP hydrolyzing activity. Consequently, the exchange of GDP to GTP mediated by a cellular GEF leads to an irreversible activation of RhoA (43). Simultaneous treatment of the cells with PACAP38 and CNFy diminished the RhoA activity enhanced by CNFy, suggesting that PACAP38 interfered with the activation of RhoA.
More than 60 mammalian GEF proteins are already known, several of which specifically activate Rho. The subcellular localization of a GEF protein seems to be important for its activity and may regulate the effect of the activated Rho GTPase (66). The GEFs expressed in astroglial cells are not known so that we can only speculate on the mechanisms of their regulation. Of the signal transduction pathways initiated by PACAP38, only the cAMP-PKA cascade has been linked to Rho activation. AKAP-Lbc is a PKA anchoring protein with an additional Rho GEF domain (67). Upon binding of active PKA, the GEF activity is reduced (68,69). However, activation of PKA was not responsible for the antimitogenic effect of PACAP38 in our astroglial cells. Stimulation of PACAP receptors may also activate phospholipase C and thus increase intracellular free Ca 2ϩ levels as well as protein kinase C activity (3). However, PACAP38 decreased the nuclear BrdUrd uptake induced by activation of protein kinase C with TPA, indicating that this kinase was not involved in the inhibitory effect of PACAP38 on Rho. Whether PACAP38-induced changes in intracellular Ca 2ϩ concentrations affected the activation of Rho remained open.
High concentrations of RhoA encoding mRNAs are present in neocortical growth zones during the embryonic and early postnatal period (70). The present data suggest that the GTPase plays a role in the regulation of proliferation of astroglial cells. Which extra-or intracellular factors maintain the activation of RhoA and thus cell proliferation is an important question to be studied in future experiments. In addition, we have found a link between the neuropeptide growth factor PACAP38 and RhoA, which seems to be specific for astroglial cells. In cultured neocortical neuroblasts, PACAP38 inhibits proliferation by activating PKA and the CKI p57 Kip2 . It decreases CDK2 kinase activity (64). In cerebellar granule cell precursors, it inhibits the proliferation induced by the Soni-Hedgehog pathway (71). In cultured astroglial cells, however, PACAP38 does not act via cAMP and PKA, but reduces the activity of RhoA, which may lead to the disinhibition of the CKI p27 Kip1 .