PAC1 Receptor Activation by PACAP-38 Mediates Ca2+ Release from a cAMP-dependent Pool in Human Fetal Adrenal Gland Chromaffin Cells*

Previous studies have shown that human fetal adrenal gland from 17- to 20-week-old fetuses expressed pituitary adenylate cyclase-activating polypeptide (PACAP) receptors, which were localized on chromaffin cells. The aim of the present study was to identify PACAP receptor isoforms and to determine whether PACAP can affect intracellular calcium concentration ([Ca2+] i ) and catecholamine secretion. Using primary cultures and specific stimulation of chromaffin cells, we demonstrate that PACAP-38 induced an increase in [Ca2+] i that was blocked by PACAP (6–38), was independent of external Ca2+, and originated from thapsigargin-insensitive internal stores. The PACAP-triggered Ca2+ increase was not affected by inhibition of PLCβ (preincubation with U-73122) or by pretreatment of cells with Xestospongin C, indicating that the inositol 1,4,5-triphosphate-sensitive stores were not mobilized. However, forskolin (FSK), which raises cytosolic cAMP, induced an increase in Ca2+ similar to that recorded with PACAP-38. Blockage of PKA by H-89 or (R p)-cAMPS suppressed both PACAP-38 and FSK calcium responses. The effect of PACAP-38 was also abolished by emptying the caffeine/ryanodine-sensitive Ca2+ stores. Furthermore, treatment of cells with orthovanadate (100 μm) impaired Ca2+reloading of PACAP-sensitive stores indicating that PACAP-38 can mobilize Ca2+ from secretory vesicles. Moreover, PACAP induced catecholamine secretion by chromaffin cells. It is concluded that PACAP-38, through the PAC1 receptor, acts as a neurotransmitter in human fetal chromaffin cells inducing catecholamine secretion, through nonclassical, recently described, ryanodine/caffeine-sensitive pools, involving a cAMP- and PKA-dependent phosphorylation mechanism.

Pituitary adenylate cyclase-activating polypeptide is a 38residue ␣-amidated neuropeptide (PACAP-38) 1 originally isolated from the ovine hypothalamus for its ability to stimulate cAMP formation in rat anterior pituitary cells. Processing of PACAP-38 can generate a 27-amino acid amidated peptide (PACAP-27) that exhibits 68% sequence identity with vasoactive intestinal polypeptide (VIP), thus identifying PACAP as a member of the VIP/secretin/glucagon superfamily of regulatory peptides (1,2).
The effects of PACAP are mediated through interaction with two types of high affinity receptors: type I receptors are selectively activated by PACAP, whereas type II receptors bind PACAP and VIP with similar affinity (3). Three isoforms of PACAP receptors have now been cloned and designated as PACAP-specific receptor I (PAC 1 -R) (4,5) and VIP/PACAP mutual receptors 1 and 2 (VPAC 1 -R and VPAC 2 -R) (6,7). Both PAC 1 -R (type 1 receptors) and VPAC 1 -R/VPAC 2 -R (type 2 receptors) belong to the seven-transmembrane domain, G-protein-coupled receptor family, and are all positively coupled to adenylyl cyclase (2). Eight isoforms of PAC 1 -R, resulting from alternative splicing, have been characterized to date. These variants display differential signal transduction properties with regard to adenylyl cyclase and phospholipase C (PLC) stimulation (1,2). In addition to these classical signaling pathways, PACAP has been found to stimulate a Ca 2ϩ -calmodulin nitric oxide synthase (8) and mitogen-activated protein kinase activity (9). These various transduction mechanisms are involved in the neurotrophic activities exerted by PACAP (i.e. inhibition of apoptosis and stimulation of neurite outgrowth) during development (9 -11).
PACAP and its receptors are actively expressed in the adrenal medulla (12)(13)(14). In particular, we have previously demonstrated the occurrence of PACAP-38 (15) and PACAP binding sites (16) in chromaffin cells from 16-to 20-week-old fetal human adrenal glands. Activation of these receptors by PACAP-38 causes stimulation of cAMP production and induces a modest increase in inositol 1,4,5-triphosphate (IP 3 ) formation (16), suggesting a role for the neuropeptide in the developing adrenal gland. During the process of adrenal gland development, pheochromoblasts originating from the neural crest migrate throughout the fetal cortex, acquiring progressive differentiation through contact with the steroidogenic cells (17,18) (for review see Ref. 19). However, the neuroendocrine regulation of catecholamine release by chromaffin cells in the human fetus has not been investigated.
Although PACAP is known to be a potent activator of catecholamine secretion from rat and porcine adrenochromaffin cells (20,21), the differential coupling of PACAP receptor variants to the various Ca 2ϩ sources and to the adenylyl cyclase and PLC signaling pathways is still poorly understood. In particular, the effect of PACAP on fetal chromaffin cells has never been investigated. The aim of the present study was therefore to identify the PACAP receptor isoforms expressed in the human fetal adrenal gland and to analyze the signaling pathways responsible for PACAP-evoked [Ca 2ϩ ] i increase and catecholamine secretion.
Retrieval and Preparation of Glands-Fetal adrenal glands were obtained from fetuses aged 14 -20 weeks (post fertilization) at the time of therapeutic abortion. Fetal ages were estimated by foot length and time after menstruation, according to Steeter et al. (23). The project was approved by the human subject review committee of our institution. After retrieval, glands were cleansed of fat and processed immediately for cellular or RNA preparation.
PCR Amplifications-cDNA samples (2 l) were used for subsequent PCR amplifications in 50 l of 1ϫ PCR buffer (10 mM Tris, pH 8.3, 50 mM KCl, 1.5 mM MgCl 2 ) containing 200 M deoxy-NTPs, 10 pmol of each of the sense and antisense primers, and 2.5 units of Taq DNA polymerase. Primers used for the amplification of PAC 1 -R splice variant cDNAs were sense, 5Ј-CTTGTGCAGAAACTTCAGTCTCCAGACATG, and antisense, 5Ј-TCGGTGCTTGAAGTCCACAGCGAAGTAACGGTT-CACCTT (25), corresponding to base sequences 1235-1264 and 1499 -1537 of human PAC-R (4). These primers flank the insertion site of 84-bp cassettes (5) and were designed to give a 303-bp amplicon for the normal (null) PACAP receptor, a 387-bp amplicon for the SV-1 or SV-2 splice variants (insertion of one 84-bp cassette), and a 471-bp amplicon for the insertion of two cassettes. PCR was carried out in a PerkinElmer Life Sciences (geneAmp PCR System 2400) thermocycler at 94°C for 2 min followed by 35 cycles at 94°C for 30 s, 61°C for 30 s, 72°C for 1 min, and a final extension at 72°C for 5 min. Primers used for the amplification of VPAC 1 -R cDNA were the sense, 5Ј-ATGTGCA-GATGATCGAGGTG, and antisense, 5Ј-TGTAGCCGGTCTTCACAGAA (26), corresponding, respectively, to base sequences 127-146 and 431-450 of human VPAC 1 -R (6). PCR was carried out at 94°C for 2 min followed by 30 cycles at 94°C for 30 s, 55°C for 30 s, 72°C for 1 min, and a final extension at 72°C for 5 min, giving a 324-bp amplicon. Primers used for the amplification of VPAC 2 -R cDNA were the sense, 5Ј-TCAAACAGAAAAACACAAAGC, and antisense, 5Ј-ACCTGTTCCT-GTCCTTCATCC (7), corresponding, respectively, to base sequences 294 -314 and 653-673 of human VPAC 2 -R (7). PCR was carried out at 95°C for 2 min, followed by 35 cycles at 95°C for 1 min, 58°C for 1 min, 72°C for 1 min, and a final extension at 72°C for 5 min, giving a 380-bp amplicon.
In every PCR experiment, amplification in the absence of cDNA and in the presence of 2 g of RNA was performed as a control (not shown). Negative controls using cDNA templates from NCI-H295R and Chinese hamster ovary cells were performed (not shown). PCR products (10 l) were analyzed on 2% (w/v) agarose gel and visualized by ethidium bromide staining. Length of PCR products was estimated using a 100-bp DNA ladder. Identification of the PCR products was confirmed by enzymatic digestion and electrophoresis on 3% (w/v) agarose with a 25-bp DNA ladder (not shown). Experiments were performed with RNA isolated from three different adrenal glands from 14-, 17-, and 20-weekold human fetuses. Three experiments were performed for each age.
Cell Culture-Glands were processed as described previously (16). Whole tissues from one or two glands were used for each cell preparation, without separation of fetal zone, neocortex, or chromaffin cells. Briefly, small portions of glands (1-2 mm 3 ) were dissociated with collagenase (2 mg/ml) and DNase (25 g/ml) in Eagle's minimal essential medium containing 2% antibiotics. After three 20-min incubations, cells were dissociated, filtered, and centrifuged for 10 min at 100 ϫ g. The cell pellet was suspended in OPTI-MEM medium containing 2% fetal calf serum, antibiotics, and antimycotics. Cells were plated at a density of ϳ2 ϫ 10 5 on plastic coverslips (25 mm). Cells were grown for 3 days in a humidified atmosphere of 95% air/5% CO 2 , at 37°C.
Calcium Measurement-For dye loading, cells were incubated for 30 min at 37°C in the physiological medium OPTI-MEM containing 4 M of the fluorescent calcium indicator Fluo-4/AM. Hydrolysis was performed for 30 min at 37°C in a medium containing: 140 mM NaCl; 5.4 mM KCl; 2 mM CaCl 2 ; 1 mM MgCl 2 ; 10 mM HEPES, pH 7.4; and 1 g/liter glucose. The coverslips were then mounted on the stage of an inverted microscope (Nikon Diaphot, Mississauga, Ontario, Canada). The light source was generated by a 100-watt mercury lamp. Band-pass filters (450 -490 nm) and (520 -560 nm) were used for excitation and emission, respectively. The emitted light was recorded by a photon-counting unit. Calcium calibration was performed as described previously (27). However, because calcium dye properties may be different in cytoplasmic and nucleoplasmic compartments (28), [Ca 2ϩ ] i should be considered only as semiquantitative data (29).
Catecholamine Secretion-An amperometric technique was used to record secretion from chromaffin cells as described previously (30). Briefly, a 5-m carbon fiber electrode (Ala Scientific Instrument Inc., Westbury, NY) connected to a Patch Clamp PC-501A amplifier (Warner Instrument Corp., Hamden, CT) modified for voltametry was positioned in close proximity of a cell. The potential of the electrode was fixed to 800 mV, and current traces were filtered at 4 kHz and recorded on a DAS-75 digital recorder (Dagan Corp., Minneapolis, MN).
Data Analysis-Curves were fitted with SigmaPlot (version 7.0, Chicago, IL). The data are presented as means Ϯ S.E. from the number of experiments indicated in the legends or in the text. Statistical analyses of the data were performed using the one-way analysis of variance test. Homogeneity of variance was assessed by Bartlett's test, and p values were obtained from Dunnett's tables.

RESULTS
Molecular Identification of PACAP Receptors-Using autoradiography, we have previously shown that PACAP receptors in the human fetal adrenal gland were localized only on chromaffin cells (16). An RT-PCR approach was used to identify the PACAP receptor subtypes expressed in adrenal glands from 14to 20-week-old fetuses. As indicated under "Materials and Methods," PAC 1 -R splice variants were discriminated using primers flanking the insertion site of the 84-bp cassettes. As shown in Fig. 1A, two bands were detected: a 303-bp band corresponding to the short form of the receptor and a 387-bp band corresponding to an isoform containing a single insertion cassette. Amplicons of 324 and 380 bp corresponding, respectively, to VPAC 1 -R and VPAC 2 -R sequences were also detected in all fetal adrenal glands. However, based on our previous results, these receptors have a more diffuse distribution throughout the adrenal gland, whereas PAC 1 -Rs are only detected in chromaffin cells (16). Expression of the PAC 1 -R, VPAC 1 -R, and VPAC 2 -R RNAs was also observed in 18-weekold fetal brains. Amplification of glyceraldehyde-3-phosphate dehydrogenase ( Fig. 1D) confirmed that cDNA was present in each reaction.
Effect of PACAP on Cytosolic Ca 2ϩ Concentration-Chromaffin cells, which gather into small clusters with numerous processes, could be easily distinguished from the large individual fetal steroidogenic cells (Fig. 2, white arrows). Measurements of [Ca 2ϩ ] i were performed on peripheral cells of the clusters, selected by a pinhole placed in the optical path (Fig. 2, asterisk). Mean [Ca 2ϩ ] i in chromaffin cells incubated in medium containing 2 mM Ca 2ϩ was 77 nM Ϯ 7 (n ϭ 6) in resting conditions. Application of 1 ϫ 10 Ϫ9 M PACAP-38 induced a transient elevation of [Ca 2ϩ ] i (Fig. 3A). This [Ca 2ϩ ] i increase was characterized by a rapid upstroke followed by a slower decrease to basal levels. The time course of the decrease could be fitted by a monoexponential function. At a concentration of 1 ϫ 10 Ϫ9 M PACAP-38, the time constant was 57.3 Ϯ 7.4 s (n ϭ 9). In Ca 2ϩ -free medium (0 Ca 2ϩ plus 1 mM EGTA), application of 1 ϫ 10 Ϫ9 M PACAP-38 provoked an increase in [Ca 2ϩ ] i similar to that recorded in 2 mM Ca 2ϩ -containing medium (Fig. 3B) with a time constant (49.7 Ϯ 3.5 s; n ϭ 11) that was not significantly different (p ϭ 0.34), suggesting that PACAP-38 causes mobilization of intracellular Ca 2ϩ stores. Indeed, application of Ni 2ϩ , a known Ca 2ϩ channel blocker, during the decreasing phase of the Ca 2ϩ response, had no effect on the kinetics of the Ca 2ϩ spike induced by application of PACAP-38 (1 ϫ 10 Ϫ9 M) in 2 mM Ca 2ϩ -containing medium, whereas a second application of PACAP-38 in the presence of Ni 2ϩ provoked a Ca 2ϩ response similar to that obtained in the absence of the blocker (Fig. 3C). Application of increasing concentrations of PACAP-38 resulted in a dose-dependent rise in the amplitude of the [Ca 2ϩ ] i response with an ED 50 value of 5 nM (Fig. 3D), a value similar to that obtained (2.  even when applied at very short intervals (Fig. 4A). The amplitudes of the responses to any of the four pulses applied were not significantly different (Fig. 4B). To further confirm that the PACAP-induced [Ca 2ϩ ] i increase was due to the activation of the PAC 1 -R type receptor, we used PACAP (6 -38), a PAC 1 receptor-specific antagonist shown to inhibit the activation of adenylyl cyclase with a K i of 7 nM (32). A first application of PACAP-38 (1 ϫ 10 Ϫ7 M) triggered a Ca 2ϩ response as previously described. The antagonist PACAP (6 -38) was then applied (10 M) for a 10-min period and followed by a second application of PACAP-38, at the same concentration, which induced a lower Ca 2ϩ increase (Fig. 4C). For concentrations of PACAP (6 -38) of 0.3, 1, and 10 M, the Ca 2ϩ response to PACAP-38 (1 ϫ 10 Ϫ7 M) was reduced to 48.5 Ϯ 12% (n ϭ 4), 25 Ϯ 2% (n ϭ 2), and 6.9 Ϯ 3.4% (n ϭ 4) of the control, respectively (Fig. 4D).
Nature of the Intracellular Ca 2ϩ Pool-We have previously shown that PACAP-38 induces a 3.4-fold increase in cAMP production and a modest increase in IP 3 formation in fetal human chromaffin cells (16). In the present study, we first investigated whether or not the PACAP-sensitive Ca 2ϩ pool was responsive to thapsigargin (TG), a known inhibitor of the sarco(endo)plasmic reticulum Ca 2ϩ -ATPase pumps but without effect on plasma membrane Ca 2ϩ -ATPase activity (33). In one series of experiments, the cells were bathed in Ca 2ϩ -free medium. Application of PACAP-38 (1 ϫ 10 Ϫ9 M) produced an increase in Ca 2ϩ as described previously (Fig. 5A). The subsequent application of TG (4 M) triggered an additional increase in [Ca 2ϩ ] i resulting from blockage of the sarco(endo)plasmic reticulum Ca 2ϩ -ATPase pumps. When PACAP-38 was applied further, the amplitude and kinetics of the Ca 2ϩ increase were similar to those obtained prior to TG application, indicating that TG and PACAP-38 did not mobilize the same Ca 2ϩ pool(s). In a second series of experiments, cells were preincubated for 30 min in a calcium-free medium containing 8 M TG. Under these conditions, the Ca 2ϩ response to TG (4 M) was greatly reduced, whereas the response to PACAP-38 (2.5 ϫ 10 Ϫ9 M) was not affected (Fig. 5B).
In some other cell types, it has been shown that PACAP receptors are coupled to PLC through a G q/11 protein to produce diacylglycerol and IP 3 (34). In the specific case of human fetal chromaffin cells, we previously found that IP 3 production is relatively low (16). Hence, experiments were performed to assess the putative role of IP 3 -sensitive calcium pools in the PACAP-induced increase in [Ca 2ϩ ] i . Cells were treated with U-73122 compound, a specific PLC␤ inhibitor (35), at a concentration of 1 M for 18 h. PACAP-38 (1 ϫ 10 Ϫ9 M) was then added in the presence or absence of Ca 2ϩ in the external medium as described above. In five different cells, the amplitude as well as the kinetics of the Ca 2ϩ response to PACAP-38 were similar to those obtained in control cells (Fig. 6A), indicating that the Ca 2ϩ increase was not dependent on IP 3 -sensitive pools. Further confirmation of these results was provided from experiments using Xestospongin C (XeC), a potent specific blocker of the inositol 1,4,5-triphosphate (IP 3 ) receptors (36). In this experimental design, PACAP-38 (1 ϫ 10 Ϫ9 M) was first applied to induce a [Ca 2ϩ ] i increase. XeC (20 M) was then applied for 10 min prior to a second application of PACAP-38. Fig. 6B shows that the amplitude of the Ca 2ϩ increase was not affected by XeC, thus indicating that IP 3 -sensitive Ca 2ϩ pools are not involved. Similar results were obtained from five different cells.
Involvement of a cAMP-sensitive calcium pool in the PACAP response was tested by directly activating cAMP production with forskolin (FSK). Fig. 7A demonstrates that FSK (1 ϫ 10 Ϫ5 M) and PACAP-38 (1 ϫ 10 Ϫ9 M) triggered identical calcium increases. Similar results were obtained in the absence of Ca 2ϩ in the bathing medium (data not shown, n ϭ 3). The involvement of a PKA-dependent phosphorylation step in the increase in Ca 2ϩ triggered by PACAP was assessed by using H-89, an inhibitor of PKA. When cells were pretreated with H-89 (10 M) for 15 min, subsequent stimulation with PACAP-38 failed to trigger a Ca 2ϩ increase (Fig. 7B, n ϭ 7). The FSK-triggered Ca 2ϩ increase was also abolished (Fig. 7B, n ϭ 4). When (R p )-cAMPS (1 mM), a more specific membrane-permeant inhibitor of PKA was used, the response to PACAP-38 (2 ϫ 10 Ϫ7 M), but not to TG, was abolished (Fig. 7C, n ϭ 4).
Caffeine is known to activate ryanodine channels and to induce [Ca 2ϩ ] i increase in numerous cell types (37). Application of caffeine (20 mM) to chromaffin cells induced an increase in [Ca 2ϩ ] i (Fig. 8A). More importantly, the Ca 2ϩ response to PACAP-38 (1 ϫ 10 Ϫ9 M) was significantly decreased following caffeine application to chromaffin cells (Ϫ51 Ϯ 6.7%, n ϭ 5). Similar results were obtained with ryanodine (20 M, n ϭ 4, data not shown).
A recent study reported that neuroendocrine cells contained dense core secretory vesicles that could constitute a dynamic Ca 2ϩ store, whereby the P-type Ca 2ϩ pump was responsible for Ca 2ϩ uptake in these secretory vesicles (38). To assess such putative participation, we used orthovanadate at 100 M, a concentration known to inhibit the ATP-dependent P-type Ca 2ϩ pump. In Ca 2ϩ -free medium, cells were first challenged with PACAP-38 (1 ϫ 10 Ϫ9 M) followed by application of Na 3 VO 4 (100 M) for 10 min. Thereafter, the first addition of PACAP-38 induced a Ca 2ϩ response not significantly different in amplitude (9.2 Ϯ 9% decrease, n ϭ 6) from that obtained prior to adding Na 3 VO 4. However, subsequent applications of PACAP-38 gave rise to a lower amplitude of Ca 2ϩ increase (Ϫ63% and Ϫ86% for the second and third PACAP-38 applications, respectively) (Fig. 8B). Data obtained from three different cell cultures are summarized in Fig. 8C.
Functional Properties of Chromaffin Cells-Activation of cholinergic receptors in chromaffin cells is the main stimulus for mobilizing Ca 2ϩ to induce catecholamine secretion (39). Hence, experiments were undertaken to determine if chromaffin cells from 17-to 20-week-old human fetuses express functional cholinergic receptors. Release of Ca 2ϩ from IP 3 -sensitive Ca 2ϩ pools was tested using methacholine, a muscarinic receptor agonist. Fig. 9A shows that application of methacholine ( (Fig. 9B) similar to that obtained after depolarization using KCl (30 mM) (Fig. 9C).
Because PACAP-38 increased [Ca 2ϩ ] i in human chromaffin cells, the question arises as to whether these cells are capable of catecholamine secretion. A carbon fiber (5 m) for amperometric detection (41) was used to monitor catecholamine secretion from cell clusters. Stimulation with PACAP-38 (5 ϫ 10 Ϫ7 M) in a 2 mM Ca 2ϩ medium, triggered the quantal release of catecholamine by chromaffin cells as demonstrated by the occurrence of spikes under the carbon electrode (Fig. 10A, n ϭ 5). If the cells are preincubated for 20 min with the PAC 1 -R antagonist PACAP (6 -38) (10 M), application of PACAP-38 (5 ϫ 10 Ϫ7 M) failed to stimulate catecholamine release (Fig. 10B, n ϭ  4). Similar results were obtained if the cells are preincubated with (R p )-cAMPS (1 mM, 10 min) prior application of PACAP-38 (data not shown, n ϭ 2). DISCUSSION Our study demonstrates that the fetal human adrenal gland expresses both type I and type II PACAP receptors. Activation of chromaffin PACAP receptors with PACAP-38 induced a transient increase in [Ca 2ϩ ] i originating exclusively from intracellular calcium pools and did not involve Ca 2ϩ influx from the external medium. Moreover, the PACAP-sensitive Ca 2ϩ pool was not mobilized by IP 3 or TG. More importantly, PACAP-38 activated a ryanodine/caffeine-sensitive pool, which involved cAMP and a phosphorylation step by PKA. In addition, we were able to demonstrate that activation of the PAC 1 receptor induced secretion of catecholamine by the chromaffin cells.
The results presented herein indicate that, among the various isoforms of the PACAP receptors, the short fragment and one isoform A of an hip-hop insertion cassette of the PAC 1 receptors are present in the human fetal adrenal gland at the second trimester of gestation. By using PACAP (6 -38), a PAC 1 receptor-specific antagonist (32), we demonstrate that the Ca 2ϩ increase induced by PACAP-38 is mediated by the activation of PAC 1 -R coupled to adenylyl cyclase (2).
One important finding of this study is the observation that the PACAP-induced Ca 2ϩ rise was not affected by the absence of Ca 2ϩ in the external medium. This clearly indicated that PACAP-38 does not trigger a Ca 2ϩ influx through channels or exchangers, in agreement with data obtained in rat hippocampal neurons (31). However, a number of reports indicate that PACAP activates a Ca 2ϩ influx through various pathways, including Ca 2ϩ channels (29,(42)(43)(44). In chromaffin cells from adult human adrenal glands, voltage-dependent Ca 2ϩ channels have been described (45), but no electrophysiological studies have been performed in fetal cells. KCl depolarization, or application of nicotine, were able to induce an increase in [Ca 2ϩ ] i in human fetal chromaffin cells suggesting the presence of functional voltage-dependent Ca 2ϩ channels. However, our results clearly show that these Ca 2ϩ channels were not involved in PACAP-induced Ca 2ϩ increase.
The second most important feature of the present data was that PACAP-responsive Ca 2ϩ pools were insensitive to TG. It has been reported that TG triggers the release of Ca 2ϩ from major nonmitochondrial Ca 2ϩ stores (46), including the IP 3sensitive Ca 2ϩ pool (47). Our data conclusively demonstrate that preincubation of the cells with TG or chronic application of high concentrations of TG (8 M) have no effect on the Ca 2ϩ response triggered by PACAP-38. Previous experiments have shown that, in human fetal chromaffin cells, PACAP-38 triggered low production of IP 3 (16). However, the present results, using either the PLC inhibitor U73132 (35) or Xestospongin C, a blocker of IP 3 receptors (36), demonstrated that Ca 2ϩ release from IP 3 -sensitive Ca 2ϩ stores did not contribute to the increase in [Ca 2ϩ ] i triggered by PACAP-38.
Several lines of evidence pointed toward involvement of the cAMP/PKA pathway in the [Ca 2ϩ ] i increase triggered by PACAP-38 in human fetal chromaffin cells. FSK, which directly activates adenylyl cyclase to produce cAMP, triggered an increase in Ca 2ϩ that was similar in amplitude and kinetics to that induced by PACAP-38. Moreover, like PACAP-38, the response triggered by FSK was independent of the presence of Ca 2ϩ in the external medium. Additionally, blocking PKA using H-89 or (R p )-cAMPS inhibited the effects of both PACAP-38 and FSK on [Ca 2ϩ ] i . Opposite results were found in hippocampal neurons where PACAP-induced Ca 2ϩ response was not triggered by FSK or (Bu) 2 cAMP and insensitive to H-89 (31). Cyclic AMP-sensitive Ca 2ϩ stores have been described in numerous cell types, but their relationship with the IP 3 -and the ryanodine/caffeine Ca 2ϩ stores is still a matter of debate. In several cell types, it has been shown that the ryanodine receptor is subject to phosphorylation by several kinases, including PKA (48,49). Our results clearly indicated that PACAP-and FSK-induced Ca 2ϩ increases were sensitive to PKA-dependent phosphorylation. This, together with the effect of ryanodine/ caffeine, could indicate that PACAP-38 mobilizes Ca 2ϩ from caffeine-sensitive stores in human fetal chromaffin cells.
Application of caffeine to human fetal chromaffin cells triggered a moderate increase in [Ca 2ϩ ] i confirming the presence of caffeine-sensitive Ca 2ϩ stores. In bovine adult chromaffin cells, PACAP released Ca 2ϩ from a ryanodine/caffeine store (50), which was independent of IP 3 , as observed for our cell model. However, it was also reported that Ca 2ϩ rise was insensitive to (R p )-cAMPS contrary to human fetal chromaffin cells where H-89, or (R p )-cAMPS, completely inhibited both PACAP-and FSK-induced responses. The role and weightiness of the caffeinedependent Ca 2ϩ pools in chromaffin cells have been outlined by several authors (51) who report that caffeine-dependent Ca 2ϩ pools release more Ca 2ϩ than IP 3 -dependent Ca 2ϩ pools in permeabilized chromaffin cells (52,53). The signaling pathway has not yet been fully characterized, although several possibilities have been proposed (54). More recently, secretory vesicles in neuroendocrine cells have been demonstrated to constitute a dynamic Ca 2ϩ pool (38). Several features of the vesicle-Ca 2ϩ stores are similar to those defined by PACAP-dependent stores. Indeed, these pools are insensitive to TG, are not mobilized by IP 3 , and are sensitive to caffeine and ryanodine. As proposed by Mitchell et al. (38), cytosolic Ca 2ϩ is pumped into the deep vesicles by an ATP-dependent P-type Ca 2ϩ pump. Ca 2ϩ is released by primed vesicles located near the plasma membrane when [Ca 2ϩ ] i concentrations increase due to Ca 2ϩ channel opening (55). Our data indicate that treatment of chromaffin cells with Na 3 VO 4 did not empty the PACAP-sensitive Ca 2ϩ pools, as illustrated by identical levels of Ca 2ϩ increase obtained before and after Na 3 VO 4 application. However, the fact that the amplitude of the subsequent responses to PACAP-38 decreased as a function of the number of PACAP-38 applications indicated that re-pumping of Ca 2ϩ may have been impaired by Na 3 VO 4 .
By using amperometry, we also demonstrated that human fetal chromaffin cells release catecholamine in response to PACAP-38. [Ca 2ϩ ] i increase is a requirement, as outlined by the absence of secretion in cells treated with PACAP (6 -38) or (R p )-cAMPS. Catecholamine secretion could be linked to the paracrine control of secretion of dehydroepiandrosterone sulfate and cortisol by fetal steroidogenic cells. Indeed, in cell cultures, PACAP-38 stimulates DHEA and cortisol secretion, an effect abolished by preincubation with the ␤-adrenoreceptor antagonist, propranolol (15).
In summary, stimulation of chromaffin cells with PACAP-38 induced a cAMP-dependent increase in intracellular Ca 2ϩ whose properties and regulation exhibit particular features. This model has thus enabled the identification of a novel pathway in the regulation of cAMP-dependent Ca 2ϩ release, which could be specific to fetal tissues. In addition to its role in catecholamine secretion, PACAP-38 could act as a survival factor for chromaffin cells, because, in contrast to fetal cells, chromaffin cells do not undergo apoptosis, either in vivo or in vitro (56,57). Recent studies conducted in knock-out mice for PACAP, revealed that PACAP was not essential for normal development of the adrenal gland, nor for basal catecholamine secretion; however, its presence was essential for adaptive responses. Indeed, mice lacking PACAP were unable to survive in response to a metabolic stress (58). Taken together, these results clearly indicate that PACAP is one of the important factors necessary in maintaining adequate interaction between catecholamine and steroid-producing cells, an interaction that occurs during normal development.