Gastrin-Releasing Peptide in the Porcine Pancreas

GASTROENTEROLOGY 1987;92:1153-8

Gastrin-Releasing Peptide in the Porcine Pancreas SVEND KNUHTSEN, JENS J. HOLST, FURIO G. A. BALDISSERA, TINE SKAK-NIELSEN, STEEN S. POULSEN, STEEN L. JENSEN, and O. V AGN NIELSEN Institutes of Medical Physiology C and Medical Anatomy B. The Panum Institute. and Department of Surgery C. Rigshospitalet. University of Copenhagen. Copenhagen. Denmark

The presence of gastrin-releasing peptide (GRP) was studied in extracts of porcine pancreata. Gel filtration and high-pressure liquid chromatographic profiles of these extracts as monitored with both C-terminally and N-terminally directed radioimmunoassays against GRP showed pancreatic GRP to consist of one main form, namely the 27-amino acid peptide originally extracted from porcine stomach, and small amounts of a C-terminal fragment identical with the C-termina1 10-amino acid peptide. Gastrin-releasing peptide-like immunoreactivity released from the isolated perfused porcine pancreas during electrical vagal stimulation was shown by gel filtration to consist of the same two forms. By use of immunocytochemical techniques employing an antiserum directed against its N terminus, GRP was localized to varicose nerve fibers in close association with the exocrine tissue of the porcine pancreas in particular. Some fibers were found penetrating into pancreatic islets also. Immunoreactive nerve cell bodies as well as fibers were found within intrapancreatic ganglia. The potency of GRP in Received June 13. 1985. Accepted November 17. 1986. Address requests for reprints to: Svend Knuhtsen. c/o Associate Professor Jens J. Holst. Institute of Medical Physiology C. The Panum Institute. DK-2200 Copenhagen. Denmark. This study was supported by grants from the Danish Medical Research Council, Nordisk Insulinfond. Fonden til Lregevidenskabens Fremme. and Den Almindelige Danske Lregeforenings Fond. Svend Knuhtsen was supported by a scholarship from the Novo Research Institute. The authors thank Trine Eidsvold. Merete Hagerup. Letty Klarskov. Vibeke Jerris. and S0rn Hagen Nielsen for excellent technical assistance. Synthetic GRP(18-27) was a generous gift from Dr. J. E. Shively. City of Hope Research Institute. and Dr. J. R. Reeve. Center of Ulcer Research and Education. California. Synthetic GRP(15-27) and GRP(8-27) were generously given by Professor H. Yajima. Kyoto University. Japan. Synthetic GRP(1-13) was most kindly donated by Professor N. Yanaihara. Shizuoka College of Pharmacy. Japan. © 1987 by the American Gastroenterological Association 0016-5085/871$3.50

stimulating exocrine as well as endocrine secretion from the porcine pancreas, its presence in close contact with both acini and islets, and its release during vagal stimulation indicate that GRP may have a role in the parasympathetic regulation of endocrine and exocrine secretion from the pig pancreas. Gastrin-releasing peptide (GRP)-like and bombesinlike immunoreactivity have been shown to reside in varicose nerve fibers in the porcine pancreas (1,2). Recently, immunoreactive GRP was shown to be released into the bloodstream during electrical stimulation of the vagal nerves to the isolated perfused porcine pancreas (3). As GRP has potent effects on both endocrine (4-6) and exocrine (3) pancreatic functions, the peptide is likely to be of importance in the neural regulation of these functions. We therefore decided to study in more detail the localization and the molecular nature of tissue and vascularly released forms of pancreatic GRP in the pig, the animal from which GRP was originally isolated.

Materials and Methods Extraction of Tissue Biopsy specimens were obtained from four parts of the pancreas (head. body, tail, and duodenal lobe) of 3 anesthetized pigs. The specimens were immediately frozen between blocks of dry ice before extraction (7). In short, tissues were homogenized in a Waring kitchen blender (Waring Products Division, New Hartford, Conn.) in a mixture of ethanol, HCI, and water (42.3:1:10, vollvollvol; 4.4 mllg tissue) at -20°C. After 4 h at 4°C the homogenate was centrifuged. The supernatant was mixed with diethylether (E. Merck, Darmstadt. F.R.G.; 5 mllg tissue) at - 20°C and allowed to stand until two phases had

Abbreviation used in this paper: GRP. gastrin-releasing peptide.

1154

KNUHTSEN ET AL.

formed. The precipitate of the aqueous phase was dissolved in 8 moliL of urea.

Preparation of Samples Samples of venous effluent were collected during a 5-min electrical stimulation of the vagal nerves (8 Hz, 8 rnA, 4 ms, square wave impulses) from three isolated perfused porcine pancreata prepared and perfused as previously described with intact vagal innervation (8,9). The venous effluent was collected for I-min periods in chilled tubes (Minisorp, NUNC, Roskilde, Denmark) and was later assayed for GRP-like immunoreactivity with antiserum 1267. Twenty milliliters of the perfusate from each experiment was applied onto SepPak C-18 cartridges (Millipore, Waters Associates, Milford, Mass.). The columns were washed with 5 ml of 1% trifluoroacetic acid (Pierce Chemical Co., Rockford, Ill.) in water and adsorbed substances were eluted with 3 ml of 1 % trifluoroacetic acid in 96% ethanol. The samples were then evaporated to dryness under vacuum.

Chromatographic Procedures Gel filtration studies were performed using G-50 (fine grade) Sephadex columns measuring 16 x 1000 mm that were equilibrated and eluted at 4°C at a flow rate of 18 mllh. Pancreatic extracts were eluted in 0.5 mollL acetic acid and the fractions were lyophilized before assay for GRP-like immunoreactivity. Concentrated effluent from isolated perfused pancreata was eluted with 40 mmollL sodium phosphate, pH 7.4, containing in addition 1 giL human serum albumin (Behring-Werke, Marburg, F.R.G.) 66 mg/L aprotinin (Novo, Copenhagen, Denmark), 0.6 mmollL thiomersal, 6 giL NaCl. and 20 mmoliL ethylenedinitrilotetraacetic acid. For internal calibration, trace amounts of 125I-albumin and 22Na (both from The Radiochemical Centre, Amersham, U.K.) were applied with each sample. The coefficient of distribution (Kd) was calculated as Ve - Vo/(Vt - Vo)' where Ve is the volume of elution, Vo is the volume in which the albumin peak appeared, and Vt is the volume of the 22Na peak. Fractions containing GRP-like immunoreactivity from pancreatic extracts were collected and subjected to reverse-phase high-pressure liquid chromatography on a 4 x 250-mm Nucleosil C-18 column (particle size 5 fLm), using LKB (Bromma, Sweden) equipment (2150 HPLC pump, 2152 HPLC controller, 2040-203 low-pressure mixer valve). Elution phases were as follows: (a) 0.1 % trifluoroacetic acid in double-distilled water and (b) 0.05% trifluoroacetic acid in 60% ethanol (Merck, Art. No. 11727). Flow rate was 0.5 ml/min and eluted fractions, collected at I-min intervals, were evaporated to dryness under vacuum before assay for GRP-like immunoreactivity.

Radioimmunoassays C-terminal immunoreactivity was monitored using a previously described radioimmunoassay (10) employing antiserum 1267. For the production of antisera directed against the N-terminal part of the GRP molecule, 2 mg of GRP(1-16) (Peninsula Laboratories, San Carlos, Calif.) in 1

GASTROENTEROLOGY Vol. 92, No.5, Part 1

ml of 40 mmoliL sodium phosphate buffer, pH 7.0, was coupled to 5 mg of bovine serum albumin by addition of 50 p,l of 50% glutaraldehyde. After rotation for 3 h at room temperature the mixture was subjected to gel filtration on a 10 x 50-mm G-10 Sephadex column. The void volume was diluted to 11 ml with saline. One milliliter of an emulsion containing equal amounts of Freund's complete adjuvant and coupled peptide was injected intracutaneously at multiple sites on the backs of 3 rabbits. After six immunizations, antiserum 1652 was used in radioimmunoassay (final dilution 1:10,000). Gastrin-releasing peptide(1-27) (Peninsula Laboratories) was used both as standard and [iodinated as previously described (10)] as tracer. The assay conditions were similar to those described for the C-terminal assay (10). The 50% inhibition dose of binding of tracer in the assay was 560 and 660 pmol/L for GRP(1-27) and GRP(1-16), respectively. No cross-reaction was observed with GRP(8-27). GRP(15-27)' vasoactive intestinal polypeptide, cholecystokinin(1-33), or substance P in concentrations up to 10 nmollL. The detection limit was <25 pmollL. Interassay and intraassay variations were 12% (n = 6) and 8% (n = 10), respectively.

Immunohistochemistry Pancreata were isolated with preserved vascular supply from 3 anesthetized pigs (weighing 16 kg) (8). The pancreata were perfused from the aorta through the celiac trunk and the superior mesenteric artery with ice-cold saline followed by 4% paraformaldehyde (Merck) in 0.1 mol/L phosphate buffer, pH 7.4, for 10 min. Biopsy specimens of the organs were postfixed in the fixative for 24 h, after which they were transferred to 20% sucrose in 0.1 mollL phosphate buffer, pH 7.4. After another 24 h the tissue was frozen in melting Freon. The tissue was cut on a cryostat into slices of 10-16 p,m. Gastrin-releasing peptide-like immunoreactivity was visualized according to the principles of the peroxidase-anti peroxidase technique (11). Tissue was incubated for 20 h at 4°C with antiserum 1652 in a dilution of 1:1600 in a 50 mmol/L Tris buffer, pH 7.4, with 150 mmol/L of NaCI and 1 % Triton X-100. After another 2 h of incubation at room temperature the tissue was washed and incubated for 1 h with swine antirabbit immunoglobulin G (DakoPatt, Glostrup, Denmark). The tissue was washed again and incubated for 30 min with peroxidase-rabbit antiperoxidase complexes (DakoPatt). The peroxidase activity was developed with diaminobenzidine (5 mg/10 ml in Tris buffer with 0.01 % H 20 2 and without Triton) for 30 min. Controls included conventional staining and absorption controls. Cross-reactions with GRP(1-27), GRP(8-27), GRP(15-27). GRP(1-16), vasoactive intestinal polypeptide, and substance P were tested by mixing the antiserum with the peptides in concentrations of 10 and 100 p,g/ml and allowing a preincubation period of 24 h at 4°C before use. In the porcine pancreas the pancreatic islets are hard to identify without specific staining. After being stained for GRP-like immunoreactivity, slices from all parts of the pancreas (head, body, tail, and duodenal lobe) were then incubated at 4°C with an antiserum directed against insulin (antiserum 2004) produced in a guinea pig. The antiserum was used in a

GASTRIN-RELEASING PEPTIDE IN PORCINE PANCREAS

May 1987

pmol/g

""I

GRPI18-27)

GRPII-27)

I I

1.0

1155

pmol/l

O.

200

r

'loB

100

80

0.5

......................

0

pmollt

II

b.

100

'loB

TO 10

O~ I

o Figure 1. Average gel filtration profile of three extracts of pancreatic heads. Each run. has been divided into intervals of Kd = 0.05, and results are presented as average concentration of immunoreactivity within each interval. Filled dots represents N-terminal (non-bombesinlike) immunoreactivity, open dots represent C-terminal (bombesinlike) immunoreactivity. The column was calibrated with GRP(1-27) and GRP(18-27). Values are presented as mean ± SEM (n = 3).

dilution of 1:50,000 in Tris buffer. After 20 h the tissue was washed and incubated with peroxidase-labeled goat antiguinea pig immunoglobulin G (DakoPatt) for 30 min. After washing, the tissue was developed for 30 min with 3amino-9-ethyl-carbazole (20 mg dissolved in 2.5 ml N,Ndimethylformamide and diluted to 50 ml with a 50 mmollL acetate buffer, pH 5.0, with 0.01% H 2 0 2 ). During this procedure the f3-cells were stained red.

Table 1. Contents of Gastrin-Releasing Peptide-Like Immunoreactivity in Gel Filtration Profiles of Extracts of Different Parts of Porcine Pancreas Q GRP (1-27) Head Body Tail Duodenal lobe

5.9 5.4 6.5 6.4

± ± ± ±

2.3 1.6 1.7 1.9

GRP (18-27) 0.69 0.84 0.87 0.95

± ± ± ±

0.18 0.37 0.37 0.11

GRP, gastrin-releasing peptide. a From each run the content of gastrin-releasing peptide (1-27) (picomoles per gram tissue, wet weight) was calculated as total amount of immunoreactivity within Kd values of 0.35-0.55. The content of gastrin-releasing peptide (18-27) was calculated as total amount of immunoreactivity within Kd values of 0.65-0.95. Values are given as mean ± SEM (n = 3).

50

------",..-"

o

..........~ 20

30

40

50

60

70

80min

Figure 2. High-pressure liquid chromatographic profiles of GRPlike immunoreactivity from gel filtrated pancreatic extract. In (a) the applied sample was pooled fractions collected at Kd = 0.44. The major peak eluted close to the position of synthetic GRP(1-27) (marked by 1) and a mixture of equal amounts of synthetic GRP(1-27) and immunoreactivity collected at Kd = 0.44 resulted in one peak of the expected magnitude in the position of synthetic GRP(1-27). In (b) the applied sample consisted of pooled fractions collected at Kd = 0.80. The major peak of immunoreactivity eluted in the position of synthetic GRP(18-27) (marked by II) and application of a mixture of synthetic GRP(18-27) and immunoreactivity collected at Kd = 0.80 eluted as one peak of the expected magnitude in the position of synthetic GRP(18-27).

Results The gel filtration profile of the pancreatic extracts as determined with the C-terminal radioimmunoassay showed two immunoreactive forms: one eluting at the position of synthetic GRP(1-27) (Kd = 0.44) and the other at the position of synthetic GRP(18-27) (Kd = 0.80). The assay of N-terminal GRP-like immunoreactivity revealed two peaks: the first at Kd = 0.1 and the second at the position of

1156

GASTROENTEROLOGY Vol. 92, No.5, Part 1

KNUHTSEN ET AL.

.... ,

N

b

c 'iii .a

., E o

Vagal stirn.

CD

,...----,

a

recovered

I I

Immunoreactivity pmoLI L 20

'10

20

'10

10

10

;\

\

a

a

a

10

15

20 min

0-0

a

0.5

GRP(1-27). No immunoreactivity was found at Kd = 0.80 (Figure 1). Addition of synthetic GRP(1-27) and GRP(18-27) to the mixture of tissue and acid/ethanol medium before homogenizing resulted in an overall recovery of 109% ± 29% and 107% ± 28% (mean ± SEM, n = 3), respectively, after the extraction and gel filtration procedures. The contents of GRP-like immunoreactivities in the different parts of the pancreas did not differ (Table 1). Reverse-phase high-pressure liquid chromatography of fractions from Kd = 0.44 monitored with the C-terminal assay showed the major peak of immunoreactivity to elute in the same position as synthetic GRP(1-27) (Figure 2a). Several minor peaks eluted in front of the main peak. The immunoreactivity from Kd = 0.80 eluted in the high-pres~ure liquid chromatography system in the same position as synthetic GRP(18-27) (Figure 2b). Gastrin-releasing peptide-like immunoreactivity is released during electrical vagal stimulation (Figure 3a). The gel filtration profile of the released immunoreactivity exhibits two peaks, one corresponding to synthetic GRP(1-27) and the other to synthetic GRP(18-27) (Figure 3b). The immunohistochemical study showed N-terminal GRP-like immunoreactivity to reside in nerve cell bodies in both the pancreatic body and head. Absorbing the antiserum with GRP(1-27) removed all immunostaining in the slice immediately adjacent to the one with the immunoreactive ganglion cells. Some ganglia were found in which GRPcontaining nerve fibers surrounded immunonegative ganglion cells (Figure 4b). The vast majority of

1.0

Figure 3. (a) Concentration of gastrin-releasing peptide in venous effluent of isolated perfused pancreata before, during, and after electrical stimulation of the vagus nerves. Values significantly different (p < 0.05) from all values in the preceding 5-min period of "rest" as judged by the extended Friedman test (Reference 12) are marked by asterisks. Values are presented as mean ± SEM (n = 3). (b) Average gel filtration profile of three concentrated samples of venous effluent from the isolated perfused pancreata presented in (a) taken during electrical stimulation of the vagus nerves. Each run was monitored with the C-terminal assay of GRP and divided into intervals of Kd = 0.05. Results are given as percentage of total recovered immunoreactivity within each 0.05 interval of Kd and are presented as mean ± SEM (n=3).

GRP-containing nerve fibers in the pancreas were seen spreading out in the exocrine tissue to surround acini and exocrine ducts (Figure 4c), but some fibers were found in close contact with or even penetrating into pancreatic islets (Figure 4d). Preincubating the antiserum with GRP(1-27) or GRP(1-16) before incubation with the tissue abolished the immunostaining completely, whereas preincubating with GRP(1527), substance P, or vasoactive intestinal polypeptide had no effect on the immunostaining.

Discussion The present study shows porcine pancreatic GRP to consist of two forms as previously suggested by Ghatei et al. (13). One form is identical to GRP(1-27) as it was originally isolated from porcine stomach (14) and the other is identical to GRP (18-27), which has been isolated from dog intestine (15) and from porcine spinal cord (16). Gastrinreleasing peptide(4-27) was reported to be present (15) in the dog intestine. We have been unable to demonstrate this peptide in the porcine pancreas, although we cannot exclude the possibility that this peptide is represented by one of the minor peaks of immunoreactivity eluting in front of GRP(1-27) in the high-pressure liquid chromatography run shown in Figure 2a. The presence of these peaks may also represent small amounts of GRP(1-27) modified by a {3-aspartyl shift as described by McDonald et al. (17). At Kd = 0.1, a peak with only N-terminal GRP-like immunoreactivity occurs. This peptide could represent a biosynthetic GRP precursor. In a recent study

May 1987

GASTRIN-RELEASING PEPTIDE IN PORCINE PANCREAS

1157

b

d

Figure 4. Gastrin-releasing peptide-containing nerve fibers and nerve cell bodies immunostained according to the peroxidaseantiperoxidase method usi~g antiserum 1652 against the N-terminal (non-bombesinlike) sequence of GRP. The tissue was slightly counterstained with hematoxylin. (a) Intrapancreatic GRP-containing ganglion cells (magnification, x 100), (b) nerve fibers within an intrapancreatic ganglion (magnification, x 120), (c) GRP-containing nerve fibers surrounding pancreatic acini (magnification, x400), and (d) GRP-containing nerve fiber (arrow) crossing through pancreatic islet (,B-cells have been immunostained according to the method described in the text) (magnification, x400).

(18) of complementary deoxyribonucleic acid encoding for human GRP, it was shown that the precursor of GRP consists of GRP and a large C-terminal extension. The C-terminal extension may hinder the recognition of this precursor in the C-terminal assay, and thus explain the lack of C-terminal immunoreactivity at Kd = 0.1. As ribonucleic acid blot analysis of the tumor ribonucleic acid employed by Spindel et al. (18) shows the messenger ribonucleic acid for GRP to consist of two different forms, the possibility also exists that the peak at Kd = 0.1 represents the

final product of a posttranscriptional modification of the GRP ribonucleic acid-precursor lacking the sequence coding for the bombesinlike part of the molecule. As judged by gel filtration, GRP is being released from the porcine pancreas during stimulation of the vagus nerves in the same two forms as those found in tissue extracts, namely GRP(1-27) and GRP(18-27). Gastrin-releasing peptide(1-27) and GRP(20-27) are equally potent in stimulating pancreatic exocrine secretion (3) and there is no reason to believe that

1158 KNUHTSEN ET A1.

GRP(18-27) is less potent. Thus both molecular forms may be involved in the regulatory functions of GRP in the pancreas. Studies in calves (19) have shown bombesinlike immunoreactivity to occur in large quantities in peripheral plasma during electrical stimulation of the splanchnic nerves, whereas stimulation of the vagus nerves was without effect (20). We therefore find it of importance that our gel filtration studies have documented that the vag ally released pancreatic GRP-like immunoreactivity is due to release of GRP(1-27) and its C-terminal decapeptide, and not to some as yet uncharacterized bombesinlike substance. Our studies of the localization of GRP using an antiserum directed against the N-terminal (nonbombesinlike) part of GRP confirm the findings of a recent study (2) where an antiserum directed against the C-terminal (bombesinlike) part of the peptide was used. We find GRP in nerve fibers innervating the exocrine tissue, in nerve cell bodies, and in fibers contacting ganglion cells. Our finding that some GRP-containing fibers contact the pancreatic islets is slightly at variance with the findings of Moghimzadeh et al. (2), who were unable to identify nerve fibers in contact with the pancreatic islets. We have studied in particular the relationship between the GRP-containing nerve fibers and the islets by staining both the nerve fibers and the J3-cells together. Although the vast majority of the nerve fibers are found in the exocrine tissue, there seems to be no doubt that a GRP-ergic innervation of the islets also exists. Furthermore, by using an antiserum directed against the N-terminal part of the peptide instead of an antiserum directed against the C terminus, possible cross-reactions with other bombesinlike molecules [e.g., neuromedin B (21)] and with substance P have been avoided. Taken together, the potency .of GRP to stimulate both exocrine and endocrine pancreatic functions, its presence in two well-defined molecular forms in nerve fibers in close contact with both acini and islets, and its release from the organ during vagal stimulation make it reasonable to consider that GRP has a role in the parasympathetic regulation of the porcine pancreas.

References 1. Yanaihara N,Yanaihara C, Mochizuki T, Imura K, Fujita T, Iwanaga T. Immunoreactive GRP. Peptides 1981;2(Suppl 2):185-92. 2. Moghimzadeh E, Ekman R, Hakanson R, Yanaihara N,

GASTROENTEROLOGY Vol. 92, No.5, Part 1

Sundler F. Neuronal gastrin-releasing peptide in the mammalian gut pancreas. Neuroscience 1983;10:553-63. 3. Knuhtsen S, Holst n, Jensen SL, Knigge U, Nielsen OV. Gastrin releasing peptide: effect upon exocrine secretion, and release from isolated perfused porcine pancreas. Am J Physiol 1985;248:G281-6. 4. McDonald TJ, Ghatei MA, Bloom SR, et al. A qualitative comparison of canine plasma gastroenteropancreatic hormone responses to bombesin and the porcine gastrin releasing peptide (GRP). Regul Pept 1981;2:293-304. 5. Bloom SR, Edwards AV, Ghatei MA. Effects of certain metabolites on pancreatic endocrine responses to gastrin-releasing peptide in conscious calves. J Physiol 1984;346:547-55. 6. Knigge U, Holst n, Knuhtsen S, et al. Gastrin-releasing peptide: pharmacokinetics and effects on gastro-entero-pancreatic hormones and gastric secretion in normal men. J Clin Endocrinol Metab 1984;59:310-5. 7. Newgard CB, Holst n. Heterogeneity of somatostatin like immunoreactivity in extracts of porcine, canine, and human pancreas. Acta Endocrinol (Copenh) 1981;98:564-72. 8. Jensen SL, Kuhl C, Nielsen OV, Holst n. Isolation and perfusion of the porcine pancreas. Scand J Gastroenterol (Suppl) 1976;37:57-61. 9. Holst n, Fahrenkrug J, Knuhtsen S, Jensen SL, Poulsen SS, Nielsen OV. Vasoactive intestinal peptide (VIP) in the pig pancreas: role of VIPergic nerves in control of fluid and bicarbonate secretion. Regul Pept 1984;8:245-59. 10. Knuhtsen S, Holst n, Knigge U, Olesen M, Nielsen OV. Radioimmunoassay, pharmacokinetics, and neuronal release of gastrin-releasing peptide in anesthetized pigs. Gastroenterology 1984;87:372-8. 11. Sternberger 1. Immunohistochemistry. Englewood Cliffs, N.J.: Prentice Hall, 1974:129-71. 12. Conover WJ. Practical nonparametric statistics. 2nd ed. New York: John Wiley & Sons, 1980. 13. Ghatei MA, George SK, Major JH, et al. Bombesin-like immunoreactivity in the pancreas of man and other mammalian species. Experientia 1984;40:884-6. 14. McDonald TJ, J0rnwall H, Nilsson G, et al. Characterization of a gastrin releasing peptide from porcine non-antral tissue. Biochem Biophys Res Commun 1979;90:227-33. 15. Reeve JRJr, Walsh JH, Chew P, et al. Amino acid sequences of three bombesinlike peptides from canine intestine extracts. J BioI Chern 1983;258:5582-8. 16. Minamino N, Kangawa K, Matsuo H. Neuromedin C: a bombesin-like peptide identified in porcine spinal cord. Biochem Biophys Res Commun 1984;119:14-20. 17. McDonald T, J0rnwall H, Tatemoto K, Mutt V. Identification and characterization of variant forms of the gastrin releasing peptide (GRP). FEBS Lett 1983;156:349-56. 18. Spindel ER, Chin WW, Price J, et al. Cloning and characterization of cDNAs encoding human gastrin-releasing peptide. Proc Natl Acad Sci USA 1984;81:5699-703. 19. Bloom SR, Edwards AV, Ghatei MA. Neuroendocrine responses to stimulation of the splanchnic nerves in bursts in the conscious adrenalectomized calf. J Physiol 1984;346: 519-31. 20. Adrian TE, Bloom SR, Edwards AV. Neuroendocrine responses to stimulation of the vagus nerves in bursts in conscious calves. J Physiol 1983;344:25-35. 21. Minamino N, Kangawa K, Matsuo H. Neuromedin B: a novel bombesin-like peptide identified in porcine spinal cord. Biochem Biophys Res Commun 1983;114:541-8.