P(IC

KV

Dl

Ag(Cic(gccPtnrTPtpotssqnaPsmtavsswtNpol

*EAtspP(3N

Neuroscience 126 (2004) 335–346

0d

ITUITARY ADENYLATE CYCLASE ACTIVATING POLYPEPTIDEPACAP) DECREASES NEURONAL SOMATOSTATINMMUNOREACTIVITY IN CULTURED GUINEA-PIG PARASYMPATHETICARDIAC GANGLIA

etnr

Kc

Iswftdagloaosera(B2

gcfiipPoIsipmrbawpm

. M. BRAAS, T. M. ROSSIGNOL, B. M. GIRARD,

. MAY AND R. L. PARSONS*

epartment of Anatomy and Neurobiology, University of Vermont Col-ege of Medicine, 89 Beaumont Avenue, Burlington, VT 05405, USA

bstract—Postganglionic parasympathetic neurons inuinea-pig cardiac ganglia exhibit choline acetyltransferaseChAT)-immunoreactivity, and a large fraction (60%) of thehAT-positive cardiac neurons co-express somatostatin-

mmunoreactivity. This co-expression remained when theardiac ganglia explants were maintained in culture for 72 h40% somatostatin-immunoreactive). The guinea-pig cardiacanglia neurons express the high affinity pituitary adenylateyclase activating polypeptide (PACAP)-selective PAC1 re-eptor, and treatment of the ganglia explants with 20 nMACAP27 for 72 h to evaluate PACAP regulation of soma-

ostatin expression revealed a dramatic 85% decrease in theumber of somatostatin-IR neurons (6% somatostatin-IR neu-ons) compared with untreated control explant preparations.he decrease in percentage of somatostatin-IR neurons byACAP27 was time- and concentration-dependent, and selec-

ive for PACAP27; PACAP38 and vasoactive intestinalolypeptide were less effective. PACAP6-38, a PACAP antag-nist, eliminated the PACAP27-induced change in soma-ostatin positive neurons. The PACAP-mediated decrease inomatostatin-IR neurons was eliminated in calcium-deficientolutions and by the addition of nifedipine, indicating a re-uirement for calcium influx through L-type calcium chan-els. The addition of either the calmodulin inhibitor N-(4-minobutyl)-1-naphthalenesulfonamide or the MEK inhibitorD98059, also eliminated the PACAP27-induced decrease inomatostatin-IR cells. The PACAP27-mediated effect on so-atostatin expression was not affected by inhibitors of pro-

ein kinase A or phospholipase C, but was reduced by thedenylyl cyclase inhibitor SQ22356, suggesting cAMP in-olvement. Semiquantitative and quantitative reverse tran-cription PCR prosomatostatin transcript measurementshowed that cardiac ganglia prosomatostatin mRNA levelsere not diminished by chronic PACAP27 exposure despite

he dramatic decrement in somatostatin-expressing neurons.euronal peptide-IR content represents a balance betweenroduction and secretion. These results suggested that onef the primary effects of PACAP exposure may be enhanced

evels of neuropeptide release that exceeded production lev-

Corresponding author. Tel: �1-802-656-2230; fax: �1-802-656-8704.-mail address: rodney.parsons@uvm.edu (R. L. Parsons).bbreviations: ANOVA, analysis of variance; ChAT, choline acetyl-

ransferase; CT, threshold cycle; NPY, neuropeptide Y; PAC1, PACAPelective receptor; PACAP, pituitary adenylate cyclase activatingolypeptide; PCR, polymerase chain reaction; PKA, protein kinase A;LC, phospholipase C; U73122, 1-[6-((17 b-3-methoxyestra-1,3,5-

10)triene-17-yl)amino)hexyl]-1H-pyrrole-2,5,-dione; PD98059, 2�-amino-�-methoxyflavone; VIP, vasoactive intestinal polypeptide; W-13,

e-(4-aminobutyl)-1-naphthalenesulfonamide.

306-4522/04$30.00�0.00 © 2004 IBRO. Published by Elsevier Ltd. All rights reseroi:10.1016/j.neuroscience.2004.04.007

335

ls, resulting in somatostatin depletion and a decrement inhe number of identifiable somatostatin-expressing cardiaceurons. © 2004 IBRO. Published by Elsevier Ltd. All rightseserved.

ey words: neuropeptide, autonomic neuron, L-type calciumhannels, peptide secretion.

ntrinsic cardiac ganglia are clusters of cholinergic, para-ympathetic postganglionic neurons located close to orithin cardiac tissues. These ganglia, believed initially to

unction as simple relay stations, are now considered to behe final local neuronal integrative centers involved in car-iac regulation (Loffelholz and Pappano, 1985; Randallnd Wurster, 1994; Ardell, 2001; Parsons, 2004). In theuinea-pig, all of the cardiac ganglia neurons exhibit cho-

ine acetyltransferase (ChAT)-immunoreactivity, indicativef acetylcholine as their primary transmitter. As in otherutonomic ganglia, the cardiac ganglia neurons also dem-nstrate diversity in neurotransmitter/neuropeptide expres-ion; subpopulations of the ChAT-immunopositive neuronsxhibit immunoreactivity for a number of bioactive neu-opeptides, including somatostatin, neuropeptide Y (NPY),nd pituitary adenylate cyclase activating polypeptidePACAP), and nitric oxide synthase (Steele et al., 1994;raas et al., 1998; Kennedy et al., 1998; Calupca et al.,000a,b).

In addition to vagal cholinergic innervation of theuinea-pig cardiac ganglia, virtually every postganglionicardiac neuron is innervated by PACAP-immunoreactivebers. The co-localization of PACAP and ChAT with-

n pericellular fiber networks surrounding individualostganglionic cardiac neurons indicated that theACAP-containing fibers were parasympathetic pregangli-nic terminals (Braas et al., 1998; Calupca et al., 2000b).n addition, all of the cardiac neurons express PACAP-elective PAC1 receptors and PACAP27 depolarized and

ncreased membrane excitability of essentially all guinea-ig cardiac neurons (Braas et al., 1998). The PACAP-ediated increase in cardiac neuron excitability is likely

esponsible for the negative chronotropic effects producedy exogenous application of PACAP to isolated guinea-pigtrial preparations (Seebeck et al., 1996). Thus, previousork suggested that PACAP peptides released fromreganglionic fiber terminals may be excitatory neuro-odulators amplifying the parasympathetic inhibitory influ-

nce on cardiac tissues.

ved.

tt(PgD2pssmpp

terncnlo1eadgapfaeAmec(2tsgmtuodirpnisssstsc

lr

G

Ea(ipiNoaaKT

gKasietpadumpt3edtb

dm(t(p3maici

sb(m2ftiMtf

I

F

K. M. Braas et al. / Neuroscience 126 (2004) 335–346336

PACAP peptides, in addition to their actions as impor-ant intercellular signaling molecules, function as potentrophic factors and modulators of gene transcriptionArimura, 1998; Hansel et al., 2001; Waschek, 2002).ACAP peptides stimulate sympathetic neuroblast mito-enesis, neuritogenesis and survival (Lu et al., 1998;iCicco-Bloom et al., 2000; Nicot and DiCicco-Bloom,001). In sympathetic ganglia, the PACAP peptides alsootently regulate transmitter and neuropeptide production,ecretion and mRNA levels (May and Braas, 1995). NPYecretion by PACAP peptides in sympathetic neurons isediated by activation of PAC1 receptors and requires thereferential activation of phospholipase C (PLC) signalingathways (Braas and May, 1999).

To better appreciate the scope of PACAP function inhe cardiac ganglia, the present study was initiated tovaluate the ability for PACAP peptides to regulate neu-opeptide production and secretion in parasympatheticeurons. In the guinea-pig, estimates of the number ofardiac ganglia neurons range from a few hundred to 2500eurons. The neurons are distributed essentially as mono-

ayers of neuron clusters within the epicardial muscle layerf the posterior atria (Mawe et al., 1996; Leger et al.,999). All guinea-pig cardiac ganglia neurons are cholin-rgic and in addition, a large percentage of these neuronslso express somatostatin (Steele et al., 1994). The diffuseistribution and limited number of neurons in cardiac gan-lia preclude direct biochemical measurements of cellularnd secreted neuropeptide levels. However, the guinea-ig cardiac ganglia whole mount preparation is well suitedor quantitative immunocytochemical analysis (Steele etl., 1994; Mawe et al., 1996; Braas et al., 1998; Kennedyt al., 1998; Leger et al., 1999; Calupca et al., 2000a,b).lso, the cardiac ganglia whole mount preparations can beaintained readily in explant organ culture which allowsxperimental assessments of the PACAP-inducedhanges in neuronal neuropeptide expression over timeKennedy et al., 1998; Lynch et al., 1999; Calupca et al.,000a,b). Accordingly, the aims of the present study wereo quantify PACAP-induced changes in numbers ofomatostatin-immunoreactive neurons in the cardiac gan-lia whole mount explant preparations, to establish theechanisms underlying these changes, and to correlate

hese variations with ganglia somatostatin transcript levelssing reverse transcription PCR approaches. The resultsf this study show that PACAP27 produced an apparentecrease in the number of cardiac ganglia somatostatin-

mmunoreactive neurons in a time-dependent manner thatequired PAC1 receptor activation of cAMP/MEK signalingathways, and calcium influx through L-type calcium chan-els. The marked decrease in numbers of somatostatin-

mmunoreactive neurons was not accompanied by mea-urable changes in pro-somatostatin transcript levels, con-istent with the interpretation that PACAP did not alter theomatostatin-expressing neuronal population but alteredignificantly neuronal neuropeptide production and secre-ion homeostasis. The diminished detection ofomatostatin-containing neurons is postulated to be the

onsequence of accentuated neuronal somatostatin re- l

ease causing a depletion of cellular secretory granules inesponse to PACAP stimulation.

EXPERIMENTAL PROCEDURES

eneral methods

xperiments were performed in vitro on atrial whole mount prep-rations containing the cardiac ganglia from Hartley guinea-pigsmixed sex; 250–400 g). Guinea-pigs were killed by halothane orsofluorane overdose followed by exsanguinations using animalrotocols approved by the University of Vermont Institutional An-

mal Care and Use Committee and methods described in theational Institutes of Health Guide for the Care and Use of Lab-ratory Animals. All efforts were made to minimize the number ofnimals used and their suffering. The heart was quickly removednd placed in cold standard Krebs’ solution (in mM: 121 NaCl, 5.9Cl, 2.5 CaCl2, 1.2 MgCl2, 25 NaHCO3, 1.2 NaH2CO4, 8 glucose).he pH was maintained at 7.4 by aeration with 95% O2–5% CO2.

Atrial whole mount preparations containing the cardiac gan-lia were prepared as described previously (Mawe et al., 1996;ennedy et al., 1998; Calupca et al., 2000a,b). In guinea-pigs, tworeas of atrial tissue are rich in cardiac ganglia-an area around theuperior vena cava and an area surrounding the coronary sinus,nteratrial septum and small cardiac vein (Mawe et al., 1996; Legert al., 1999). For the present study, the cardiac ganglia prepara-ion consisted of the region around the coronary sinus, the inferiorortion of the interatrial septum and small cardiac vein (Lynch etl., 1999). The number of ChAT-positive neurons identified inifferent whole mount preparations ranged from 200 to 1500. Thenderlying atrial muscle layers were removed, and the ventricularuscle, except for a small piece used to pin out the whole mountreparation, was removed also. The preparations were main-ained in Sylgard-lined dishes with oxygenated Krebs solution at5–37 °C. Control and cultured explanted preparations were fixedither in 4% paraformaldehyde for 24 h at 4 °C or 2% paraformal-ehyde containing 0.2% picric acid for 2 h at room temperature;he immunocytochemical staining patterns were identical underoth protocols.

For explant studies, the cardiac ganglia preparations wereissected under sterile conditions. The explant preparations wereaintained at 37 °C in culture media consisting of DMEM-F12

1:1) containing 10% horse serum, gentamicin (10 �g/ml), ampho-ericin B (3.75 �g/ml), penicillin (100 units/ml) and streptomycin100 �g/ml; Sigma Chemical Company, St. Louis, MO, USA). Thereparations were placed on a slowly rocking shaker table in a7 °C, 95% O2 and 5% CO2 incubator. The preparations wereaintained for 24, 48 or 72 h. The culture media for vehicle controlnd peptide-treated explants was changed every 24 h. When

ntracellular signaling pathway inhibitors or the L-type calciumhannel inhibitor, nifedipine, was used, the culture media contain-ng these drugs were changed every 12 h.

The following signaling pathway inhibitors were prepared astock solutions and added directly to the culture media: N-[2-((p-romocinnamyl)amino)ethyl]-5-isoquinolinesulfonamide-HCl25 �M, protein kinase A [PKA] inhibitor); 1-[6-((17 b-3-ethoxyestra-1,3,5-(10)triene-17-yl)amino)hexyl]-1H-pyrrole-

,5,-dione (U73122; 10 �M, PLC inhibitor); 9-(tetrahydro-2�-uryl)adenine (50 �M adenylyl cyclase inhibitor); N-(4-aminobu-yl)-1-naphthalenesulfonamide (W-13; 70 �M, calmodulinnhibitor); and 2�-amino-3�-methoxyflavone (PD98059; 50 �M,

EK inhibitor). Nifedipine (10 �M) was included in some cultureso inhibit L-type calcium channels. Control experiments were per-ormed to ensure that addition of vehicle alone was ineffectual.

mmunohistochemistry

or immunocytochemical studies, the preparations were double-

abeled using antisera produced against somatostatin and ChAT

fKTpoACTbatUwv(Hi

M

Cspaflmc(aobisaotCrna2c

fcyac

R(

Cct(Bw(sopuwsiasVrpsop

spgtorCprcdgcc4ppBr(

T

O

SSSSSSSSSS

a

as

K. M. Braas et al. / Neuroscience 126 (2004) 335–346 337

ollowing procedures described previously (Braas et al., 1998;ennedy et al., 1998; Lynch et al., 1999; Calupca et al., 2000a,b).he fixed atrial tissue was washed in phosphate-buffered saline,ermeabilized with 0.5% Triton X-100 and incubated at 4 °Cvernight with mouse anti-somatostatin (1:400; SOMA 10; Dr.lison Buchan, University of British Columbia, Vancover, B.C.,anada) and goat anti-ChAT (1:1000; Chemicon International,emecula, CA, USA). The preparations were washed and incu-ated for 2 h at room temperature with FITC-conjugated donkeynti-goat and Cy3-conjugated donkey anti-mouse secondary an-

isera (Jackson ImmunoReseach Laboratories, West Grove, PA,SA). Each cardiac ganglia preparation was washed, mountedith Citifluor (UKA Chemical Laboratory, Canterbury, UK) andiewed with a Zeiss (Carl Zeiss, Thornwood, NY, USA) or NikonNikon, Melville, NY, USA) fluorescence photomicroscope withBO 100 W UV light source and filters for FITC and Cy3. Spec-

ficity of each antiserum has been characterized extensively.

orphometric analyses

omputerized morphometric analyses of immunocytochemicallytained parasympathetic neurons present in the cardiac gangliareparations followed previously published procedures (Lynch etl., 1999). In brief, a motorized X-Y stage mounted on a Nikonuorescence photomicroscope equipped with an HBO 100 Wercury arc lamp light source, Lucivid video hardware and a

omputer equipped with Windows 95 and Neurolucida softwareMicroBrightfield, Colchester, VT, USA) were used. Neurolucida iscomputerized morphometric system that enabled the digitizationf serial regions of the cardiac ganglia preparations and assem-led a map of the preparation. Numbers of postganglionic neurons

n each preparation were determined by the numbers of ChAT-tained cells (Mawe et al., 1996; Braas et al., 1998; Kennedy etl., 1998; Lynch et al., 1999; Calupca et al., 2000a,b). The numberf cardiac ganglia neurons that expressed somatostatin was de-ermined as a fraction of the total number of neurons that werehAT-immunoreactive [(somatostatin-IR neurons/ChAT-IR neu-

ons)�100�%]. Whole mounts processed for dual ChAT immu-oreactivity and Hoechst 33258 nuclear staining failed to revealpoptotic profiles in cardiac ganglia in either the 72 h untreated or0 nM PACAP-treated explants. These observations assured thathanges in neuronal populations were not due to cell death.

Studies for the different experimental treatments were per-ormed in parallel and shared the same quantitative data forontrol and PACAP-treated explants. Student’s t-test or an anal-sis of variance (ANOVA) was used to evaluate differencesmong groups. Differences were considered statistically signifi-

able 1. RT-PCR oligonucleotide primers for somatostatina

ligo Sequence

ST #4 5�-AGA TGC TGT CCT GCC GTC TCC AGT GST #2 5�-ACA GTC TTC AAT TTC TAA CGC AGG GTST #4 5�-AGA TGC TGT CCT GCC GTC TCC AGT GST #3 5�-AGA GCA TCG TTC TCC GTC TGG TTG GGST #4 5�-AGA TGC TGT CCT GCC GTC TCC AGT GST #5 5�-CCG GGT TCG AGT TCG CAG ACC T-3�

ST #4 5�-AGA TGC TGT CCT GCC GTC TCC AGT GST #6 5�-GAA GAA ATT CTT GCA GCC CGC TTT G-ST #7 5�-TCA GTT TCT GCA GAA GTC TCT-3�

ST #8 5�-GCT AAC AGG ATG TGA ATG TCT-3�

Primer positions are given as the nucleotide sequences for the Ratccession number J00787. Primers were selected to regions of high homatostatin gene.

ant if P�0.05. e

everse transcription-polymerase chain reactionPCR)

ardiac ganglia preparations were dissected under RNase-freeonditions, and total RNA was extracted from individual prepara-ions using the RNA STAT-60 total RNA/mRNA isolated reagentTel-Test B, Inc., Friendswood TX, USA; May and Braas, 1995;raas et al., 1998; Braas and May, 1999). Complementary DNAas synthesized using the SuperScript II Preamplification System

Invitrogen, Carlsbad, CA, USA) and oligo-dT primers. Single-tranded cDNA was amplified 30–40 cycles as described previ-usly with Expand High Fidelity PCR System thermostable DNAolymerase mixture (Boehringer Mannheim). PCR was conductedsing primer templates specific to pro-somatostatin transcriptsith high interspecies homology (Table 1). The primers targetequences on different exons of the somatostatin gene to facilitatedentification of products from genomic origins. The expectedmplified cDNA products were resolved on 1.6% agarose gels,tained with ethidium bromide and viewed under UV illumination.erification of the PCR products was performed using diagnostic

estriction digest analyses and direct sequencing of the amplifiedroducts. Routine controls included cDNA synthesis in the ab-ence of either RNA or reverse transcriptase; amplification withmission of template, primers or DNA polymerase failed to yieldroducts.

Quantitative PCR for somatostatin was performed as de-cribed previously (Girard et al., 2002). Briefly, amplified guinea-ig somatostatin DNA product from SST4/SST6 primers was li-ated into pCR2.1 TOPO using TOPO TA cloning kit (Invitrogen)o generate plasmids for peptide transcript standards. The nucle-tide sequences of the inserts were verified by automated fluo-escent dideoxy dye terminator sequencing (Vermont Cancerenter DNA Analysis Facility) and 10-fold serial dilutions of stocklasmid were prepared to generate assay standard curves. Foreal-time quantitative PCR, generation of the guinea-pig cardiacDNA was performed using random hexamers for primers asescribed previously (Girard et al., 2002). Amplification of theuinea-pig cardiac cDNA templates and plasmids standards wasarried out using SYBR Green JumpStart Taq Ready Mix (Sigma)ontaining 3.5 mM MgCl2, 200 �M dATP, dGTP, and dCTP,00 �M dTTP, 0.64 U TaqDNA polymerase and 300 nM of eachrimer in a final volume of 25 �l. Real-time quantitative PCR waserformed on an ABI Prism 7700 Sequence Detector (Appliediosystems, Norwalk, CT, USA) after optimization of cycling pa-

ameters which included SYBR Green I melting point analysesGirard et al., 2002). Data were analyzed at the termination of

Location Ta Product size(bp)

J00787 540U2562 476

J00787 1610L27J00787 540U25

63 205J00787 1340L26J00787 540U25

63 285J00787 1424L22J00787 540U25

62 329J00787 1465L25J00787 637U21

63 257J00787 1494L21

gicus somatostatin precursor protein gene sequence with GenBankwith other species and span intron 1 within the coding region of the

-3�

C-3�

-3�

-3�

-3�

-3�

3�

tus norveomology

ach assay using Sequence Detector 1.7a software (Applied Bio-

spw�itNS

Ptn

Analr1suemcsi4tt

nmptae

agssprmtprnPnn

Pvs

Pgp

Ffcse

K. M. Braas et al. / Neuroscience 126 (2004) 335–346338

ystems). The increase in dye fluorescence intensity (�Rn) waslotted as a function of cycle number and the threshold cycle (CT)as determined by the software as the amplification cycle at whichRn first intersected the default baseline setting. Transcript levels

n each sample were calculated from the CT by interpolation fromhe standard curve to yield the relative change in expression.ormalization of the samples was done using 18S quantification.tatistical analyses were performed using one-way ANOVA.

RESULTS

ACAP27 produced a time-dependent decrease inhe number of cardiac ganglia somatostatin-IReurons

large fraction of postganglionic guinea-pig cardiac gangliaeurons express somatostatin. Quantitated as the percent-ge of ChAT-immunopositive neurons, 60%�1% of the cho-

inergic neurons in freshly dissected cardiac ganglia prepa-ations also exhibited somatostatin-immunoreactivity (Fig.A; n�4). A substantial population of cardiac gangliaomatostatin-containing neurons can be maintained in vitro;nder the current conditions, greater than 40% of the cholin-rgic neurons in control explant preparations expressed so-atostatin after 72 h of culture (Fig. 1B). Treatment of the

ardiac ganglia explants with 20 nM PACAP27 for 24 to 72 hignificantly decreased the percentage of ChAT-mmunoreactive neurons expressing somatostatin from1%�3% in untreated controls (n�3) to 6%�3% (n�5) in

he 72 h PACAP-treated preparations (Fig. 2). Representa-

ig. 1. PACAP27-induced a decrease in the number of cardiac ganglor ChAT- and somatostatin-IR neurons in freshly dissected control (Aontrol preparations, virtually all cardiac neurons exhibited ChATomatostatin-IR (A2). After 72 h in explant culture, many cardiac ganxplant preparations exposed to 20 nM PACAP27, very few ChAT-IR

ive photomicrographs of explant cultured cardiac ganglia e

eurons immunocytochemically processed for ChAT and so-atostatin are shown in Fig. 1B and 1C. In control explantreparations, there were numerous ChAT-positive neurons

hat also stained for somatostatin (Fig. 1B); by contrast, onlyfew neurons in explants chronically treated with PACAP27

xhibited somatostatin-immunoreactivity (Fig. 1C).The effect of PACAP27 was not simply related to its

bility to depolarize the neurons. Treatment of cardiacanglia explants with substance P for 72 h did not alteromatostatin expression. Substance P was previouslyhown to produce a comparable depolarization of guinea-ig cardiac neurons through activation of NK3 tachykinineceptors (Hardwick et al., 1995). Following the 72 h treat-ent with 20 nM substance P, the percentage of soma-

ostatin/ChAT-immunoreactive neurons in the explantreparations was 38%�3% (n�3), a percentage compa-able to untreated control levels. Thus, the decreasedumber of cardiac ganglia somatostatin neurons inACAP27 was not a nonspecific effect shared by othereuropeptides capable of modulating guinea-pig cardiaceuronal function.

ACAP27 was more effective than PACAP38 orasoactive intestinal polypeptide (VIP) in modulatingomatostatin expressing neurons

ACAP peptides are members of the VIP, secretin, andlucagon peptide superfamily and the precursor molecule isosttranslationally processed in a tissue-specific manner to

statin-IR neurons. Guinea-pig cardiac ganglia were dually processedxplants (B), and 72 h explants treated with 20 nM PACAP27 (C). Inand approximately 60% of the ChAT-IR neurons also exhibited-IR neurons (B1) retained somatostatin-IR (B2). By contrast, in 72 h(C1) were also somatostatin-IR (C2). Scale bar�50 �m.

ia somato), 72 h e

-IR (A1)glia ChATneurons

ither of the two -amidated PACAP38 or PACAP27 peptide

f1pigVP1estrP

(cuPtt

tpPtPtastwPcn(dnt

Ps

Ceegcwvwqmmgtt

Fs(nmnwerc

FCGwe

K. M. Braas et al. / Neuroscience 126 (2004) 335–346 339

orms (Miyata et al., 1989; Kimura et al., 1990; Arimura,998). PACAP actions are mediated by at least three G-rotein coupled PACAP receptors: the PAC1 receptor, which

s PACAP-selective and demonstrates a 100- to 1000-foldreater affinity for PACAP than VIP; and the VPAC1 andPAC2 receptors, which share similar high affinities forACAP and VIP peptides (May and Braas, 1995; Arimura,998; Braas and May, 1999; Vaudry et al., 2000; McCullocht al., 2002). We tested whether the PACAP27-induced re-ponse was mediated, at least in part, through activation ofhe PAC1 receptor. The population of somatostatin-immuno-eactive neurons in cardiac ganglia explants co-treated withACAP27 (20 nM) and the PACAP antagonist PACAP(6-38)

ig. 2. PACAP27 produced a time-dependent decrease inomatostatin-IR neurons in explanted cardiac ganglia preparations.�) After 72 h in culture without PACAP, approximately 41% of theeurons also exhibited immunoreactivity for somatostatin as deter-ined from the relationship: (somatostatin-positive/ChAT-positiveeurons)�100�%. (● ) For other guinea-pig cardiac ganglia treatedith 20 nM PACAP27 for the times shown, the percentage of neuronsxhibiting somatostatin-IR progressively decreased. Each data pointepresents mean�S.E.M. for results averaged from at least threeardiac ganglia whole mount preparations.

ig. 3. PACAP27 was more effective than either PACAP38 or VIPardiac ganglia explants were treated with 20 nM PACAP27 in theuinea-pig cardiac ganglia explants were treated with PACAP27 (P27ere dually processed for ChAT- and somatostatin-IR; the number of

xamined�100. Data represent mean�S.E.M. of at least three cardiac ganglia

100 nM or 1 �M) for 72 h was 40%�3% (n�5) of allholinergic neurons, which was similar to that observed inntreated control preparations (41%�3%; n�3; Fig. 3A).ACAP(6-38) alone had no effects; the percentage of soma-

ostatin-immunoreactive neurons after 72 h in explant cul-ures treated with PACAP(6-38) alone was 46%�4% (n�3).

Unlike many PAC1 receptor expressing tissues in whichhe two PACAP peptide forms demonstrate near equal highotency and receptor affinity, the guinea-pig cardiac neuronAC1 receptors exhibited significant selectivity not only be-

ween PACAP and VIP, but also between PACAP27 andACAP38 (Braas et al., 1998). PACAP27 more potently al-

ered cardiac ganglion membrane excitability than PACAP38nd VIP (Braas et al., 1998). Accordingly, in the presenttudy, we examined whether the same peptide potency rela-ionships in altering ganglion cell somatostatin expressionere evident. Treatment of the explants with 20 nMACAP27, PACAP38 or VIP for 72 h all resulted in a de-rease in the population of somatostatin-immunoreactiveeurons; however, PACAP27 was significantly more effectiveFig. 3B). Treatment with 200 pM PACAP27 consistentlyecreased the percentage of somatostatin-immunoreactiveeurons to less than 20% and 20 nM PACAP27 decreased

he population of stained neurons to less than 10%.

ACAP27 treatment did not alter cardiac ganglia pro-omatostatin mRNA levels

ellular peptides in neuroendocrine cells may reflect eitherndogenous production or exogenous peptide uptake. Tostablish endogenous somatostatin production in cardiacanglia neurons, pro-somatostatin mRNA expression inardiac ganglia neurons was evaluated. Total RNA inhole mount ganglion preparations was prepared for re-erse transcription PCR. As the oligonucleotide primersere designed specifically to target somatostatin se-uences on different exons, analyses of the amplified frag-ents allowed the discrimination of cDNA products fromaterial of genomic origins (Fig. 4). Amplification ofuinea-pig cardiac ganglia cDNA for somatostatin usinghe various primer pairs produced the anticipated products;he same products for somatostatin were amplified from

ting the population of cardiac ganglia somatostatin-IR neurons. (A)or absence of the PACAP antagonist PACAP(6-38) [P(6-38)]. (B)

P38 (P28) or VIP at the concentrations shown for 72 h. The explantsith somatostatin-IR was expressed as a fraction of ChaT-IR neurons

in modulapresence), PACA

neurons w

whole mount preparations.

goeaogptcaosd

c

PossltqtpRcweds

F(ssisB

Fpspmns

K. M. Braas et al. / Neuroscience 126 (2004) 335–346340

uinea-pig brain and rat superior cervical ganglion to dem-nstrate transcript organizational identity among the differ-nt tissues. Diagnostic restriction endonuclease analysesnd direct sequencing of the PCR band verified the identityf the amplified product (Fig. 4). The region within theuinea-pig transcript encoding the mature somatostatineptide contained six to seven degenerate base substitu-ions (93% nucleotide identity) compared with other spe-ies resulting in the preservation of all of the peptide aminocid residues (Fig. 5). Larger amplified products indicativef potential amplification of genomic DNA were not ob-erved; PCR controls as described in Experimental Proce-ures all failed to produce amplified products.

Changes in tissue neuropeptide expression are typi-ally well correlated with cellular transcript levels. As

ig. 4. Guinea-pig cardiac ganglia explant preparations expressed sCG), guinea-pig brain or rat superior cervical ganglion (SCG) preparaet shown was used for PCR as described in methods. Shading withhading with accompanying explanatory label in the resulting transcdentification of product derived from genomic origins. The sites of theets amplified one unique product from the different tissues. Represen(SST4/SST5) and C (SST4/SST3) are shown and the correspondin

ig. 5. Guinea-pig somatostatin is homologous to the other mammaliaroducts were purified, ligated into pCR-Script cloning vector for transfequencing was performed in both directions using the T3 and T7 porcellus) transcript encoding the mature somatostatin peptide is aligneouse (Mus musculus), dog (Canis familiaris), rat (Rattus norvegicus)ucleotides that differ from consensus are shown in bold red. The

ubstitutions were degenerate resulting in conservation of all amino acid resid

ACAP27 appeared to decrease dramatically the numberf somatostatin-containing neurons, the somatostatin tran-cripts were anticipated to decrease concomitantly in theame 72 h chronic peptide treatment paradigm. Since the

imiting amounts of cardiac ganglia tissue precluded quan-itative mRNA assessments by Northern analyses, a semi-uantitative PCR approach, to discern relative changes inissue somatostatin transcript levels during the logarithmichase of the amplification process, was used initially. TotalNA was analyzed from two groups of four individualardiac ganglia preparations; four explant preparationsere not treated and represented control, and another fourxplants were treated with 20 nM PACAP27. From twoifferent primer sets, a single product was amplified in allamples from each group of explants (Fig. 6). For each

tin transcripts. Total RNA from individual guinea-pig cardiac gangliareverse-transcribed, and the prosomatostatin oligonucleotide primer(black background) and exon 2 (white background) corresponds tooligonucleotide primer was targeted for different exons to obviate

within the prosomatostatin gene and mRNA are identified. All primeridium bromide-stained agarose gels from primer sets A (SST4/SST6),t sizes are indicated.

Guinea-pig cardiac ganglia cDNA was amplified for somatostatin; theinto competent cells. Automated fluorescence dideoxy dye terminators. The nucleotide sequence for the region of the guinea-pig (Caviaalogous regions of the monkey (Macaca mulatta), bovine (Bos taurus),Ovis aries), human (Homo sapiens) and pig (Sus scrofa). Guinea-pigamino acid sequences for somatostatin are show above. All base

omatostations wasin exon 1ript. Eachprimers

tative ethg produc

n forms.ormationrimer sited with an, sheep (predicted

ues encoding somatostatin among species.

adpP

ndlSaa

tctasimsspC

Fgdpi

FupalputaivA

K. M. Braas et al. / Neuroscience 126 (2004) 335–346 341

ssay, the relative expression of pro-somatostatin mRNAid not appear significantly different among individual sam-les within each group, or between the control andACAP27-treated groups (Fig. 6).

As these results appeared inconsistent with the immu-ocytochemical data, a quantitative PCR approach waseveloped to evaluate changes in somatostatin transcript

evels. To demonstrate rigorously the specificity of theST4/SST6 primer set under the assay conditions, themplified product was subjected to SYBR Green I meltingnalyses (Fig. 7B). Upon binding to amplified dsDNA,

ig. 6. PACAP27 did not alter cardiac ganglia prosomatostatin transcanglia explant preparations maintained in the presence or absence ofifferent oligonucleotide primers sets specific for prosomatostatinrosomatostatin mRNA levels between the 72 h untreated (vehicle) a

ndependent explant preparation.

ig. 7. Real-time quantitative PCR of guinea-pig cardiac ganglia somantreated control and PACAP27-treated cardiac ganglia explants wererimers SST4/SST6 under optimized conditions as described in Expmplification of templates from a single explant, and the CT for each g

ight green, orange and yellow tracings; PACAP27: pink, blue and deeprimers on guinea-pig templates was assessed by SYBR Green I melnique DNA dissociation curve at 89.8 °C. The amplification parameteemplate control. (C) Wells containing 10-fold dilutions of the somatosta

standard curve. (D) The CT for each plasmid dilution was determinnversely related to plasmid concentration, the color for each CT pointalues from amplification tracings between control (�) and PACAP27

NOVA.

here was a large enhancement of SYBR Green fluores-ence. After completion of the amplification reaction, theemperature of the reaction sample was ramped to 95 °Cnd the melting of the amplified DNA product at a definedpecific temperature was accompanied by a sharp decline

n sample fluorescence (Fig. 7B). A single unique DNAelting profile was produced indicating amplification of a

ingle product and the validity of the assay. When theame amplification parameters were applied to templatesrepared from control and PACAP27-treated explants, the

T values from the two sample groups did not segregate

in explant preparations. Total RNA extracted from individual cardiac7 for 72 h was reverse transcribed, and the cDNA was amplified usings described in Fig. 4. Under linear PCR amplification conditions,

PACAP27 treated explants were similar. Each lane represents an

ranscripts. (A) Complementary DNA templates synthesized from 72 hfor somatostatin transcript levels by real-time quantitative PCR using

l Procedures. Each fluorescence intensity (�Rn) tracing represents3) were determined using the Sequence Detector software. Vehicle:) The uniqueness of the amplicon synthesized using the somatostatinsis. Ramping of the temperature from 65 to 95 °C produced a singlet generate anomalous products or primer dimers as shown for the noard plasmid were assayed with the experimental samples to generatexpressed as a function of initial plasmid quantity. As CT values are

responds to the tracing of the same color in (C). Normalized mean CT

(■ ) explant samples in (A) were not statistically different by one-way

ript levelsPACAP2mRNA and 20 nM

tostatin tanalyzederimentaroup (n�green. (B

ting analyrs did notin standed and e

value cor-treated

bbtaool7sduts

Tir

WscawpstscnpbtsitcdaVtcmusstnse(PamPm1T(fs

iPptaPc

Pt

Neaicamptnda4(v

Fsi(SUcPTid

K. M. Braas et al. / Neuroscience 126 (2004) 335–346342

ut clustered between 28 and 31 cycles (Fig. 7A). This wasetter appreciated when the CT data were plotted againsthe standard curve (Fig. 7D). The standard curve, gener-ted by plotting the starting plasmid material as a functionf CT, demonstrated a dynamic range spanning five ordersf magnitude with high amplification efficiency and corre-

ation coefficient (slope, 3.7; corr. coefficient, �0.99; Fig.C and 7D). As with the semiquantitative PCR results,tatistical analyses of the quantitative data failed to revealifferences in pro-somatostatin mRNA levels between thentreated control and PACAP27-treated explant prepara-

ions despite a near seven-fold decrease in the number ofomatostatin-immunoreactive neurons.

he PACAP-induced decrease in somatostatin-mmunoreactive neurons was mediated by specificeceptor signaling pathways

hile the decrement in the number of somatostatin-taining neurons following PACAP treatment without aorresponding decrease in somatostatin mRNA levelsppeared disparate, these observations were consistentith experimental and physiological studies for othereptidergic systems where stimulated bioactive peptideecretion greatly exceeded cellular peptide biosynthe-ic rates. Hence the PACAP-induced decrease inomatostatin-containing neurons may not reflecthanges in the population of somatostatin expressingeurons, but rather a depletion of neuronal dense coreeptidergic granules due to enhanced secretion to levelselow immunocytochemical detection. The PAC1 recep-

or belongs to the superfamily of G-protein-coupledeven transmembrane receptors that initiate their phys-ological functions through the activation of diverse in-racellular signaling cascades, including adenylyl cy-lase, PLC and MEK. The signaling cascade activated isependent on the specific receptor PAC1 receptor vari-nts expressed (Arimura, 1998; Braas and May, 1999;audry et al., 2000; McCulloch et al., 2002), and all

hree pathways have been associated with cellular se-retory processes. Accordingly, experiments using phar-acological second messenger pathway inhibitors weresed to determine whether the PACAP-induced re-ponse was dependent on activation of one of theseignaling cascades. Inhibition of cAMP production withhe adenylyl cyclase inhibitor SQ223356 (50 �M) elimi-ated the PACAP27-induced decrease in percentage ofomatostatin-immunoreactive neurons (Fig. 8); how-ver, inhibition of the downstream PKA with H8925 �M) had no apparent effect. Similarly, inhibition ofLC/protein kinase C signaling with U73122 (10 �M)lso failed to attenuate the decrease in the PACAP27-ediated somatostatin response. More recently, theAC1 receptor has been shown to be coupled to theitogen-activated protein kinase pathway (Arimura,998; Vaudry et al., 2000; McCulloch et al., 2002).reatment with the potent MEK1 inhibitor PD9805950 �M) to block downstream ERK phosphorylation ef-ectively blocked the PACAP27-induced decrease in

omatostatin-immunoreactive neurons (Fig. 8) suggest- P

ng involvement of ERK intracellular signaling in theACAP effect. PLC/protein kinase C signaling did notarticipate in the PACAP-induced decrease in soma-

ostatin positive neurons, an observation consistent withpotential preferential coupling of the guinea-pig

AC1(veryshort) null (neither HIP nor HOP cassetteontaining) receptor to the cAMP signaling pathway.

ACAP27-induced response requires calcium influxhrough L-type channels

eurotransmitter/neuropeptide release requires influx ofxternal calcium through voltage-gated calcium channelsnd MEK signaling has been implicated in modulating

ntracellular calcium levels to affect optimal regulated se-retion (Weck et al., 1998; Gomez et al., 2002). To evalu-te the role of extracellular calcium influx in the PACAP-ediated somatostatin response, cardiac ganglia explantreparations treated with 20 nM PACAP27 were main-ained for 72 h in a calcium-deficient media in which mag-esium was substituted for calcium (Fig. 9). In the calcium-eficient solution, PACAP27 did not decrease the percent-ge of somatostatin neurons (calcium-deficient media,3%�1%, n�3) compared with control cultures41%�3%, n�3), consistent with requirements for ele-ated intracellular calcium levels and secretion for the

ig. 8. PACAP27-mediated decrease in guinea-pig cardiac gangliaomatostatin expression is blunted by adenylyl cyclase and MEKnhibitors. Cardiac ganglia explants were treated with 20 nM PACAP27P27) for 72 h in the absence or presence of adenylyl cyclase inhibitorQ23356 (SQ, 50 �M), PKA inhibitor H89 (25 �M), PLC inhibitor73122 (U7, 10 �M) or MEK inhibitor PD98059 (PD, 50 �M). Theulture media was changed every 24 h for control (veh) andACAP27-only explants, and every 12 h when inhibitors were used.he explants were fixed after 72 h for immunocytochemical process-

ng. Data represent mean�S.E.M. from three to five ganglia. Asterisksenote statistical difference from control at P�0.05 by ANOVA.

ACAP effects. As calmodulin commonly is an intracellular

mwP7u2ietate

vitttsgb2s

Sgcgscngchefi(tt

aTispwlccc

dPnnassptngparcisctsIPsaPtov

FGoim

K. M. Braas et al. / Neuroscience 126 (2004) 335–346 343

ediator of calcium actions, additional experiments testedhether calmodulin activation was required for theACAP27-induced responses (Fig. 9). Co-treatment of2 h PACAP-treated explant preparations with the calmod-lin antagonist W-13 (70 �M), blocked completely the0 nM PACAP27-induced decrease in somatostatin-

mmunoreactive neurons to control untreated explant lev-ls (54%�6%, n�3). In aggregate, these results indicatedhat calcium influx and concomitant calcium/calmodulinctivation of target enzymes were essential elements forhe PACAP-mediated decrease in somatostatinxpression.

Mammalian cardiac neurons express a number ofoltage-gated calcium channels which can affect calciumnflux and potentially induce neurotransmitter/neuropep-ide secretion (Jeong and Wurster, 1997). Experimentsested whether activation of a specific calcium channelype was required for the PACAP-induced modulation ofomatostatin expression. Co-treatment of 72 h cardiacanglia explant preparations with nifedipine (10 �M) tolock L-type calcium channels completely eliminated the0 nM PACAP27-induced changes in the cardiac gangliaomatostatin population (46%�7%, n�3; Fig. 9).

DISCUSSION

omatostatin is highly expressed in guinea-pig cardiacanglia neurons. Previous studies, based on an analysis ofolchicine-treated cardiac ganglia preparations, have sug-ested that approximately 80% of the cardiac neurons areomatostin-IR (Steele et al., 1994). In good agreement, oururrent results estimate that more than 60% of the neuro-al population in freshly dissected guinea-pig cardiac gan-lia preparations expresses somatostatin. Prior immuno-ytochemical studies have demonstrated innervation ofeart muscle by somatostatin-immunopositive fibers (Dayt al., 1985). Recent studies indicate the highest density ofber staining occurs around cardiac valves and nodesSteele and Choate, 1994; Steele et al., 1996). Somatosta-in receptors are present in guinea-pig heart and soma-

ig. 9. PACAP-mediated somatostatin responses in guinea-pig cardiauinea-pig cardiac ganglia explants were treated with 20 nM PACAP2r L-type calcium channel blocker nifedipine (nif, 10 �M, C) for 72 h,

n magnesium-supplemented, calcium-deficient medium (B). The explaean�S.E.M. from three ganglia explants. Asterisks denote statistica

ostatin can decrease voltage-gated calcium currents in e

trial myocytes (Feniuk et al., 1993; Ohmura et al., 1999).ogether, the immunocytochemical and electrophysiolog-

cal findings suggest that somatostatin released from para-ympathetic postganglionic cardiac nerves might partici-ate in the regulation of atrial contractile function alongith acetylcholine. Thus, we hypothesize that PACAP re-

eased from preganglionic terminals during vagal activityould potentially facilitate postganglionic neuron acetyl-holine and somatostatin secretion for concerted amplifi-ation of inhibitory activities on cardiac tissues.

Our previous studies have shown that guinea-pig car-iac ganglia express high levels of the PACAP-selectiveAC1 receptor, and that PACAP27 peptides depolarizeeurons and increase neuronal membrane excitability inearly all postganglionic cardiac ganglia neurons (Braas etl., 1998). In sympathetic neurons, PACAP has beenhown to modulate NPY and tyrosine hydroxylase expres-ion (May and Braas, 1995; Braas and May, 1999). Theserior observations led us to evaluate the ability of PACAPhrough PAC1 receptor signaling cascades to modulateumbers of somatostatin-IR neurons in guinea-pig cardiacanglia. As only a few hundred cardiac neurons areresent in the atrial whole mounts, direct neuropeptidessay measures of cardiac ganglia somatostatin content/elease proved untenable. Therefore, immunocytochemi-al techniques provided an appropriate means of assess-

ng the changes in cardiac neuron somatostatin expres-ion. The present study demonstrated that exposure of theardiac ganglia explants with PACAP caused a concentra-ion- and time-dependent decrease in the percentage ofomatostatin-IR neurons within guinea-pig cardiac ganglia.n addition, PACAP27 was more effective than eitherACAP38 or VIP and the PACAP27-induced decrease inomatostatin-IR neurons was eliminated by the PACAPntagonist PACAP6-38. We suggest that activation ofAC1 selective receptors mediated much of the soma-

ostatin response, which is consistent with the results ofur earlier studies demonstrating that PAC1 receptor acti-ation mediates increases in guinea-pig cardiac neuron

require extracellular calcium influx through L-type calcium channels.n the absence or presence of calmodulin antagonist W-13 (70 �M, A)bed in Fig. 7. Additional PACAP27-treated explants were maintainedfixed and processed for somatostatin immunoreactivity; data representce from control at P�0.05 by ANOVA.

c ganglia7 (P27) ias descrints werel differen

xcitability (Braas et al., 1998). However, as we have not

tpPh

PCcrsimsPntncotcctieto

ta1ePVtssgtntgsclcnctfagsroacch

u1aadiiPserngd

aLPcgnwccliTsiPsmtespwrc1rsec

Pmr(caccPrtgw

K. M. Braas et al. / Neuroscience 126 (2004) 335–346344

ested for the presence of VPAC receptors on the guinea-ig cardiac neurons, it remains possible that some of theACAP effect on somatostatin immunoreactivity couldave been mediated through activation of VPAC receptors.

Chronic treatment of the cardiac ganglia explants withACAP27 decreased somatostatin-IR neurons by 85%.ellular peptide expression levels are typically correlatedlosely with transcript levels, but in parallel explant prepa-ations, the dramatic decrease in cardiac gangliaomatostatin-IR neurons was not mirrored by decrements

n ganglia pro-somatostatin transcript expression by PCRethods. In fact, quantitative PCR may have suggested a

light increase in somatostatin transcript expression in theACAP-treated explants, but overall, the changes wereot statistically different. Total cellular neuropeptide con-

ent, as measured by either radioimmunoassay or immu-ocytochemical approaches, reflects the net change inellular biosynthetic and secretory rates. Accordingly, thebserved decrease in cardiac ganglia somatostatin pep-

ide levels following PACAP treatment without apparenthanges in transcript expression, may result from de-reased prosomatostatin biosynthesis and posttransla-ional processing, enhanced somatostatin degradation orncreased PACAP-stimulated cellular somatostatin releasexceeding the rate of somatostatin production, all leading

o diminished cytosolic somatostatin levels below thresh-ld for immunocytochemical detection.

As for many peptidergic systems, somatostatin produc-ion can be upregulated by transcriptional (Montminy etl.1996) and/or translational mechanisms (Spiegel et al.,990; Rage et al., 1994). However, as second messengersspecially cAMP stimulate somatostatin expression, andACAP, a potent activator of cellular cAMP production atPAC/PAC receptors, produced an apparent decrease in

he number of somatostatin expressing neurons, we con-idered PACAP-mediated somatostatin secretion to repre-ent the primary mechanism for our results. Theuinea-pig cardiac ganglia neurons express predominantlyhe PAC1veryshort null receptor isoform with neither HIPor HOP cassette inserts in the third cytoplasmic loop ofhe G-protein-coupled receptor (Braas et al., 1998). Theuinea-pig PAC1 receptor variant appears to demonstratepecificity for PACAP27 over PACAP38, and is potentiallyoupled to both adenylyl cyclase and PLC. It is well estab-ished that the somatostatin gene is positively regulated byAMP/PKA-mediated CREB Ser133 phosphorylation in aumber of different endocrine and neuronal cell types, butan be tempered by the concomitant activity of phospha-ases, including PP1, in the removal of phosphate groupsrom CREB to mitigate the activation process (Montminy etl., 1996). With these characterizations of the somatostatinene promoter, the apparent stability in somatostatin tran-cripts levels after PACAP/PAC1 receptor signaling mayeflect these dephosphorylation and inactivation eventsver the long PACAP-treatment periods. By contrast, bothdenylyl cyclase and/or PLC signaling can elicit dramaticellular neuroendocrine hormone and peptide release. Cy-lic AMP/PKA or PLC/PKC activation in islet cells results in

igh levels of somatostatin release that can be synergistic d

pon stimulation of both signaling pathways (Patel et al.,991). Enhanced secretory processes can result in anpparent decrease in cellular hormone content. Potentctivation of the ACTH system by CRH for example, elicitsramatic ACTH release resulting in an apparent decrease

n anterior pituitary corticotrope population both in vivo andn vitro (Westlund et al., 1985; May and Eipper, 1986).ACAP stimulation of catecholamine and neuropeptideecretion has been documented in many neuronal andndocrine cells, and PACAP-induced sympathetic NPYelease results in an apparent decrement in sympatheticeuronal NPY content (May and Braas, 1995). An analo-ous process appears to mediate the PACAP-inducedecrease in somatostatin positive cardiac neurons.

Mammalian parasympathetic cardiac neurons expressnumber of voltage-gated calcium channels including the

-, N-, Q- and R-subtypes (Jeong and Wurster, 1997). TheACAP-induced decrease in somatostatin-IR neurons wasompletely blocked following the addition of nifedipine sug-esting that calcium influx through L-type calcium chan-els, rather than other voltage-gated calcium channels,as essential for the PACAP effect. This observation isonsistent with other studies that demonstrate L-type cal-ium channel involvement in PACAP peptide-induced re-

ease of catecholaminergic and peptidergic containing ves-cles (Martı́nez-Fuentes et al., 1998; Shibuya et al., 1998;aupenot et al., 1998; Fukushima et al., 2001). Whetherufficient L-type channels were opened by a PACAP-

nduced depolarization or activated by channel coupling toACAP receptors cannot be established from the presenttudy. However, a PACAP-induced depolarization aloneay not be sufficient as substance P, which also consis-

ently depolarizes guinea-pig cardiac neurons to the samextent (Hardwick et al., 1995), did not affect the number ofomatostatin-IR neurons in 72 h explant cardiac gangliareparations. Recently, a unique PAC1 receptor variantith a mutation in the transmembrane domain (PAC1 TM4

eceptor) has been described that directly gates L-typealcium channels in pancreatic � islets (Chatterjee et al.,996). The mechanisms underlying the expression andegulation of the PAC1 TM4 receptor have not been welltudied, and whether this particular receptor variant isxpressed in guinea-pig cardiac neurons for L-type cal-ium channels activation remains to be established.

From other signaling inhibitor studies, the observedACAP-mediated somatostatin responses appeared to beore consistent with the actions of the PAC1 null or HOP

eceptor variants expressed in guinea-pig cardiac neuronsBraas et al., 1998). These PAC1 receptors are coupledommonly to multiple intracellular signaling cascades (Braasnd May, 1999; Vaudry et al., 2000; McCulloch et al., 2002);onsequently, experiments tested whether activation of spe-ific second messenger cascades was a requisite for theACAP-induced decrease in somatostatin-IR cardiac neu-

ons. Although inhibition of adenylyl cyclase activity blockedhe PACAP-mediated decrease in the number of cardiacanglia somatostatin neurons, the addition of PKA inhibitorsas ineffective suggesting that the response was cAMP-

ependent but PKA-independent. Furthermore, the addition

omi

aMPTtcpAitncGtasicCcecrLnrpc

cniPwrcoPget

ANAts

A

A

A

B

B

C

C

C

C

D

D

D

F

F

G

G

G

H

H

J

K

K

K. M. Braas et al. / Neuroscience 126 (2004) 335–346 345

f PLC inhibitor U73122 had no effect suggesting that ele-ents of the phosphatidylinositol pathway were not involved

n the PACAP effect on somatostatin immunoreactivity.PAC1 receptors also can be coupled to the mitogen

ctivated kinase signaling cascade and treatment with theEK kinase inhibitor PD98059 effectively eliminated theACAP-induced decrease in somatostatin-IR neurons.he PAC1 receptor and MAPK activation could be coupled

hrough a number of signaling cascades including theAMP/GEF/Rap/MEK or Ca2�/CaM/RasGRF/Ras/MEKathway (Agell et al., 2002; Stork and Schmitt, 2002).lthough MAPK activation was required for the PACAP-

nduced decrease in somatostatin-IR neurons, results ofhe present study do not delineate the potential mecha-isms of MAPK involvement. However, MAPK has bothytoplasmic and nuclear targets (Sugden and Clerk, 1997;rewal et al., 1999). Phosphorylated and activated MAPK

ranslocation into the nucleus is known to be coupled toctivation of transcription factors to stimulate gene expres-ion; cytosolic phosphorylation targets of MAPK, whichnclude other enzymes, cytoskeletal elements and ionhannels, may regulate protein function (Sugden andlerk, 1997; Grewal et al., 1999). L-type calcium channelonduction properties can be regulated by phosphorylationvents (Catterall, 2000). Cytokines enhance cardiac myo-yte L-type calcium channel through MAPK signaling (Mu-ata et al., 1999); PACAP has been shown to facilitate-type calcium channel conductance in suprachiasmaticeurons in a MAPK-dependent manner (Dziema and Ob-ietan, 2002). Whether L-type calcium channels are directhosphorylation targets of MAPK activity in guinea-pigardiac neurons remains to be established.

In summary, PACAP27 potently produced a time andoncentration decrease in the percentage of somatostatin-IReurons in explanted guinea-pig cardiac ganglia. We suggest

n aggregate with our previous studies that this effect ofACAP was mediated through the PAC1 receptor, althoughe have not excluded potential contributions from VPAC

eceptor signaling. In addition, the effect of PACAP requiredalcium influx through L-type calcium channels and activationf adenylyl cyclase and the MAPK signaling pathway. Thus,ACAP peptides modulate the neurochemical profile ofuinea-pig cardiac neurons, suggesting that PACAP mayxert trophic effects on these parasympathetic neuronshrough activation of specific intracellular signaling pathways.

cknowledgements—This work was supported by NIH grantsS23978, NS37179, HD27468 and HL65481 and AHA Grant-In-ids 94615540 and 975043N. We thank Kristin Schutz for expert

echnical assistance and Dr. Allison Buchan for providing theomatostatin antiserum.

REFERENCES

gell N, Bach O, Rocamora N, Villalonga P (2002) Modulation of theRas/Raf/MEK/ERK pathway by Ca(2�), and calmodulin. Cell Signal14:649–654.

rdell JL (2001) Neurohumoral control of cardiac function. In: Heartphysiology and pathophysiology (Sperelakis N, ed), pp 45–59. SanDiego, CA: Academic Press.

rimura A (1998) Perspectives on pituitary adenylate cyclase activat-

ing polypeptide (PACAP) in the neuroendocrine, endocrine, andsystems. Jpn J Physiol 48:301–331.

raas KM, May V, Harakall SA, Hardwick JC, Parsons RL (1998)Pituitary adenylate cyclase-activating polypeptide expression andmodulation of neuronal excitability in guinea pig cardiac ganglia.J Neurosci 18:9766–9779.

raas KM, May V (1999) Pituitary adenylate cyclase-activatingpolypeptides directly stimulate sympathetic neuron neuropeptide Yrelease through PAC(1) receptor isoform activation of specific in-tracellular signaling pathways. J Biol Chem 274:27702–27710.

alupca MA, Vizzard MA, Parsons RL (2000a) Origin of pituitaryadenylate cyclase-activating polypeptide (PACAP)-immunoreactivefibers innervating guinea pig parasympathetic cardiac ganglia.J Comp Neurol 423:26–39.

alupca MA, Vizzard MA, Parsons RL (2000b) Origin of neuronal nitricoxide synthase (NOS)-immunoreactive fibers in guinea pig para-sympathetic cardiac ganglia. J Comp Neurol 426:493–504.

atterall WA (2000) Structure and regulation of voltage-gated Ca2�

channels. Annu Rev Cell Dev Biol 16:521–555.hatterjee TK, Sharma RV, Fisher RA (1996) Molecular cloning of a

novel variant of the pituitary adenylate cyclase-activating polypep-tide (PACAP) receptor that stimulates calcium influx by activation ofL-type calcium channels. J Biol Chem 271:32226–32232.

ay SW, Gu J, Polak JM, Bloom SR (1985) Somatostatin in the humanheart and comparison with guinea pig and rat heart. Br Heart J53:153–157.

iCicco-Bloom E, Deutsch PJ, Maltzman J, Zhang J, Pintar JE, ZhengJ, Friedman WF, Zhou X, Zaremba T (2000) Autocrine expressionand ontogenetic functions of the PACAP ligand/receptor systemduring sympathetic development. Dev Biol 219:197–213.

ziema H, Obrietan K (2002) PACAP potentiates L-type calcium chan-nel conductance in suprachiasmatic nucleus neurons by activatingthe MAPK pathways. J Neurophysiol 88:1374–1386.

eniuk W, Dimach J, Humphrey PP (1993) Characterization of soma-tostatin receptors in guinea-pig isolated ileum, vas deferens andright atrium. Br J Pharmacol 110:1156–1164.

ukushima Y, Nagayama T, Kawashima H, Hikichi H, Yoshida M,Suzuki-Kusaba M, Hisa H, Kimura T, Satoh S (2001) Role ofcalcium channels and adenylate cyclase in the PACAP-inducedadrenal catecholamine secretion. Am J Physiol Regul Int CompPhysiol 281:R495–R501.

irard BM, May V, Bora SH, Fina F, Braas KM (2002) Regulation ofneurotrophic peptide expression in sympathetic neurons: quantita-tive analysis using radioimmunoassay and real-time quantitativepolymerase chain reaction. Reg Peptides 109:89–101.

omez E, Pritchard C, Herbert TP (2002) cAMP-dependent proteinkinase and Ca2� influx through L-type voltage-gated calcium chan-nels mediate Raf-independent activation of extracellular regulatedkinase in response to glucagon-like peptide-1 in pancreatic beta-cells. J Biol Chem 277:48146–48151.

rewal SS, York RD, Stork PJ (1999) Extracellular-signal-regulatedkinase signalling in neurons. Curr Opin Neurobiol 9:544–553.

ansel DE, May V, Eipper BA, Ronnett GV (2001) Pituitary adenylylcyclase-activating peptides and alpha-amidation in olfactory neuro-genesis and neuronal survival in vitro. J Neurosci 21:4625–4636.

ardwick JC, Mawe GM, Parsons RL (1995) Tachykinin-induced ac-tivation of non-specific cation conductance via NK3 neurokinin re-ceptors in guinea-pig intracardiac neurons. J Physiol (Lond) 504:65–74.

eong S-W, Wurster RD (1997) Calcium channel currents in acutelydissociated intracardiac neurons from adult rats. J Neurophysiol77:1769–1778.

ennedy AL, Harakall SA, Lynch SW, Braas KM, Hardwick JC, MaweGM, Parsons RL (1998) Expression and physiological actions ofneuropeptide Y in guinea pig parasympathetic cardiac ganglia. JAuton Nerv Syst 71:190–195.

imura C, Ohkubo S, Ogi K, Hosoya M, Itoh Y, Onda H, Miyata A,

Jiang L, Dahl RR, Stibbs HH, Arimura A, Fujino M (1990) A novel

L

L

L

L

M

M

M

M

M

M

M

M

N

O

P

P

R

R

S

S

S

S

S

S

S

S

T

V

W

W

W

K. M. Braas et al. / Neuroscience 126 (2004) 335–346346

peptide which stimulates adenylate cyclase: molecular cloning andcharacterization of the ovine and human cDNAs. Biochem BiophysRes Commun 166:81–89.

eger J, Croll RP, Smith FM (1999) Regional distribution of andextrinsic innervation of intrinsic neurons in the guinea pig. J CompNeurol 407:303–317.

offelholz K, Pappano AJ (1985) The parasympathetic neuroeffectorjunction of the heart. Pharmacol Rev 37:1–24.

u N, Zhou R, DiCicco-Bloom E (1998) Opposing mitogenic regulationby PACAP in sympathetic and cerebral cortical precursors corre-lates with differential expression of PACAP receptor (PAC1-R)isoforms. J Neurosci Res 53:651–662.

ynch SW, Braas KM, Harakall SA, Kennedy AL, Mawe GM, ParsonsRL (1999) Neuropeptide Y (NPY) expression is increased in ex-planted guinea pig parasympathetic cardiac ganglia neurons. BrainRes 827:70–78.

artı́nez-Fuentes AJ, Castano JP, Gracia-Navarro F, Malagon MM(1998) Pituitary adenylate cyclase-activating polypeptide (PACAP)38 and PACAP27 activate common and distinct intracellular signal-ing pathways to stimulate growth hormone secretion from porcinesomatotrophs. Endocrinology 139:5116–5124.

awe GM, Talmage EK, Lee KP, Parsons RL (1996) Expression ofcholine acetyltransferase immunoreactivity in guinea pig cardiacganglia. Cell Tissue Res 285:281–286.

ay V, Braas KM (1995) Pituitary adenylate cyclase-activatingpolypeptide (PACAP) regulation of sympathetic neuron neuropep-tide Y and catecholamine expression. J Neurochem 65:978–987.

ay V, Eipper BA (1986) Long term culture of primary rat pituitaryadrenocorticotropin/endorphin-producing cells in serum-free me-dium. Endocrinology 118:1284–1295.

cCulloch DA, MacKenzie CJ, Johnson MS, Robertson DN, HollandPJ, Ronaldson E, Lutz EM, Mitchell R (2002) Additional signalsfrom VPAC/PAC family receptors. Biochem Soc Trans 30:441–446.

iyata A, Arimura A, Dahl RR, Minamino N, Uehara A, Jiang L, CullerMD, Coy DH (1989) Isolation of a novel 38 residue-hypothalamicpolypeptide which stimulates adenylate cyclase in pituitary cells.Biochem Biophys Res Commun 164:567–574.

ontminy M, Brindle P, Arias J, Ferreri K, Armstrong R (1996) Regu-lation of somatostatin gene transcription by cAMP. Adv Pharmacol36:1–13.

urata M, Fukuda K, Ishida H, Miyoshi S, Koura T, Kodama H,Nakazawa HK, Ogawa S (1999) Leukemia inhibitory factor, a po-tent cardiac hypertrophic cytokine, enhances L-type Ca2� currentand [Ca2�]i transient in cardiomyocytes. J Mol Cell Cardiol 31:237–245.

icot A, DiCicco-Bloom E (2001) Regulation of neuroblast mitosis isdetermined by PACAP receptor isoform expression. Proc Natl AcadSci USA 98:4758–4763.

hmura T, Nishio M, Kigoshi S, Muramatsu I (1999) Somatostatindecreases the calcium inward current in guinea pig atria. Br JPharmacol 99:587–591.

arsons RL (2004) Mammalian cardiac ganglia as local integrationcenters: histochemical and electrophysiological evidence. In: Neu-ral mechanisms of cardiovascular regulation (Dun N, Machado B,Pilowsky P, eds), pp 335–356. Boston, MA: Kluwer Academic

Publishers.

atel YC, Papachristou DN, Zingg HH, Farkas EM (1991) Regulationof islet somatostatin secretion and gene expression: selective ef-fects of adenosine 3�,5�-monophosphate and phorbol esters innormal islets of Langerhans and in a somatostatin-producing ratislet clonal cell line 1027 B2. Endocrinology 128:1754–1762.

age F, Lazaro JB, Benyassi A, Arancibia S, Tapia-Arancibia L (1994)Rapid changes in somatostatin and TRH mRNA in whole rat hypo-thalamus in response to acute cold exposure. J Neuroendocrinol6:19–23.

andall WC, Wurster RD (1994) Peripheral innervation of the heart. In:Vagal control of the heart: experimental basis and clinical implica-tions (Levy MN, Schwartz PJ, eds), pp 21–32. Armonk, NY: Futura.

eebeck J, Schmidt WE, Kilbinger H, Neumann J, Zimmermann N,Herzig S (1996) PACAP induces bradycardia in guinea-pig heart bystimulation of atrial cholinergic neurones. Naunyn SchmiedebergsArch Pharmacol 354:424–430.

piegel K, Wong V, Kessler JA (1990) Translational regulation ofsomatostatin in cultured sympathetic neurons. Neuron 4:303–311.

teele PA, Choate JK (1994) Innervation of the pacemaker in guinea-pig sinoatrial node. J Auton Nerv Sys 47:177–187.

teele PA, Gibbins IL, Morris JL, Mayer B (1994) Multiple populationsof neuropeptide-containing intrinsic neurons in the guinea-pigheart. Neuroscience 62:241–250.

teele PA, Gibbins IL, Morris JL (1996) Projections of intrinsic cardiacneurons to different targets in the guinea-pig heart. J Auton NervSyst 56:191–200.

tork PJ, Schmitt JM (2002) Crosstalk between cAMP and MAPkinase signaling in the regulation of cell proliferation. Trends CellBiol 12:258–266.

ugden PH, Clerk A (1997) Regulation of the ERK subgroup of MAPkinase cascades through G protein-coupled receptors. Cell Signal9:337–351.

hibuya I, Noguchi J, Tanaka K, Harayama N, Inoue Y, Kabashima N,Ueta Y, Hattori Y, Yamashita H (1998) PACAP increases the cyto-solic Ca2� concentration and stimulates somatodentritic vasopres-sin release in rat supraoptic neurons. J Neuroendocrinol 10:31–42.

aupenot L, Mahata SK, Wu H, O’Connor DT (1998) Peptidergicactivation of transcription and secretion in chromaffin cells-cis andtrans signaling determinants of pituitary adenyl cyclase-activatingpolypeptide (PACAP). J Clin Invest 101:863–876.

audry D, Gonzalez BJ, Basille M, Yon L, Fournier A, Vaudry H (2000)Pituitary adenylate cyclase-activating polypeptide and its receptors:from structure to functions. Pharmacol Rev 52:269–324.

aschek JA (2002) Multiple actions of pituitary adenylyl cyclase acti-vating peptide in nervous system development and regeneration.Dev Neurosci 24:14–23.

eck J, Fallest PC, Pitt LK, Shupnik MA (1998) Differential gonadotropin-releasing hormone stimulation of rat luteinizing hormone subunit genetranscription by calcium influx and mitogen-activated protein kinase-signaling pathways. Mol Endocrinol 12:451–457.

estlund KN, Aguilera G, Childs GV (1985) Quantification of morpho-logical changes in pituitary corticotropes produced by in vivocorticotropin-releasing factor stimulation and adrenalectomy. Endo-

crinology 116:439–449.

(Accepted 6 April 2004)(Available online 8 May 2004)