SCIENTIFIC PAPERS

In Vitro Release of Gastrin from Isolated Perfused Antrum

Robin Dretler, BA, Boston, Massachusetts

Robert I. C. Wesdorp, MD, Boston, Massachusetts

Josef E. Fischer, MD, Boston, Massachusetts

Peter B. Soeters, MD, Boston, Massachusetts

Amin M. Ebeid, MD, Boston, Massachusetts

Although the primary stimulus for the release of gastrin remains physiologic, such as ingestion of food,

“antral distension,” or vagal activity [l-3], a variety

of other substances have been shown to release gas- trin in vivo as well, including calcium and glycine

[4-61. In vivo studies involving gastrin release are

handicapped by a series of technical problems, not the least of which is the feedback mechanism by

which acid produced by the gastrin-stimulated pa- rietal cells inhibits further antral release of gastrin [7,8]. Antral exclusion and a variety of other com- plicated physiologic technics have been derived in

vivo to isolate the antrum from inhibiting mecha-

nisms [3,9,10]. To further investigate gastrin physi-

ology, it would be desirable to devise a physiologic

preparation in which reliable and reproducible re- lease of gastrin from the gastric antrum or any other part of the small bowel could be obtained. An isolated

perfusion technic used for catecholamine studies [Ii] has been adapted for this purpose. We believe the following experiments suggest that an isolated per- fused antrum preparation may be very useful in the elucidation of some of the physiologic characteristics of gastrin. It is well documented that in vivo instil-

lation of acetylcholine into the isolated antral pouch causes an immediate release of gastrin [3,12]. In these

perfusion studies, acetylcholine, glycine, and calcium apparently also cause an immediate release of gastrin in vitro.

From the Department of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts.

Reprint requests should be addressed to Josef E. Fischer, MD, Department of Surgery, Massachusetts General Hospital. Boston. Massachusetts 02114.

Methods

Isolated Antral Perfusion

Fasting female Sprague-Dawley rats, weighing 200 to 250 gm, were killed by cervical fracture and their gastric antra quickly removed, washed in Krebs-Ringer’s A solution, everted, and placed in the perfusion chamber.

The perfusion chamber was a glass chamber (30 X 15

mm) with a capacity of 2 ml, much of which was occupied by the perfused tissue. Perfusion was carried out from several reservoir bottles (Figure l), each of which contained an oxygenated Krebs-Ringer’s A solution [13] either as a control perfusion solution or with the releasing agent to be tested. A large series of solutions may thus be used and are connected to the perfusion chamber by a series of switching valves. (Figure 1.) From the reservoir bottles, the solutions were first allowed to flow through glass-jacketed heat ex- change systems, the outer jackets of which were perfused at a constant temperature of 40°C before perfusing the antral tissue. Regulation of flow rate was carried out by a 4 or 2 channel perfusion pump (Polystaltic Pump, Buchler Instruments@) generally at a rate of 30 ml/hr (0.5 mUmin). Effluent samples were collected every 2 minutes (1 ml). After an initial perfusion (12 minutes) with oxygenated Krebs-Ringer’s A solution to establish a basal level of gastrin efflux, the valves were turned to allow the testing substance in the same Krebs-Ringer’s A solution to perfuse the isolated tissue. After a stimulatory interval of 22 min- utes, the chamber was again perfused with the basal solu- tion to reestablish the baseline level of secretion.

Perfusions were carried out with 50 mM acetylcholine (+O.l mM eserine), 50 mM glycine, or 10 mM or 30 mM calcium chloride added to the Krebs-Ringer’s A solution. Eserine, an anticholinesterase, was added to prevent de- struction of acetylecholine by cholinesterase. Control perfusions were carried out with 0.1 mM eserine and with Krebs-Ringer’s A solution without an added substance.

Volume 134, August 1977 237

Dretler et al

Each perfusion was repeated on at least six different antra. To test the ability of the perfused tissue to release gastrin repeatedly, we restimulated the tissue after a 10 minute rest in four instances. Effluent samples were stored at -20% before being assayed. Recovery of known quan- tities of synthetic human-line gastrin SHG: 1-17 (Imperial Chemical Industries, LM) was approximately 91 f 3.7 per cent.

Gastrin Radloimmunoaaaay

A radioimmunoassay technic similar to that employed by McGuigan (141 and previously described [Is], was used to measure gastrin in the perfusate. Antigastrin antibody was produced by rabbits by serial injection of antigen prepared from SHG: 2-17 coupled to bovine serum albu- min. SHG: 1-17 was labeled with iodine 125 (1251) by the technic reported by Hunter and Greenwood [ 161. Fifty ~1 of the effluent sample to be assayed were incubated for four days at 4OC with 250 ~10.02 M barbital buffer (pH 8.4), 50 gl of charcoal-treated human plasma, 2,000 counts/min ‘251~labeled SHG: 1-17 in 100 ~1, and 100 ~11:30,000 rabbit antigastrin antibody, making a final dilution of 1:150,000. Free lz51-labeled SHG: 1-17 was separated from that bound to antibody and plasma proteins by mixing with 200 ~1 of dextran-coated charcoal for 15 minutes. After cen- trifuging at 3,000 rpm for 20 minutes, the supernatant was

Krebs Ringe

ebs Ringer

i Szbsfance

Figure 1. Experfmental apparatus. The Krebs-Rfnger’s A solution, wtth or without test substance, flows from the reservoir bottles through a heat exchange system. Regu- latfon of the various flows into the perhrslon chamber is car&d out by a perktalttc pump. After the ~lfactkm per&d using control sotuttoq the valves are changed ariii the teai solution Infused. At the end of the test period, valves are again changsd, aIrOwIng a series of penirsbns to be canted out wltftout interrupthrg the flow of perfusing solutfon.

decanted and the two fractions counted separately in a gamma scintillation spectrometer. Bound/free ratio was corrected by a computer program, and the gastrin con- centration was ascertained by reading off an accompanying standard curve.

Samples were assayed in triplicate on two different oc- casions. The coefficient of variation for the assay using the technic described is 5 f 1.3 per cent, and gastrin can be measured accurately between levels of 20 and 1,000 pg/ml. Concentrations of gastrin during the perfusion fell ap- proximately in the middle of this range, that is at the steepest and most sensitive part of the standard curve [ 2 71, generally between 80 and 400 pg/ml, although there existed some variation between basal levels of different perfusions. Results were analyzed using the Student t test in which each perfusion period was compared with basal levels.

Results

Upon perfusion of unminced antrum with 50 mM acetylcholine (+O.l mM eserine), the gastrin level in the effluent increased within 2 minutes to 4 times the

basal level (Figure 2) and decreased rather sharply. A smaller second peak appeared after 20 minutes (10 ml). When the perfusate was switched to Krebs- Ringer’s A solution, the gastrin concentrations re- turned to the basal level. The large standard error can be explained by the variation between basal levels of different perfusions, with basal levels varying from 70 to 300 pg/ml.

Glycine, a known local stimulator of gastrin release [5], elicited a smaller reaction from the antral tissue. During perfusion with 50 mM glycine, gastrin levels increased to 2 times the basal level in 4 minutes

400r

300-

e

F - zoo-

s

E

2 100-

O-

50 mM ACH t 0.1 mM ESERINE

in KRE& RINGER A

I P(OO1 II

2 4 6 8 10 (2 (4 16 1s 20 22

ml P&w USED

Figure 2. Mean effluent gastrfn levels f SEM after nine separate perfustons of unmmcqd rat antra wfth 50 mM acetytcf?ogne +O. 1 mM eserine in Krebs-Ringer’s A solu- tion.

238 The American Journal of Surgery

Antral Gastrin Release

(Figure 3), after which they gradually decreased. No second peak was seen. However, a consistent feature during perfusion with glycine was increased spon- taneous leakage of gastrin, even after the perfusate had been changed back to Krebs-Ringer’s A solution. This may, however, represent a delayed second peak, although this occurred after changing back to control solution.

Upon perfusion with 10 mM or 30 mM calcium chloride in Krebs-Ringer’s A solution, gastrin levels in the effluent increased within 4 minutes to 2 times the basal level, similar to perfusion with 50 mM gly- tine. (Figure 4.) There was no apparent difference between the effect of high and low concentration of calcium chloride. No secondary release of gastrin was seen.

In six control experiments, perfusion with 0.1 mM eserine alone as well as Krebs-Ringer’s A solution without an added substance failed to increase gastrin levels in the effluent. (Figure 5.) In four other ex- periments, the ability of the tissue to release gastrin at various intervals was tested by restimulating the tissue after a 10 minute rest after the initial release by acetylcholine. As shown in Figure 6, large amounts of gastrin are again released, confirming a stable re- producible model.

In perfusion with acetylcholine, a constant feature was the apparent biphasic release of gastrin. After the initial release of gastrin from the isolated perfused

50mM GLYCINEin

KREES RINGER A

soor

700-

600 -

5 500 -

B

s 4oo a h 300-

? Q 200-

100-

o- 2 4 6 8 IO 12 14 I6 I6 20

m/ PERFUSED

Fb@lre3.Mean-ga&blh?veLFfSEM%llerpetWkn of six individual unminced rat antra with 50 mM glycine In Krebs-Ringer’s A solution. Note that after the termlnatkn of glycine perfusion and when the antra are perfused with Krebs-Ringer’s A solutkn, a high spontansous effluent persists, as ii the tksues were damaged by giycine.

antrum, there was inevitably a secondary peak, smaller than the first, suggesting secondary release of gastrin. To better demonstrate this phenomenon, the isolated antrum was minced, allowing better contact between the antral cells containing gastrin and the perfusion fluid. However, as can be seen in Figure 7, the primary release of gastrin was the same as upon perfusion of unminced antrum, although the secondary release was greater and lasted for a longer period of time. Secondary release of gastrin was again

10 mM CALCIUM CHLORIDE in KREB’S RINGER A

0 2 4 6 6 10 12 14 i6 18 20 22

ml PERFUSED

Figure 4. Mean effluent gastrh levels f SEM after six pertuskns of unmkced gastrk antra with 10 mM caklum chlortde In Krebs-Ringer’s A solution. Only one peak was seen. Similar results occurred affer pertitsion with 30 mM calcium chloride.

0. I mM ESERINE in

KREEk RINGER A

I

. 604 I

200-

IOO-

O- 2 4 6 6 IO 12 I4 I6 16 20

m/ PERFUSED

Figure 5. Mean effluent gastrln levels f SEM atler six conbvlperirusconsof- rat anbum fn Kn?bs*k&Ws A solutkn with 0.1 mM eserke. No increase in effluent gastrin level occurred.

vohllne 134, Au9usl1977 239

Dretler et al

seen only after perfusion with acetylcholine in the minced preparation.

Gastric antra were assayed after perfusion and compared with twelve gastric antra which had been perfused without stimulation for 10 minutes. There was great individual variation between antra. No consistent difference between “pre- and postinfu- sion” antra was detected.

Comments

The results of the present study confirm the results of previously reported in vivo studies in which antral gastrin release was strongly stimulated by bathing the antral mucosa with acetylcholine (3,121 and stimulated to a lesser extent by glycine [5] and cal- cium solutions [18]. In these experiments, the po- tency of acetylcholine is considerably greater than that of glycine and the two calcium chloride solu- tions. The apparent biphasic release of gastrin from the isolated antrum perfused with acetylcholine has been reported in an in vivo preparation [3]. Other polypeptide hormones have demonstrated a secon- dary release with continuing stimulation, but not after a single stimulus [19,20].

One can only speculate on the meaning of the secondary release of gastrin. It may well be that the two peaks elicited from the isolated perfused rat antrum represent different pools of stored gastrin, the first being more.available and the second being less available but ultimately finding its way to the surface of the cell. Alternatively, the first peak may

50mMACH + O.lmMESERlNEin KRE6’S RINGER A

I I I I 1

400 -

OL 2 4 6 6 IO 12 14 16 16 20 22 24 26

ml PERFUSED

Figure 6. Effluent gas&in levels after two lndlvtduai 10 mkndepertWonper&dsofanktokdedrat~antrum with 50 mM acetytcholhe + 0.1 mM eserlne In Krebs- Ringer’s A sotutlon. Nde the kqe secondpeak sqgestlngt that repeated stimutatlon brings about a new release of gastrlnhabnosteq4aiamounk memsuttls~alo?dlbw such experiments.

represent stored gastrin and the secondary peak newly synthesized gastrin. Supporting this concept is the fact that perfusion with glycine and the two calcium chloride solutions stimulated only mono- phasic release. The only substance that results in biphasic release of gastrin appears to be the physio- logic releaser, acetylcholine, and it is not unreason- able to assume that perhaps it stimulates synthesis as well. This is in accordance with the finding of a glucose-induced biphasic release of insulin in vitro [20], which is enhanced by adding acetylcholine. A similar biphasic release of insulin is obtained from isolated pancreas after vasoactive intestinal peptide infusion [21].

After antrum is minced, allowing more of the an- tral cells to come into contact with the perfusing so- lution, it is of interest that the initial peak remains the same but secondary release is increased, suggesting that perhaps the second pool may not be on the surface, but deeper in the cells. Synthesis may take place in cells not as readily available.

That eserine failed to release gastrin suggests that cholinergic nerves are not involved in gastrin release at least in this preparation, where such nerves, if necessary, are no longer viable.

This preparation simplifies the study of antral gastrin release, doing away with many of the inhib- iting feedback mechanisms, and may be profitably applied to the study of other endocrine organs as well.

50mM ACH i- 0. I mM ESERI NE inKREEk RINGER A

I

“Or T

; p<OOl

i T I I

P(O.05 ; T

IOO-

o- 2 4 6 6 IO I2 I4 I6 I6 20

m/ PERFUSED

Figurn 7. Mean effkmnt gastrin levekr f SEM after SIX dit- ferent p&Wons of mtnced rat antra in Krebs-Ringer’s A soiutiot% Noto that the secondary peak Is much more kientltiabh9 In the minced than In the unminced rat antrum preparatkm. ResuKs represent mean f SEM for nlne d/t- ferent experiments.

240 The Amuican Journal cl Surgafy

Antral Gastrin Release

Summary

In an in vitro preparation, isolated rat antra were perfused with different releasing agents in oxygen- ated Krebs-Ringer’s A solution. Upon perfusion with 50 mM acetylcholine (+O.l mM eserine), the mean gastrin levels in the effluent increased within 2 minutes to 4 times the basal level, while there was a smaller secondary release after 29 minutes. During perfusion with 50 mM glycine and 10 mM or 30 mM calcium chloride in Krebs-Ringer’s A solution, gas- trin levels in the effluent increased within 4 minutes to twice the basal level, without a secondary release of gastrin. Perfusion with eserine alone or with Krebs-Ringer’s A solution failed to release gastrin.

The apparent biphasic release of gastrin with acetylcholine suggests that antral gastrin may exist in two populations, one of which is readily accessible, the other less so. Alternatively, the initial peak may consist only of stored gastrin and the second peak newly synthesized gastrin, stimulated only by the physiologic stimulus (acetylcholine) but not by other releasers.

References

1. Korman MG, Soveny C, Hansky J: Effect of food on serum gastrin evaluated by radioimmunoassay. Gut 12: 619. 1971.

2. Jaffe BM, McGuigan JE, Newton WT: lmmunochemical mea- surement of the vagal release of gastrin. Surgery 68: 196, 1970.

3. Jackson BM, Reeder DD, Thompson JC: Dynamic character- istics of gastrin release. Am J Surg 123: 137. 1972.

4. Barreras RI: Calcium and gastric secretion. Gastroenferology 64: 1168,1973.

5. Elwin CE, Nilsson G: Comparison of the effect of gastric acid secretion of some protein compounds releasing gastrin. Acfa Physiol Stand 59: 37, 1963.

6. Csendes A. Grossman MI: D and L isomers of serine and alanine equally effective as releasers of gastrin. Experientia 28: 1306, 1972.

7. McGuigan JE, Jaffe EM, Newton WT: lmmunochemical mea- surement of endogenous gastrin release. Gsstmenterology 59: 499, 1970.

8. &gory RA: Gastric secretion: a review of its chief nervous and hormonal mechanisms. Surgical Physiology of the Gas- trointestinal Tract (Smith AN, ed). Edinburgh, The Royal College of Surgeons, 1962, p 57.

9. Nyhus LM, Chapman ND, DeVito RV, et al: An experimental study illustrating a new concept. Gastfoenterobgy 39: 582, 1960.

10. Bedi BS, Debas HT, Gillespie G, et al: Effect of bile salts on antral gastrin release. Gastroenterology 60: 256, 1971.

11. Baldessarini RJ, Kopin IJ: The effect of drugs on the release of norepinephrine-H3 from central nervous system tissues by electrical stimulation in vitro. J pharmacol Exp Ther 156: 31, 1967.

12. Robertson CR, Langlois K. Martin GG, et al: Release of gastrin in response to bathing the pyloric mucosa with acetylcholine. Am J Physioll63: 27, 1950.

13. Krebs HA: Body size and tissue respiration. Biochim Siophys Acta 4: 249, 1950.

14. McGuigan JE: lmmunochemical studies with synthetic human gastrin. Gashoenterology 54: 1005, 1968.

15. Dent RI, James JH, Wang CA, et al: Hyperparathyroidism, gastric acid secretion and gastrin. Ann Surg 176: 380, 1972.

16. Hunter WM, Greenwood FC: Preparation of iodine 131 labelled human goti hormone of high specific activity. Nature 192: 495. 1962.

17. Wesdorp RIG, Fischer JE: Assay of Gastrin in Hormones in Human Blood (Antoniades HN, ed). Cambridge, Harvard University Press, 1976, p 657.

18. Levant JA, Walsh JH, lsenberg JI: Stimulation of gastric se- cretion and gastrin release by single oral doses of calcium carbonate in man. N Engl J Med 289: 555. 1973.

19. Kolts BE, McGuigan JE: Radioimmunoassay of secretin: serum concentrations and half life in man. Gastroenterobgy 66: 849, 1974.

20. Sharp R, Culbert S, Cook J, et al: Cholinergic modification of glucose induced biphasic insulin release in vitro. J C/In Invest 53: 710, 1974.

21. Schebalin M, Brooks AM, Said SI. et al: The insulintropic effect of vasoactive intestinal peptide (VIP). Direct evidence from in vitro studies (abstract). Gastroenterology 66: 772. 1974.

volume 134, Augu8t 1977 241