Proc. Natl. Acad. Sci. USAVol. 93, pp. 10377-10382, September 1996Medical Sciences

A novel nonneuronal catecholaminergic system: Exocrinepancreas synthesizes and releases dopamine

(monoamine transporters/mucosal healing/dopamine receptor)

EVA MEZEY*t, GRAEME EISENHOFER*, GYONGYI HARTA*, STEFAN HANSSONt, LYDIA GOULDt, BELA HUNYADYt§,AND BETH J. HOFFMANt*Clinical Neuroscience Branch, National Institute of Neurological Diseases and Stroke, and tLaboratory of Cell Biology, National Institute of Mental Health,Building 36, 3A17, Bethesda, MD 20892

Communicated by Bernhard Witkop, National Institutes of Health, Bethesda, MD, June 27, 1996 (received for review May 9, 1996)

ABSTRACT Cells of the exocrine pancreas produce di-gestive enzymes potentially harmful to the intestinal mucosa.Dopamine has been reported to protect against mucosalinjury. In looking for the source of dopamine in the smallintestine, we found that the duodenaljuice contains high levelsof dopamine and that the pancreas itself has a high dopamine[and dihydroxyphenylalanine (dopa)] content that does notchange significantly after chemical sympathectomy. Further-more, we were able to demonstrate tyrosine hydroxylase (TH)activity in control pancreas as well as in pancreas from ratsafter chemical sympathectomy. Immunostaining and in situhybridization histochemistry confirmed both the presence ofTH, dopamine, and the dopamine transporter, and themRNAs encoding TH and dopamine transporter, and thepresence ofboth types ofvesicular monoamine transporters inthe exocrine cells of the pancreas. Since there are no cat-echolaminergic enteric ganglia in the pancreas, the aboveresults indicate that pancreatic cells have all the character-istics of dopamine-producing cells. We suggest that the pan-creas is an important source of nonneuronal dopamine in thebody, and that this dopamine has a role in protecting theintestinal mucosa and suggests that dopamine Dlb receptoragonists might be used to help mucosal healing in the gas-trointestinal tract.

Dopamine protects against both gastric and intestinal mucosalinjury. For instance, gastric and duodenal ulcers heal signifi-cantly faster after administration of dopamine agonists (1-4).However, the possible source of dopamine in the gastrointes-tinal system remained to be determined.The exocrine pancreas produces and secretes digestive en-

zymes and bicarbonate and releases them into the duodenum,while endocrine cells in the islets of Langerhans synthesize andrelease hormones (such as insulin, glucagon, etc.) and areembedded in the exocrine pancreas. If dopamine in fact playsa protective role in the duodenum, then corelease with diges-tive enzymes from the exocrine pancreas seems reasonable. Inthe cells that are outside of the central nervous system,dopamine is generally considered to be a precursor of norepi-nephrine and epinephrine. Recent studies in swine (5), how-ever, suggest that the mesenteric organs produce about half oftotal body dopamine and that dopamine in the mesentericorgans must not be exclusively a precursor to norepinephrine.The source of this dopamine is still unknown. We have recentlydiscovered that the acid-secreting parietal cells of the stomachsynthesize and release dopamine into the gastric lumen, whereit may act as a paracrine hormone at dopamine receptors onepithelial cells (unpublished results). To determine if a similarmechanism exists in other parts of the digestive system, weanalyzed both the pancreatic/duodenal secretions for the

presence of dopamine and the pancreas itself for dopaminergicmarkers.

METHODS

Sample Collection. To collect pancreatic/duodenal juicefrom rats, we ligated the duodenum near the pylorus toeliminate the contribution of dopamine from the stomach andligated the duodenum below the papilla Vateri (the opening ofthe pancreatic duct into the duodenum) and the ductuscholedochus (where bile enters the pancreatic juice) in anes-thetized rats. After 4 hr, the juice from the pancreas was takenand frozen until further processing. Collected from anesthe-tized rats, pancreatic tissue was frozen on dry ice and kept at-80°C until further processing.Catecholamine Measurements. Concentrations of dopa-

mine, dihydroxyphenylalanine (dopa), and 3,4-dihydroxyphe-nylacetic acid (DOPAC) were determined by liquid chroma-tography with electrochemical detection after alumina extrac-tion (6). Intraassay coefficients of variation were 8.1% fordopamine and 3.9% for DOPAC.Chemical Sympathectomies (CSs). CSs of adult male

Sprague-Dawley rats were performed using 6-hydroxydopa-mine (6-OHDA), according to the following schedule. Toavoid the initial side effects due to quick depletion of periph-eral norepinephrine terminals, rats were first given 6-OHDAat a low dose (5 mg/kg) intraperitoneally. After 12 hr, each ratreceived 15 mg/kg, and then two additional doses of 30 mg/kgat 24 and 36 hr. Control rats received injections of vehicle onthe same schedule. Four days after the last injection, the ratswere anesthetized, the pylorus was ligated, and samples werecollected as above. Norepinephrine content of the heart wasused to control the completeness of the CS. All treated ratsshowed >95% reduction of cardiac noradrenaline levels.

Immunohistochemistry. For immunohistochemistry, ratswere anesthetized with pentobarbital sodium (40 mg/kg bodyweight), then perfused with 4% paraformaldehyde. The pan-crei were removed, cryoprotected in 20% sucrose, frozen ondry ice, and cut in a cryostat onto silanized slides in 12-,um-thick sections. To decrease nonspecific staining, the sectionswere pretreated for 30 min at room temperature (RT) in asolution containing 0.6% Triton X-100, 5% normal serum inlx PBS (pH 7.4). Normal serum was either goat or donkey(depending on the host of the secondary antibody). Formonoclonal primary antibodies, the normal serum was re-placed with 0.1% bovine serum albumin. Primary antibodies(Table 1) were applied to the sections either for 1 hr at RT or

Abbreviations: CS, chemical sympathectomy; 6-OHDA, 6-hydroxydo-pamine; DAT, dopamine transporter; TH, tyrosine hydroxylase;VMAT, vesicular monoamine transporter; dopa, dihydroxyphenylala-nine.tTo whom reprint requests should be addressed.§On leave from: First Department of Medicine, Medical University ofPecs, Pecs, Hungary.

10377

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

10378 Medical Sciences: Mezey et al.

Table 1. Primary antibodies used in these studies

Antibody Host Dilution Source/Ref.TH Rabbit 1:1000 Eugentec (37)TH Rabbit 1:1000 8, 9TH Mouse mc 1:500 Incstar (10)TH Mouse mc 1:500 Boehringer Mannheim (11)Dopamine Rabbit 1:2000 12Dopamine Goat 1:5000 12DAT Rabbit 1:1000 13Neurofilament Mouse mc 1:4000 Chemicon (14)Insulin Guinea pig 1:2000 Chemicon

TH, tyrosine hydroxylase; DAT, dopamine transporter; mc, mono-clonal.

for 12 hr at 4°C. After several rinses in PBS, fluorescentsecondary antibody was applied for 1 hr at RT in the dark, andthe sections were rinsed, coverslipped, and viewed with a LeitzDialux 20 fluorescent microscope. For double-staining, theabove procedure was used, and then the sections were incu-bated in the second primary antibody and processed as de-scribed above. The second secondary antibody was conjugatedto a different fluorochrome than the first one. For visualizingthe TH immunostaining, a tyramide signal amplification sys-tem was used (DuPont/NEN) (15). For dopamine immuno-staining, strong reducing agents (12) were used in all solutionsthrough the primary antibody incubation to prevent dopamineoxidation. Negative controls included staining with nonim-mune rabbit serum, leaving out the primary antibody or thesecondary antibody and using several antibodies when possibleto recognize the same antigen (Table 1). In the doubleimmunostaining procedures, extra care was taken to avoid anypossible cross-reactivity between the different primary andsecondary antibodies so that the second secondary antibodydid not recognize the first primary antibody. In addition, thedouble stainings were always repeated reversing the order ofthe primary antibodies.

In Situ Hybridization Histochemistry. For all in situ hybrid-ization studies, 200 g male rats were decapitated, the pancreatawere removed, rinsed with PBS, and quickly frozen on dry ice.The tissue was kept at -80°C until sectioning. Sections (12 umthick) were cut in a cryostat (Reichert Frigocut 2500),mounted onto silanized slides, dried on a hotplate at 37°C, andthen processed for in situ hybridization histochemistry asdescribed earlier (16). The templates that were used formaking the RNA probe are listed below. The sense probe usedas a control did not hybridize to the tissue.

Riboprobes were prepared using [35S]UTP and the Maxi-script (Ambion, Austin, TX) kit. The following templates wereused. For TH, a fragment of the tyrosine hydroxylase (17)cDNA (a gift of Dona Chikarachi, Duke University MedicalCenter, Durham, NC) corresponding to nucleotides 16-1165of the coding sequence was used. For dopamine Dlb, afragment of the cDNA between nucleotide 1-1565 (GenBankaccession no. M69118) was subcloned into pBluescript KS II+.For transporter mRNAs, specific riboprobes were transcribedfrom DNA templates generated by PCR using the followingprimers: for DAT (GenBank accession no. M80570; ref. 18)(nucleotides 1088-1109) TAGAGACGCAATCATCACCA-CC (primer 1, sense) and (nucleotides 1575-1555) CACTG-AATTGCTGGACGCCGT (primer 2, anti-sense); for vesic-ular monoamine transporter 1 (VMAT1) (GenBank accessionno. M97380; ref. 19) (nucleotides 1741-1761) ACAGAGAC-CCAGATGTACACA (primer 3, sense) and (nucleotides2177-2157) GTTAGTCTCTTCTTTCCGTCC (primer 4, an-tisense); for VMAT2 (GenBank accession no. L00603; ref. 20)(nucleotides 1649-1669) GACCCTCTAACGTCGCCAAA-TG (primer 5, sense) and (nucleotides 2160-2140) ACACA-TTGGTAC TAGTTACAA (primer 6, anti-sense).

For nonradioactive in situ hybridization histochemistry (21),the above riboprobes were labeled with digoxigenin-UTP.After hybridization, the digoxigenin was developed using ananti-digoxigenin antibody conjugated to horseradish peroxi-dase. The signal then was amplified using a tyramine ampli-fication method (15) and the TSA kit (New England Nuclear).The final marker was fluorescein isothiocyanate-conjugatedtyramide.Western Blots. Tissues were homogenized in 0.2% Triton

X-100 in PBS (pH 7.5), assayed for protein content, and storedat -80°C. Either 20 ,g protein of pancreatic tissue or 2 ,ugprotein from adrenal gland was size-fractionated by SDS/PAGE and then electroblotted onto Hybond-EnhancedChemiluminesence (ECL) nitrocellulose membrane (Amer-sham) at 30 V for 18 hr in transfer buffer (20% methanol/25mM Tris/192 mM glycine). Blots were preblocked in 5%nonfat dry milk in Tris-buffered PBS and incubated withanti-TH antibodies for 1 hr at 22°C. Blots were then incubatedwith horseradish peroxidase-conjugated goat anti-mouse IgG(1:1000) for 1 hr at 22°C and developed using the ECL kit(Amersham). Prestained molecular weight markers were usedto estimate the molecular weight of immunoreactive species.Antibody dilutions used (also see Table 1) were: BoehringerMannheim, 1:100 and 1:1000; Incstar, 1:2000; Eugentec,1:5000; Thibault et al. (22) 1:1000TH Activity Measurements. For determining TH activity,

tissues were homogenized in PBS and centrifuged at 20,000 xg for 10 min; the resulting supernatant was recovered. THactivity in the soluble fraction was determined by a modifica-tion of the method of Naoi et al. (23) in which NADPH (0.5mM) and hydropteridine reductase (1 unit per reaction) wereused as the reducing system (both from Sigma).

Southern Blots/PCR Total RNA (5 ,ug) from rat pancreasor adrenal was treated with DNase I (amplification grade,GIBCO/BRL) for 15 min at 37°C. First-strand cDNA wassynthesized using random primers and 200 units of Moloneymurine leukemia virus reverse transcriptase (Superscript II,GIBCO/BRL) according to manufacturer's instructions. Sub-sequently, PCR amplification was performed using 10% of thecDNA in the presence of 0.2 ,uM of each dNTP, 8 mM MgSO4,1 ,uM primers, and 2.5 units of Pwo polymerase (BoehringerMannheim). Thermocycling conditions were: 1 min at 95°C, 2min at 55°C, and 2 min at 72°C for 35 cycles followed by 7 minat 72°C. Amplified products were size-fractionated on a 1.4%agarose gel and blotted to nitrocellulose membranes (Schlei-cher & Schuell). Oligonucleotides were labeled using[y-32P]ATP and polynucleotide kinase (New England Biolabs),and were hybridized to Southern blots in 3x SSC/2x Den-hardt's solution at 55°C for 18 hr. Final washes were performedin 3x SSC at 55°C.For amplification of VMAT cDNAs, the following primers

were used: For VMAT1, (nucleotides 399-419) CACCTTC-CTGTACGCGACAGA (primer 7, sense) and (nucleotides655-635) CTTCTTCTAAGAACTCTATCC (primer 8, anti-sense); for VMAT2, (nucleotides 2161-2181) TTAGGAATT-TACAACTCGTCA (primer 9, sense) and (nucleotides 2661-2641) GTGAAACTCATTTCTACATTG (primer 10, anti-sense). Oligonucleotides for probing PCR products: ForVMAT1, (nucleotides 603-556) GGCTTCAGTGACTG-GAGGAGGGATGGTGCCATTTGTCCAAGTTACACG(oligo 11); for VMAT2, (nucleotides 2300-2347) CATGTT-TCACCCCTTGTCGGCTTTAGTGACTGCT-GCCGATGACATAAC (oligo 12).

RESULTSFour hours after ligation of the pylorus, the duodenum, andthe bile duct, duodenal juice contained a significant amount ofdopamine [276 ± 61.8 pg/ml (n = 15)]. Then we used CS toeliminate sympathetic nerve fibers and, thus, rule out neurons

Proc. Natl. Acad. Sci. USA 93 (1996)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

Proc. Natl. Acad. Sci. USA 93 (1996) 10379

as a source of dopamine. Dopamine content did not changeafter CS [463 + 287 pg/ml (n = 5)], suggesting local, non-neuronal production of dopamine.

In the pancreatic tissue itself, 40% of the dopamine content[32 + 6 (n = 5) versus 84 + 36 (n = 5) pmol/mg tissue] wasstill present after CS, in agreement with previously publisheddata (24-26) (Fig. 1). We also measured consistently highlevels of DOPA in all samples (data not shown).The rate-limiting step in dopamine synthesis is the conver-

sion of tyrosine to dopa by TH (27). Therefore, the presenceof TH enzyme activity implies the potential for dopamineproduction. We were able to detect TH activity in the pan-creas; the activity was little changed compared with tissuenorepinephrine by CS (2.8 + 0.7 versus 1.8 + 0.8 pmol/mgprotein per min) (Fig. 1).

Using four different antibodies to TH and a signal ampli-fication method (15), we localized TH immunostaining to allexocrine cells of the pancreas (Fig. 24). The insulin-positivepancreatic islet cells are negative for TH (Fig. 2B). In additionto cells in the exocrine pancreas, numerous nerve fibers werealso TH positive, but no neuronal cell bodies were observed.All four anti-TH antibodies gave similar results. To corrobo-rate the histological results, we determined that anti-THantibodies recognize a protein in pancreatic extracts corre-sponding in apparent molecular weight to adrenal TH (Fig. 3).

0,0-

z

0,E4-

E

e

=m

E

c

2500 -

1000 -

1000

500 -

0

100 -_

80 -_

60 -_

40 -_

20 -_

0 _

4-

3.5 -

3 -

CONTROL6OHDA

O CONTROL* 6OHDA

o3 CONTROL* 6OHDA

2 5

Ce 2

E 1.5- ;

..........

0.5

FIG. 1. (Top and Middle) Norepinephrine (NE) and dopamine(DA) content, respectively, of pancreatic tissue. (Bottom) TH activityin the pancreas before and after CSs. Note that there is a significantamount of dopamine still present after CS. More than 70% of THactivity also persists after CS. Number of rats in each group: n = 5.

FIG. 2. Immunohistochemical ocalization of TH, dopamine, andDAT in pancreas. Fluorescent in situ hybridization ofVMAT mRNAs.All exocrine pancreatic cells are immunopositive for TH (A), while theinsulin-producing islet cells (asterisks) in the same section (B) arenegative. (A) An arrowhead points at a nerve fiber. (C) Similarly, allexocrine cells are immunopositive for dopamine itself in addition tosympathetic nerve fibers that are also immunopositive for dopamine(arrowheads). (D) DAT is present in the plasma membrane of allpancreatic cells in addition to the epithelium of excreting ducts (largearrow), veins (small arrows), and some connective tissue elements. (D)Arteries are negative (arrowhead). In situ hybridization histochemistrydemonstrates the VMAT1 (E) and VMAT2 (F) mRNAs in all cells(the perinuclear staining is typical of the localization of an mRNA).(Bar = 100 ,um.)

Immunostaining for dopamine itself revealed the presenceof dopamine in most exocrine cells (Fig. 2C) in addition tosympathetic nerves. We also visualized a few dopamine-positive enterochromaffin cells in the duodenal epithelium.The very scattered distribution and low number of cells makeit doubtful that these cells produce sufficient amounts ofdopamine to account for the dopamine in the pancreatic/duodenal juice.Dopamine neurons express a plasma membrane DAT that

transports released dopamine back into the cell but that mayalso function in reverse to release dopamine (for review, seeref. 28). Using an antibody to DAT (13), we found that mostexocrine and endocrine pancreatic cells and many vascularelements are DAT immunopositive (Fig. 2D). We consistentlyobserved a plasma membrane immunostaining in the endo-

Medical Sciences: Mezey et aL

^, ^^2

2

1

1I

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

10380 Medical Sciences: Mezey et al.

LA

mC L.

L. CLL

86.0

A VMAT1

Vl

_CO LI

C

bp a 21

310 _

194 -

~~~ w_

50.7-.

CLU e

C +In <

+ +

0

e

FIG. 3. TH immunoreactivity in pancreas and adrenal. Anti-THantibody (Incstar, 1:2000) recognized a protein species of the sameapparent molecular mass in the pancreas (20 ,ug protein) as in adrenalgland (5 ,ug protein), the positive control. These results were replicatedwith each of the other three TH antibodies (see refs. 7-9 and 11).Molecular mass markers are indicated in kDa.

crine and exocrine cells, in the excreting ducts of the pancreas,and in the venous part of the vasculature. No DAT immuno-staining was present in the arterial endothelium (Fig. 2D).Using in situ hybridization histochemistry, we were also able todetect the mRNA encoding DAT in both exocrine and endo-crine cells (data not shown).

In neurons and neuroendocrine cells, VMATs packagemonoamines (serotonin, dopamine, norepinephrine, epineph-rine, and histamine) into storage vesicles. Two VMAT sub-types have been identified: (i) VMAT1 (19) from adrenalchromaffin cells and (ii) VMAT2 (19, 20) from monoamineneurons, platelets, and mast cells. Our in situ hybridizationhistochemistry revealed that both VMAT1 and VMAT2mRNAs are present in the pancreas (Fig. 2 E and F, respec-tively). To prove the specificity of hybridization, two differentriboprobes were used individually with identical results foreach transporter. We confirmed the presence of both VMAT1and VMAT2 mRNA in pancreas by amplification of specificDNA fragments from pancreatic cDNA (Fig. 4). VMAT1 andVMAT2 may be present in different populations of storagevesicles (19, 28).

Finally, we looked for a potential target of the dopamineproduced by the pancreas using in situ hybridization histo-chemistry to visualize mRNAs encoding the five known dop-amine receptors. In the pancreas, the dopamine Dlb (D5)receptor (29) mRNA is abundantly and widely expressed,whereas the other dopamine receptor mRNAs were not de-tectable. Dlb receptor mRNA was also present in all of theduodenal epithelial cells as was the case for the stomachepithelial cells (unpublished results). Cells in the laminapropria as well as many smooth muscle cells also make asignificant amount ofDlb receptor mRNA (Fig. 5), in additionto some D3 receptor and D4 receptor mRNAs (data notshown).

DISCUSSIONThe pancreas contains markers usually associated with cat-echolaminergic and neuroendocrine cells. We have shown herein exocrine pancreatic cells the presence of dopamine itself byimmunocytochemistry as well as the presence of tyrosinehydroxylase, the enzyme that is responsible for the synthesis ofdopamine. Another characteristic of catecholaminergic cells isthe presence of transporter molecules. We found that bothtypes of the vesicular monoamine transporters (VMAT1 andVMAT2) are present in the pancreas. The presence of bothVMATs in many cells suggests that dopamine, or othermonoamines, may be differentially targeted to and costoredwith specific hormones or enzymes in the pancreas. We have

B VMAT2

bp

600 -

310 _

C

0C +

UL dh < <0 0-i 0 IV A

0 u~~~6. c v wXL 2. X-,

C

2s0

C

FIG. 4. Detection of VMAT mRNAs in pancreas by reversetranscription-PCR. VMAT1 (A) and VMAT2 (B) mRNAs are presentin the pancreas as determined by reverse transcription-PCR. For eachset of primers, a DNA fragment of identical size was amplified frompancreas, adrenal (mRNA control), and a positive VMAT cDNAplasmid, and was identified by hybridization to a third oligonucleotide.Ethidium bromide staining is always shown above the Southern blotanalysis. (A) A 256-bp fragment of VMAT1 cDNA was amplified withprimers 7 and 8 and hybridized to oligonucleotide 11. (B) A 500-bpfragment of VMAT2 cDNA was amplified with primers 9 and 10 andhybridized to oligonucleotide 12. Lanes: pancreas-no RT, equivalentamount (relative to cDNA) ofRNA only, no reverse transcriptase; +vecDNA + primers, positive cDNA with primers; +ve cDNA - primers,positive cDNA without primers; primers only, no cDNA added. DNAsize markers are indicated in base pairs (bp).

also demonstrated the mRNA encoding the plasma membraneDAT (data not shown) as well as the presence of DAT itself,using immunostaining.We found both dopamine content and TH activity in the

normal and in the CS pancreas and a significant amount ofdopamine present in the duodenal juice even after CS. All thesympathetic fibers are destroyed by CS (as indicated by thedramatic decrease in norepinephrine content) (Fig. 1) and, inagreement with literature data (30), we did not detect TH-positive intramural ganglionic neurons in the pancreas. Thus,there is no neuronal source for dopamine production observedafter CS. These observations raised the possibility that thepancreas itself may be the major source of dopamine in theduodenal juice.

Proc. Natl. Acad. Sci. USA 93 (1996)

'Wolu".1ilbuiLl,

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

Proc. Natl. Acad. Sci. USA 93 (1996) 10381

"4r ~44

A*J o ;h P-; ~ ~~S . w.

' ,., ; , J ' d'. 't.< r :, .

I>

f

ih mesl _

-.d'1 I

o

FIG. 5. Dopamine Dlb receptor mRNA in duodenal and pancreatic cells. Brightfield (A and C) and are the corresponding darkfield (B andD) images of the same sections. (A and B) In a cross section through the pancreas (P) and the duodenal wall (D), most pancreatic cells as wellas smooth muscle cells in the duodenum are positive for the Dlb receptor mRNA. A few enteric ganglionic cells (arrowhead) are also positive.(C and D) A transverse section of the duodenum shows a very strong labeling of all epithelial as well as many lamina propria cells (*) with thedopamine Dlb receptor mRNA probe. Arrows indicate epithelium of two villi. (Bar = 100 ,um.)

The immunostaining showed a lack of DAT in the endo-thelial cells of arteries, while it is rather abundant in excretingducts and veins. This distribution might be consistent with localproduction and usage of dopamine. Since the pancreas is alarge organ, the dopamine produced here could get into thegeneral circulation and result in high dopamine levels in theblood. This might have harmful cardiovascular effects. Onecan imagine that the presence of DAT in the veins (and not inthe arteries) serves to reuptake the dopamine from the pan-creas and metabolize it before it could leak into the generalcirculation. This suggests that the dopamine made here isintended exclusively for local use.

In looking for a target site of this locally produced dopa-mine, we demonstrated the presence of mRNA encoding thedopamine Dlb receptor in all epithelial cells of the duodenumand the pancreas itself. While the dopamine Dlb receptorsfound in abundance in gastrointestinal epithelial cells seem tobe the target of intraluminal dopamine, the pancreatic dopa-mine receptors might mediate an autocrine feedback effect.The effects of dopamine on exocrine pancreatic secretion havebeen studied in several animal species (31, 32). While a strongstimulatory effect on secretion was found in dogs (33-35), theeffect was less pronounced in other species and is unclear inhumans (36).

Our results define a novel catecholaminergic system in thepancreas and duodenum and suggest a paracrine hormonalrole for dopamine outside of the central nervous system.Along with acid-secreting parietal cells in the stomach,pancreatic exocrine cells that produce the digestive enzymessynthesize and release dopamine into the duodenum. Thepancreatic cells also produce large amounts of dopa, whichmay serve only as the precursor for dopamine. The possibilitythat dopa may have a role of its own should be considered,however.Dopamine appears to have a beneficial effect in acute

pancreatitis (37-39). Since we found a similar nonneuronalcatecholamine system in the gastric epithelium, it appears thatthe production and release of dopamine in concert withpotentially harmful agents (digestive enzymes) may be ageneral mechanism of self-defense in the gastrointestinalsystem. Our results suggest the potential efficacy of specificdopamine receptor agonists in the treatment of gastrointesti-nal diseases. Future studies should focus on the mechanismwhereby dopamine protects the pancreas and duodenum andchanges in dopamine secretion in pathological states.

We acknowledge Drs. Michael Brownstein and Miklos Palkovits forcritical reading of the manuscript; Drs. H. Steinbusch and R. A.

Medical Sciences: Mezey et aL

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

10382 Medical Sciences: Mezey et al.

Vaughan for supplying antibodies for dopamine and DAT; Dr. Sey-mour Kaufman for his valuable advice on the TH assays; and RicardoDreyfuss for expert photography.

1. Sikiric, P., Rotkvic, I., Mise, S., Petek, M., Rucman, R., Seiwerth,S., Zjacic-Rotkvic, V., Duvnjak, M., Jagic, V., Suchanek, E.,Grabarevic, Z., Anic, T., Brkic, T., Djermanovic, Z., Dodig, N.,Marovic, A. & Hernandez, D. E. (1991) Dig. Dis. Sci. 36,905-910.

2. Glavin, G. & Szab6, S. (1990) Dig. Dis. Sci. 35, 1153-1161.3. Glavin, G. & Dugani, A. (1987) Life Sci. 41, 1397-1408.4. MacNaughton, W. K. & Wallace, J. L. (1989) Gastroenterology

96, 972-980.5. Eisenhofer, G., Aneman, A., Hooper, D., Holmes, C., Goldstein,

D. & Friberg, P. (1995) Am. J. Physiol. 268, G641-G649.6. Eisenhofer, G., Goldstein, D., Stull, R., Keiser, H., Sunderland,

T., Murphy, D. & Kopin, I. (1986) Clin. Chem. 32, 2030-2033.7. Ziller, C., Mirabel, M. A., Vandenbunder, B. & Fauquet, M.

(1994) Anat. Embryol. 189, 307-315.8. Arluison, M., Dietl, M. & Thibault, J. (1984) Brain Res. Bull. 13,

269-285.9. Dietl, M., Arluison, M., Mouchet, P., Feuerstein, C., Manier, M.

& Thibault, J. (1985) Histochemistry 82, 385-389.10. Chiodo, L. A. & Kapatos, G. (1992) Synapse 11, 294-309.11. Rohrer, H., Acheson, A. L., Thibault, J. & Thoenen, H. (1986)

J. Neurosci. 6, 2616-2624.12. Steinbusch, H. W., Wouterlood, F. G., de Vente, J., Bol, J. G. &

Berkenbosch, F. (1988) Acta Histochem. Suppl. 35, 86-106.13. Vaughan, R. A. (1995) Mol. Pharmacol. 47, 956-964.14. Harris, J., Ayyub, C. & Shaw, G. (1991) J. Neurosci. Res. 30, 47-62.15. Berghorn, K. A., Bonnett, J. H. & Hoffman, G. E. (1994) J. His-

tochem. Cytochem. 42, 1635-1642.16. Bradley, D. J., Towle, H. C. & Young, W. S. (1992) J. Neurosci.

12, 2288-2302.17. Grima, B., Lamouroux, A., Blanot, F., Biguet, N. F. & Mallet, J.

(1985) Proc. Natl. Acad. Sci. USA 82, 617-621.18. Kilty, J. E., Lorang, D. & Amara, S. G. (1991) Science 254,578-579.19. Liu, Y., Peter, D., Roghani, A., Schuldiner, S., Prive, G. G., Eisen-

berg, D., Brecha, N. & Edwards, R. H. (1992) Cell 70, 539-551.20. Erickson, J. D., Eiden, L. E. & Hoffman, B. J. (1992) Proc. Natl.

Acad. Sci. USA 89, 10993-10997.21. Young, W. S. I. (1989) Neuropeptides 13, 271-275.

22. Thibault, J., Vidal, D. & Gros, F. (1981) Biochem. Biophys. Res.Commun. 99, 960-968.

23. Naoi, M., Takahashi, T. & Nagatsu, T. (1988)J. Chromatogr. 427,229-238.

24. Graffner, H., Ekelund, M., Hakanson, R. & Rosengren, E. (1985)Scand. J. Gastroenterol. 20, 1276-1280.

25. Orloff, L. A., Orloff, M. S., Bunnett, N. W. & Walsh, J. H. (1985)Life Sci. 36, 1625-1631.

26. Quansah, F. A., Klatt, C. & Feldman, J. M. (1981) Metabolism 30,242-247.

27. Nagatsu, T., Levitt, M. & Underfriend, S. (1964) J. Biol. Chem.239, 2910-2917.

28. Borowsky, B. & Hoffman, B. J. (1995) in International Review ofNeurobiology, eds. Bradley, R. J. & Harris, R. A. (Academic, SanDiego), pp. 139-199.

29. Tiberi, M., Jarvie, K. R., Silvia, C., Falardeau, P., Gingrich,J. A., Godinot, N., Bertrand, L., Yang-Feng, T. L., Fremeau,R. T. J. & Caron, M. G. (1991) Proc. Natl. Acad. Sci. USA 88,7491-7495.

30. Oomori, Y., luchi, H., Ishikawa, K., Satoh, Y. & Ono, K. (1994)Histochemistry 101, 313-323.

31. Furuta, Y., Hashimoto, K. & Washizaki, M. (1978) Br. J. Phar-macol. 62, 25-29.

32. Hashimoto, K., Oguro, K. & Furuta, Y. (1977) Arch. Histol. Jpn.40, Suppl., 129-132.

33. Iijima, F., Iwatsuki, K. & Chiba, S. (1983) Eur. J. Pharmacol. 92,191-197.

34. Takeuchi, S., Okamura, T., Shonai, K. & Imamura, S. (1990)Nippon Ika Daigaku Zasshi 57, 119-126.

35. Hashimoto, K., Sato, S. & Takeuchi, 0. (1971) Br. J. Pharmacol.43, 739-746.

36. Valenzuela, J. E., Defilippi, C., Diaz, G., Navia, E. & Merino, Y.(1979) Gastroenterology 76, 323-326.

37. Shen, J., Huang, M. K. & Wu, F. L. (1990) Chin. Med. J. (Engl.Ed.) 103, 201-207.

38. Sikiric, P., Rotkvic, I., Mise, S., Krizanac, S., Gjuris, V., Jagic, V.,Suchanek, E., Petek, M., Udovicic, I., Geber, J., Tucan-Foretic,M., Cvitanovic, B., Ivanovic, D. & Marovic, A. (1988) Eur.J. Pharmacol. 147, 321-326.

39. Nishiwaki, H., Koh, I., Kanazawa, G., Hiura, A., Satake, K. &Umeyama, K. (1988) Nippon Geka Gakkai Zasshi 89, 1990-1996.

Proc. Natl. Acad. Sci. USA 93 (1996)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0