A non-cholinergic transmitter, pituitary adenylate cyclase-activating polypeptide, utilizes a novel mechanism to evoke catecholamine secretion in rat adrenal chromaffin cells.

Pituitary adenylate cyclase-activating polypeptide (PACAP) is the most potent non-cholinergic neurotransmitter to stimulate catecholamine secretion from rat chromaffin cells; however, the mechanism of action is not clear. We used amperometric detection of exocytosis and indo-1 monitoring of [Ca2+]i to identify PACAP actions in cultured chromaffin cells. PACAP (100 nM) required external Ca2+ to evoke secretion. However, unlike nicotine and KCl which caused immediate and relatively brief secretion, PACAP has a latency of 6.8 +/- 0.96 s to the first secretory response and secretion continued for up to 2 min. PACAP elevation of [Ca2+]i showed similar latency and often remained above base line for several minutes following a brief exposure. ZnCl2 (100 microM) selectively inhibited PACAP-stimulated secretion and [Ca2+]i with little effect on nicotine-evoked responses. Nifedipine (10 microM) had little effect on PACAP-evoked secretion but inhibited nicotine-evoked secretion by more than 80%, while omega-conotoxin (100 nM) failed to affect either agonist. PACAP-stimulated cAMP levels required 5 s to significantly increase, consistent with the latency of exocytotic and Ca2+ responses. Forskolin (10 microM) caused responses similar to PACAP. PACAP-evoked exocytosis was blocked by the protein kinase A inhibitor adenosine 3'5'-cyclic monophosphorothioate Rp-diastereomer (Rp-cAMPS). These data showed that PACAP stimulates exocytosis by a mechanism distinctly different from cholinergic transmitters that appears to involve cAMP-mediated Ca2+ influx. Differences in receptor coupling mechanisms and pharmacology of Ca2+ entry stimulated by cholinergic and peptidergic agonists support the idea that the peptidergic system maintains catecholamine secretion under conditions where the cholinergic system desensitizes or otherwise fails.

Pituitary adenylate cyclase-activating polypeptide (PACAP) is the most potent non-cholinergic neurotransmitter to stimulate catecholamine secretion from rat chromaffin cells; however, the mechanism of action is not clear. We used amperometric detection of exocytosis and indo-1 monitoring of [Ca 2؉ ] i to identify PACAP actions in cultured chromaffin cells. PACAP (100 nM) required external Ca 2؉ to evoke secretion. However, unlike nicotine and KCl which caused immediate and relatively brief secretion, PACAP had a latency of 6.8 ؎ 0.96 s to the first secretory response and secretion continued for up to 2 min. PACAP elevation of [Ca 2؉ ] i showed similar latency and often remained above base line for several minutes following a brief exposure. ZnCl 2 (100 M) selectively inhibited PACAP-stimulated secretion and [Ca 2؉ ] i with little effect on nicotineevoked responses. Nifedipine (10 M) had little effect on PACAP-evoked secretion but inhibited nicotine-evoked secretion by more than 80%, while -conotoxin (100 nM) failed to affect either agonist. PACAP-stimulated cAMP levels required 5 s to significantly increase, consistent with the latency of exocytotic and Ca 2؉ responses. Forskolin (10 M) caused responses similar to PACAP. PACAP-evoked exocytosis was blocked by the protein kinase A inhibitor adenosine 3,5-cyclic monophosphorothioate R p -diastereomer (R p -cAMPS). These data show that PACAP stimulates exocytosis by a mechanism distinctly different from cholinergic transmitters that appears to involve cAMP-mediated Ca 2؉ influx. Differences in receptor coupling mechanisms and pharmacology of Ca 2؉ entry stimulated by cholinergic and peptidergic agonists support the idea that the peptidergic system maintains catecholamine secretion under conditions where the cholinergic system desensitizes or otherwise fails.
There is convincing evidence that secretion of adrenal medullary hormones in various species is regulated by peptidergic neurotransmitters. Among a long list of peptides that exist in the adrenal medulla, vasoactive intestinal polypeptide (VIP) 1 satisfied several important criteria to be classified as a noncholinergic neurotransmitter in the rat adrenal gland (1). VIPlike immunoreactivity has been detected in nerve terminals of adrenal glands of several species (1)(2)(3)(4). However, there is also evidence against the existence of VIP immunoreactive fibers in adrenal medulla (5) and splanchnic neurons (6). While VIP stimulates the cAMP as well as phosphatidylinositol pathways (7), it is difficult to draw firm conclusions about the mechanism of peptide-evoked catecholamine secretion from studies of intact adrenal gland.
At the same time that VIP was being characterized as a medullary neurotransmitter, a new member of the VIP/glucagon/secretin family of peptides (pituitary adenylate cyclaseactivating polypeptide, PACAP) was discovered and shown to be a most potent activator of adenylyl cyclase (8). PACAP occurs in both a 27 (PACAP 27 ) and 38 (PACAP 38 ) amino acid form and has about 70% homology with VIP (8,9). PACAP immunoreactive fibers have been identified in the adrenal gland 2 (but see Ref. 10), and PACAP receptors have been demonstrated on chromaffin cells (10,11). Release of PACAP-like material from perfused adrenal following splanchnic nerve stimulation has been demonstrated (12). Importantly, PACAP is one of the most potent secretagogues in chromaffin cells in culture (13)(14)(15) and in perfused adrenal gland (16). Finally, the adrenal gland has been reported to contain the second highest concentration of PACAP among peripheral organs of the rat (17). Based on these findings, we have suggested that both VIP and PACAP might function as non-cholinergic neurotransmitters controlling catecholamine synthesis and secretion in the rat adrenal medulla (Ref. 12 and see also Refs. 18 and 19). The secretory properties of PACAP have been examined in whole adrenal gland (16,20) and mass cultures of chromaffin cells (13,14); however, the mechanism of PACAP-evoked catecholamine secretion remains unclear.
We undertook the present work to identify the mechanism by which PACAP mobilizes Ca 2ϩ and stimulates secretion in primary cultures of rat chromaffin cells. Amperometric detection of exocytosis and indo-1 fluorescence measurement of [Ca 2ϩ ] i were used to define PACAP effects on single chromaffin cells. Our recently developed protocols were used to isolate and identify internal versus external sources of Ca 2ϩ used in exocytosis. PACAP-evoked exocytosis was monitored in cells during transient exposure to a Ca 2ϩ -free environment and in cells in the presence of Ca 2ϩ but depleted of internal Ca 2ϩ stores. Our findings show that a peptidergic transmitter utilizes extracellular Ca 2ϩ to evoke catecholamine secretion, but by a mechanism distinctly different from acetylcholine acting on nicotinic and muscarinic receptors in this preparation. The data indicate that PACAP increases Ca 2ϩ influx in association with cAMP production.

MATERIALS AND METHODS
Primary Cultures of Rat Chromaffin Cells-Chromaffin cells were cultured from 19-to 31-day-old Sprague-Dawley rat pups as described (21)(22)(23). Briefly, medullary fragments were incubated in 1.5 mg/ml type I collagenase and 0.5 mg/ml DNase (Worthington) in phosphate-buffered saline, followed by trypsin (1.25 mg/ml in PBS) and DNase (0.5 mg/ml), and cells dissociated by trituration. Cells were washed by centrifugation (600 ϫ g, 5 min) and plated on collagen-coated glass coverslips in M-199 medium (Life Technologies, Inc.) plus serum and 0.1 M dexamethasone and used after 2-7 days. Immunohistochemical staining showed that cultures contained greater than 90% tyrosine hydroxylase positive cells. 3 Coverslips with attached chromaffin cells were transferred to HEPES-buffered bath solution for monitoring of exocytosis or [Ca 2ϩ ] i . The solution contained (in mM) 119 NaCl, 4.7 KCl, 1.2 MgSO 4 , 2.5 CaCl 2 , 10 HEPES, 10 glucose, pH 7.4, with NaOH. Ca 2ϩ -free solution had no added CaCl 2 and was supplemented with 1 mM EGTA.
Electrochemical Detection of Exocytosis-Exocytosis from single rat chromaffin cells was detected using established micro-electrochemical techniques (15, 24 -26). Eight-m carbon fibers were cannulated into polyethylene tubing, pulled and cut (final tip diameter Ͻ 9 m), and sealed into glass capillaries. Capillaries were back-filled with 3 M KCl and a Ag/AgCl wire used for connection to an EI 400 potentiostat (holding potential: 650 mV) coupled to a personal computer for data acquisition and analysis as described (15). Carbon fiber electrodes were positioned within 1 m of a chromaffin cell to record currents resulting from oxidation of exocytotically released catecholamines (25,26).
Control of Internal and External Ca 2ϩ -Secretagogues in control or Ca 2ϩ -free bath solutions were applied by pressure ejection from pulled glass capillaries (ϳ10 m tip diameter) positioned about 20 m from the cell. Ca 2ϩ -free solution had no added CaCl 2 and was supplemented with 1 mM EGTA. Intracellular Ca 2ϩ stores were depleted by 15-min pretreatment in Ca 2ϩ -free solution. Control of the Ca 2ϩ environment was verified in two ways. First, by showing that in a Ca 2ϩ -containing bath, application of KCl (or nicotine) in Ca 2ϩ -free pipette solution failed to produce exocytosis ( Fig. 1a) or elevate [Ca 2ϩ ] i , (not shown) while mobilization of internal Ca 2ϩ by caffeine under the same conditions gave positive responses (Fig. 1a). Second, in Ca 2ϩ -depleted cells, caffeine (applied with Ca 2ϩ in the pipette) failed to stimulate the cells, while KCl depolarization continued to give positive responses (Fig. 1b).
In most experiments each cell was used as its own control.
Measurement of Cellular cAMP Content-For determination of cAMP content, chromaffin cells were plated at a concentration of about 5 medullae (250,000 cells)/dish and used after 3 days. Culture medium was removed by washing with Krebs buffer solution, and 100 nM PACAP in Krebs was added to the dish for various time periods from 0 to 30 s. The reaction was stopped and cells extracted in 1 ml of ice-cold trichloroacetic acid (6%). The extract was centrifuged (3000 ϫ g, 15 min) and the pellet used for protein determination as described (30). The supernatant was extracted with ethyl ether, evaporated to dryness, and the residue used for cAMP determination using a DuPont NEN radioimmuno assay kit according to the protocols supplied by the manufacturer. Fig. 2a shows a representative amperometric recording of exocytotic events stimulated in an individual rat chromaffin cell by a 10-s application of 100 nM PACAP from an ejection pipette aimed at the cell. Both the bath and pipette solutions contained 2.5 mM Ca 2ϩ . Exocytosis began 6.8 Ϯ 0.96 s (25 cells) after the start of PACAP application and continued for a minute or more after agonist application. This pattern of stimulation is much different than exocytosis produced by depolarization with KCl (Refs. 16 and 31 and see Fig. 1a) or stimulation of nicotinic receptors (below), both of which cause an almost immediate and brief (20 -30 s) burst of exocytotic events.

PACAP Requires Extracellular Ca 2ϩ for Exocytosis-
The role of Ca 2ϩ entry from the external medium in PACAP-induced exocytosis was examined by applying PACAP in a nominally Ca 2ϩ -free pipette solution to chromaffin cells in a 2.5 mM Ca 2ϩ -containing bath solution. PACAP (100 nM for 30 s) produced no exocytosis during the application period (Fig. 2b). However, immediately after the cessation of peptide application, there was a massive exocytotic response as the Ca 2ϩ -free pipette solution was displaced by the Ca 2ϩ -containing bath solution. Exocytosis continued for several minutes (not shown).

Exocytosis by PACAP in Cells Depleted of Internal Ca 2ϩ
Stores-To confirm the utilization of external Ca 2ϩ in PACAPevoked exocytosis experiments were performed with cells depleted of internal Ca 2ϩ by maintaining cells in Ca 2ϩ -free (1 mM EGTA) bath solution for 15 min prior to and during PACAP application. Ten second application of PACAP plus Ca 2ϩ to Ca 2ϩ -depleted cells caused a burst of exocytosis with typical latency (Fig. 2c). The duration of PACAP-evoked exocytosis in Ca 2ϩ -depleted cells was much less compared with cells with normal Ca 2ϩ stores (compare Fig. 2, a and c).
PACAP-evoked Elevation of [Ca 2ϩ ] i -The effects of PACAP on [Ca 2ϩ ] i were determined in parallel with exocytosis experiments using identical protocols. Fig. 3a shows that a 10-s application of PACAP to an indo-1-loaded chromaffin cell caused an increase in [Ca 2ϩ ] i that exhibits a latency to onset and prolonged duration consistent with PACAP stimulated exocytosis. In some chromaffin cells, the initial rise in [Ca 2ϩ ] i was followed by fluctuating [Ca 2ϩ ] i after a brief application of PACAP (see Fig. 4 for example). In six cells monitored for 1 min or longer after PACAP application, the average latency to the beginning of [Ca 2ϩ ] i elevation was 7.9 Ϯ 0.8 s. The mean peak was 398 Ϯ 35 nM and in four of the six cells [Ca 2ϩ ] i remained elevated for the duration of recording (352 Ϯ 29 nM at 60 s). The remaining two cells exhibited Ca 2ϩ fluctuations throughout the recording period.
When [Ca 2ϩ ] i was monitored after application of PACAP in a Ca 2ϩ -free pipette solution (Fig. 3b), there was no change in [Ca 2ϩ ] i until after cessation of peptide application, identical to the pattern of exocytosis produced under the same conditions. Finally, the effects of PACAP on [Ca 2ϩ ] i were determined in cells depleted of internal Ca 2ϩ stores (Fig. 3c). The rapid decline of [Ca 2ϩ ] i under these conditions is consistent with a more rapid sequestration of Ca 2ϩ into depleted internal stores and confirms the role of external Ca 2ϩ entry in PACAP-evoked catecholamine secretion.
Effects of Ca 2ϩ Channel Antagonists on PACAP Action-We used a pharmacological approach to discriminate between the delayed but long lasting responses stimulated by PACAP and the immediate brief effects of depolarizing stimuli. Among sev- eral Ca 2ϩ channel antagonists tested, only Zn 2ϩ exhibited differential effects on PACAP versus other secretagogues. ZnCl 2 (100 M) almost completely suppressed PACAP-evoked exocytosis (Fig. 4a) and elevated [Ca 2ϩ ] i (Fig. 4b) with little or no effect on nicotine-stimulated responses (Fig. 4, c and d) or KCl-evoked exocytosis (not shown). In the example in Fig. 4b, brief application of PACAP caused a fluctuating rather than sustained elevation of [Ca 2ϩ ] i . Nifedipine (10 M) had no effect on PACAP-induced secretion of catecholamines. PACAP stimulated 71 Ϯ 7 and 72 Ϯ 7 exocytotic events during 2-min recording in control and nifedipine-treated cells (n ϭ 9), respectively. -Conotoxin produced an approximate 18% decline in secretion (58 Ϯ 6 events/2 min recording, n ϭ 9), but the difference was not significant. However, when tested against nicotine-evoked secretion, nifedipine produced 83% inhibition (70 Ϯ 3 and 12 Ϯ 2 events in control and 1 M nifedipine, respectively, n ϭ 5), and -conotoxin had no effect (103 Ϯ 22%) compared with untreated controls (n ϭ 5).
Forskolin Effects on Exocytosis and [Ca 2ϩ ] i -Because PACAP is established as a potent stimulator of adenylyl cyclase and cAMP formation, we questioned whether forskolin, which also stimulates adenylyl cyclase, would have secretory actions similar to PACAP. Application of forskolin (10 M for 10 s) produced a long-lasting secretory response in 2.5 mM Ca 2ϩ medium (Fig. 5a). The delay between beginning of forskolin application and detection of the first exocytotic event was 7.5 Ϯ 0.95 s (28 cells). When forskolin was administered in a nominally Ca 2ϩ -free pipette solution, exocytotic events were detected only after cessation of the Ca 2ϩ -free solution (Fig. 5b). Application of forskolin plus Ca 2ϩ to Ca 2ϩ -depleted chromaffin cells, like PACAP, stimulated exocytosis only during the period when Ca 2ϩ was available to the cells (Fig. 5c).
When tested on indo-1 loaded chromaffin cells, forskolin caused a rise in [Ca 2ϩ ] i that was similar to that produced by PACAP. The average peak [Ca 2ϩ ] i was 322 Ϯ 29 nM with a latency of 6.7 Ϯ 1 s (n ϭ 5). However, forskolin typically produced fluctuations in [Ca 2ϩ ] i (see Fig. 6b for example) rather than the sustained elevation seen with PACAP. In this regard, it was noted that forskolin-evoked exocytosis was qualitatively similar to PACAP in terms of latency, use of external Ca 2ϩ , and sensitivity to Zn 2ϩ (Fig. 6a). However, compared with PACAP, forskolin-evoked exocytosis was less robust with fewer total exocytotic events (compare Fig. 6a and Fig. 4a). These data suggest that PACAP-evoked catecholamine secretion has a prominent, but not exclusive, cAMP-mediated mechanism.
Characteristic Delay in Onset of Exocytosis by PACAP-In all tests performed, the latency between beginning of agonist application and detection of the first exocytotic event appeared to be most closely correlated to the mechanism of agonist action (Table I). Nicotine and excess KCl, which act by depolarization and opening of voltage-dependent Ca 2ϩ channels, cause secretion to occur almost instantaneously. The observed delay of a few hundred ms with these agonists reflects the positioning of the ejection pipette 20 m from the cells and the resulting delay in changing the environment around the cell. Under these same conditions, agonists which mobilize internal Ca 2ϩ , muscarine, and caffeine, exhibited a latency between 2 and 3 s, while PACAP (or forskolin) acting presumably through a cAMP-dependent phosphorylation cascade, required approximately 7 s to evoke secretion.
To determine if the time required for cAMP elevation could account for the latency of PACAP effects, we monitored cAMP levels during PACAP exposure. PACAP required 5 s to significantly increase cAMP levels compared with unstimulated controls (Fig. 7). The small and variable increase in cAMP observed at 2.5-s exposure to PACAP was not statistically significant. cAMP levels appeared to reach a plateau at 5 s and remained at about the same level during PACAP exposure for up to 30 s (Fig. 7).
To further implicate cAMP-dependent phosphorylation, PACAP effects were tested in the presence of the protein kinase A inhibitor, adenosine 3Ј,5Ј-cyclic monophosphorothioate R pdiastereomer (R p -cAMPS) (32). R p -cAMPS (300 M for 30 min) caused a significant reduction of PACAP-evoked exocytosis with little effect on nicotine-evoked responses (Fig. 8). DISCUSSION Several lines of evidence have been presented to define the secretory mechanism of PACAP, a non-cholinergic co-transmitter in the adrenal medulla. The secretory mechanism of PACAP appears distinctly different from acetylcholine acting through nicotinic and muscarinic receptor stimulation. PACAP-evoked exocytosis and elevated [Ca 2ϩ ] i occurred only after a pronounced latency and required the presence of extracellular Ca 2ϩ . The time course of PACAP-stimulated cAMP elevation and the close correspondence between effects of PACAP and forskolin support the conclusion that PACAP actions are mediated by cAMP-dependent activation of protein kinases.
PACAP (or forskolin) stimulated exocytosis and elevated [Ca 2ϩ ] i only when external Ca 2ϩ was present, indicating that the peptide causes Ca 2ϩ entry into chromaffin cells. The absence of exocytosis and elevated [Ca 2ϩ ] i during 30-s application of PACAP without CaCl 2 also shows that Ca 2ϩ influx is required for PACAP-evoked catecholamine secretion. The pronounced increase in [Ca 2ϩ ] i and exocytosis upon cessation of the stimulus was most likely due to stimulation of intracellular second messengers during the application period, coupled with high affinity binding of PACAP and slow dissociation from its receptors, producing an exaggerated response as the pipette solution was displaced by the Ca 2ϩ containing bath solution.
The mechanism of Ca 2ϩ entry following PACAP application was different than that produced by acetylcholine. Acetylcholine effects were sensitive to block of L-type Ca 2ϩ channels by nifedipine, but PACAP effects were not. This is somewhat different from PACAP effects in mass cultures of porcine chromaffin cells, which were reported to be inhibited by nifedipine (14). Neither agonist was opposed by block of N-type channels by -conotoxin. The finding that 100 M Zn 2ϩ discriminates between PACAP and other secretagogues that stimulate Ca 2ϩ entry supports the idea that PACAP promotes Ca 2ϩ entry by a distinct mechanism. One possibility is that PACAP causes a phosphorylation-dependent recruitment of channels not active during nicotinic or KCl induced depolarization. While forskolin-evoked exocytosis was qualitatively similar to that produced by PACAP, forskolin consistently produced fewer exocytotic events. This may be due to the reported ability of PACAP 38 to stimulate the inositol lipid cascade along with elevated cAMP levels (33). Multiple signaling pathways stimulated by PACAP but not forskolin would account for the more robust secretion by PACAP.
The most intriguing characteristic of PACAP stimulation was the approximate 7-s delay between application and onset of action. Only a fraction of the latency could be attributed to time required for the pipette solution to displace bath solution around an individual cell. Delay intrinsic to the protocol was less than 0.5 s as indicated by the latency when depolarizing stimulus was administered. Receptor-mediated signal transduction is unlikely to account for much of the latency, since nicotinic receptor activation produced an almost immediate response. The observed latencies appear to be closely related to the intracellular mechanism involved in stimulating catecholamine secretion. Agents that mobilize internal Ca 2ϩ (muscarine and caffeine) acted within 2-3 s, while agents which elevate cAMP (forskolin and PACAP) required about 7 s to produce exocytosis. Receptor-mediated production of second messengers is not likely to account for the latency for two reasons. First, both muscarine and PACAP stimulate complex second messenger pathways, but produced exocytosis with significantly different latencies. Second, caffeine, which mobilizes internal Ca 2ϩ but does not act through a plasma membrane receptor, acts with the same delay as muscarine, while forskolin, which elevates cAMP independent of membrane receptors, acts with the same delay as PACAP. These observations, cou-  pled with the instantaneous action of depolarizing agents, suggest that the last step in the signaling pathway that directly causes Ca 2ϩ influx or liberation of internal stores is the factor controlling latency. The approximately 5-s latency for PACAPevoked elevation of cAMP levels in the present work is similar to the latent period (about 5 s) before cAMP-dependent enhancement of Ca 2ϩ current is observed in cardiac cells (34). The high density of chromaffin cells in the cultures (Ͼ90% tyrosine hydroxylase positive) 3 makes it unlikely that non-chromaffin cells account for the observed changes in cAMP levels. Furthermore, inhibition of protein kinase A by R p -cAMPS significantly reduced PACAP-evoked exocytosis without affecting nicotineevoked responses. These findings coupled with the observed 7-s latency to PACAP-evoked exocytosis support the idea that cAMP production and activation of protein kinase A are involved in stimulating Ca 2ϩ entry and exocytosis. However, the mechanism of this action remains unknown.
In conclusion, the results presented here demonstrate the importance of multiple transmitters acting through cholinergic and non-cholinergic pathways in the adrenal synapse. PACAP stimulation of catecholamine secretion is not redundant or simply modulatory, but occurs independent of other secretagogues via mechanisms distinct from cholinergic receptor coupled pathways. Thus peptidergic transmission is likely to maintain catecholamine secretion during periods of stress or situations where cholinergic receptors desensitize or cholinergic transmission otherwise fails.