Endocrinol. Japon. 1990, 37 (6), 857-865

Effects of Neuromedin B and GRP-10 on and Release from Cultured Tumor Cells of a Malignant Gastrinoma

KOICHI KAWAI, HIDEHITO MUKAI1, KENJI YUZAWA2, SEIJI SUZUKI, NOBUAKI KUZUYA, KEIJI FUJII3, EISUKE MUNEKATA1 AND KAMEJIRO YAMASHITA

Department of Internal Medicine, Department of Surgery2, Institute of Clinical Medicine, Department of Pathology3, Institute of Basic Medical Science, Institute of Applied Biochemistry1, University of Tsukuba

Abstract

A study relating to gastrin release from gastrinoma cells by neuromedin is and C-terminal decapeptide of gastrin releasing (GRP-10) has not yet been reported. Therefore, we studied the effects of neuromedin B and GRP- 10 on gastrin release from cultured dispersed cells prepared from both the primary tumor in the pancreas and the metastatic tumor in the liver from a case of malignant Zollinger-Ellison syndrome. Both the primary and metastatic tumors obtained by a curative operation contained similar concentrations of gastrin and glucagon, whereas the primary tumor contained 10 times more insulin than the metastatic tumor. Gastrin release from cultured cells of both tumors was suppressed by 0.1 and 10nM neuromedin B and tended to be suppressed by 0.1-10nM GRP-10. However, insulin release from cultured cells of the pancreatic tumor was stimulated by GRP-10, but not by neuromedin B. These results might suggest that receptor function for the family is abnormal in gastrinoma cells in both primary and metastatic tumors, and that a major source of insulin secretary cells is the contaminated normal islet cells in the primary tumor.

Zollinger-Ellison syndrome is charac- Japan where only about 100 cases were terized by recurrent peptic ulceration, reported prior to 1982 (Kishimoto and marked gastric acid hypersecretion and islet Miyoshi, 1983). Gastrin release from this cell tumors secreting gastrin (Zollinger and tumor is abnormal; it is not suppressed Ellison, 1955; McGuigan and Trudeau, by (Lamers and Van Tongeren, 1969). Its incidence is low, particularly in 1977; McGuigan and Wolfe, 1980), and is weakly stimulated by bombesin (Basso et Received June 22, 1990 al., 1981; Jansen and Lamers, 1982). Address for correspondence and reprint requests: Neuromedin B and C-terminal decapeptide KOICHI KAWAI, M. D. & Ph. D. Institute of Clinical Medicine, University of Tsukuba of gastrin releasing peptide (GRP-10) are Tsukuba, Tsukuba, Ibaraki-ken 305, Japan mammalian bombesin-like peptides (Mina- Endocrinol. Japon. 858 KAWAI et al. December 1990 mino et al., 1983; Okada et al., 1984;Reeve mission to the University Hospital. How- et al., 1983) which stimulate the release of ever, the duodenal ulcer recurred and a gastrin, insulin and glucagon in a similar high plasma gastrin level was found (1090 manner to bombesin (Spindel, 1986; Kawai pg/ml, normal value 40-20pg/ml). et al., 1988). However, no study relating The physician suspected Zollinger-Ellison to gastrin release from gastrinoma by these syndrome and referred her to the University peptides has yet been reported. Hospital. She had no family history of The effects of neuromedin B and GRP- peptic ulcer diseases. 10 on gastrin release from cultured dispersed The patient was 147 cm in height and cells, from both the primary tumor in the 54kg in weight, and seemed to be healthy. pancreas and the metastatic tumor in the Her blood pressure was 120/88mmHg. No liver obtained from a case of malignant abnormal findings were observed in the

Zollinger-Ellison syndrome, are reported. physical examination except for an operation scar in her abdomen. The blood cell count on admission showed her to be Case Report slightly anemic; RBC was 350•~104/mm3, Hb, 10.0g/dl, and hematocrit, 39.9%. A 43-year-old woman had suffered from Urinalysis and blood chemistry yielded epigastric pain and diarrhea for 7 years. normal values ; serum total protein was A duodenal ulcer was diagnosed on the 7.0g/dl, albumin, 4.0g/dl, Ca++ 8.9mg/di, the basis of her symptoms and antiacids and inorganic phosphate 4.5 mg/dl. prescribed by a physician. The symptoms Endocrinological data on admission are sometimes relapsed and she received sub- shown in Table 1. The basal gastrin level total gastrectomy 9 months before her ad- was extremely high and it was not sup-

Table 1. Endocrinological data on admission Gastrin: Secretin injection test;

Mixed meal loading;

Oral glucose tolerance test:

Basal hormone levels in plasma: Thyroxine 8.2ƒÊg/dl, Triiodothyronine 93ng/dl, TSH 1.7ƒÊU/ml, 40pg/ml, 0.3ng/ml (RIA with C-terminal-specific antibody), Growth hormone 3.6ng/ml, ACTH 35ƒÊg/ml, 21ng/ml, Luteinizing hormone 15.0mIU/ml, Follicule- stimulating hormone 11.2mIU/ml, Cortisol 9.1ƒÊg/dl, Aldosterone 15.0ng/dl. Vol.37, No.6 GASTRINOMA AND NEUROMEDIN B, GRP-10 859 pressed to normal by injecting secretin sound of the neck were normal. (2 U/kg Secrepan(R), Eisai, Tokyo, Japan). From these results, the diagnosis of The response of gastrin release to a mixed Zollinger-Ellison syndrome was established meal was low. A 75g oral glucose loading with a primary tumor in the pancreas and test revealed an abnormal increase in plasma multiple metastatic tumors in the liver. No glucose level, the so called "oxyhyper- positive data for multiple endocrine neo- glycemia" type, probably due to subtotal plasia were found. She received a resection gastrectomy. Plasma insulin increased with of the body and tail of the pancreas a pattern of delayed response from its high together with the spleen and residual fasting level (normal value 3-18 pU/ml). stomach, and a partial resection of the Basal levels of thyroid, parathyroid, pitui- liver and an enucleation of hepatic metastatic tary and adrenal hormone were normal. tumors. Urinary excretion of epinephrine, norepine- The primary tumor nodule was a phrine, metanephrine and normetanephrine solitary, solid and yellowish mass; 55mm were normal at 8.1, 77.4, 40 and 100ƒÊg/ in size at its largest diameter and weighing day, respectively. 69 gr. Fig. 1 shows the microscopic find- Basal acid output was 35.2mEq/hr and ings for the primary and metastatic tumors. the maximum acid output in response to tetra- Histologically, the arrangement of the gastrin (4ƒÊg/kg Gastpsin(R), Nihonkayaku, primary tumor cells exhibited combinations Tokyo, Japan) was 39.6mEq/hr. Abdominal of a trabecular pattern, solid nests and a ultrasound, CT scanning and angiograms gyriform pattern. The tumor cells varied revealed a tumor in the pancreatic body in shape, i. e. ovoid to polygonal, and the and multiple metastatic tumors in the liver. tumor cells were larger than normal islet CT scanning of the pituitary and ultra- cells. Cytoplasm of the tumor cells was

Fig. 1 A

Fig. 1. Histological observation of the

primary tumor in the pancreas (A),

capsular margin of the primary tumor

(B), and the me- tastatic tumor in

the liver (C). Hema- toxin-eosin staining

(•~100). Primary tumor (A): Solid Endocrinol. Japon. 860 KAWAI et al. December 1990

Fig. 1 B

B

Fig. 1 C

C

tumor growth with scanty stroma. Note the polygonal shaped cells and the round nuclei. Capsular margin of the primary tumor (B): The compressed remaining Langerhans' islets are seen at the right hand side of the photo. Liver metastasis (C): The tumor with thin fibrous capsule exhibiting a solid pattern and scanty stroma. Vol.37, No.6 GASTRINOMA AND NEUROMEDIN B, GRP-10 861 abundant, and the staining characteristics fetal calf serum and gentamicin (40ƒÊg/ml) in were granular and acidophilic. The nuclei humidified, 95% air and 5% CO2 at 37•Ž for were often giant hyperchromatic. Mitotic 3 days. figures were rarely seen. In the primary Hormone release from cultured cells nodule, capsular invasion was prominent, After 3 days' culture, the medium was and the tumor cells protruded in the pan- aspirated and the wells were washed twice with creatic duct, obstructing the lumen. 1ml DMEM supplemented with 10% fetal calf After the operation, her plasma gastrin serum and gentamicin (40ƒÊg/mi). The cells level fell to 50pg/ml and she was dis- were then incubated in 1ml of the above charged. medium at 37•Ž for 24 hr to determine the basal hormone release. The viability of the cells was > 90% by trypan blue exclusion. After incubation, the medium was collected and Materials and Methods washed twice with the same medium. The cells were then incubated in 1 ml of the same medium Materials containing neuromedin B or GRP-10 at 37•Ž Neuromedin B and GRP-10 used in this for 1h. After the incubation, the medium was study were synthesized by the solid phase collected and the cells were again incubated in method (Mukai et al., 1987). They were purified 1ml of the medium containing the same dose by gel filtration and reverse phase high- of neuromedin B or GRP-10 at 37•Ž for 1 hr. performance liquid chromatography described The medium was collected and stored at 40•Ž previously (Mukai et al., 1987), and proved to until the hormone assaay. be at least 98% pure. Hormone assay

Extraction of hormone from tumors The gastrin level was determined with a commercial kit (Gastrin-RIA kite, Dainabot 0.3g each of primary pancreatic tumor and hepatic metastatic tumor were placed into 3ml Radioisotope, Tokyo, Japan). The antiserum of boiling 0.1 N acetic acid for 10 min and in this immunoassay is directed against the C- homogenized with a Polytron (Brinkman). The terminal residues of gastrin. Insulin and homogenate was centrifuged for 30 min at glucagon levels were measured by RIAs accord- 100,000•~g at 4•Ž. The supernatant was ing to the method of Herbert et al.(1965), and lyophylized and dissolved in 0.2M glycine Faloona and Unger (1974) with E-7 antibody buffer pH 8.8 containing 0.25% human serum (kindly donated by Dr. H. von Schenck, von albumin and 1% sheep serum. Schenck, 1977), respectively. Neither neuromedin B nor GRP-10 cross-reacted with the immuno- assay systems. Cell culture of tumors 0.4g each of primary pancreatic tumor and Statistical analysis metastatic hepatic tumor were chopped into All data are presented as the mean-1-SEM. small pieces with scissors in Ca++-and Mg++- Statistical comparisons between groups were free Dulbecco's modified Eagle's medium performed by Student's t-test for unpaired data. (DMEM) (Gibco, Grand Island, NY, USA). After centrifugation at 600•~g for 5min at 4•Ž, P <0.05 was considered significant. the pieces were dispersed by DMEM containing 0.1% collagenase (Type IV, Worthington, Free- hold, NY, USA) for 2hr at 37•Ž. The dispersed Results cells were washed 5 times with Ca++-and Mg++- free DMEM. The washed dispersed cells were Hormone content of tumors plated in tissue culure wells previously coated There was 174ng/g (tissue wet weight) with a thin film of polymerized type a collagen gastrin in the primary tumor and 219ng/g prepared from rat tails (24mm multiwell plates, Corning Glass works, Corning, NY, USA) and in the metastatic tumor. The amount of incubated in DMEM supplemented with 10% insulin in the primary tumor was 198ng/g Endocrinol. Japon. 862 KAWAI et al. December 1990 tissue wet weight and that in the metastatic cultured primary tumor cells, but insulin tumor was estimated to be 28ng/g because was found not to be released from cultured it was close to the sensitivity limit of the metastatic tumor cells. Glucagon release RIA. The amount of giucagon in the primary was negligible in both tumor cells (data tumor and the metastatic tumors was 8.6 not shown). and 7.3ng/g tissue wet weight, respectively. The mean gastrin concentrations in the 24hr incubation medium from the primary Basal hormone release from cultured and metastatic tumors obtained from 28 tumor cells tissue culture wells were 1495•}94pg/ml Table 2 shows the hormone content in and 812•}41pg/ml, respectively. The mean the basal medium obtained after 24hr- insulin concentration in the 24 hr incubation incubation and in the mediums of following medium of the primary tumor from 28 incubation twice for 1hr each time. Gastrin tissue culture wells was 3657•}549pg/ml. and insulin were clearly released from That of the metastatic tumor was un-

Table 2. Hormones in the medium of control experiment (without neuromedin B or GRP-10) obtained after 24 hr incubation and following incubation twice for 1hr each time.

Mean of values from 4 tissue culture wells, *Nd; not detectable.

Fig. 2. Effects of neuro- medin B (NMB) and

GRP-10 on gastrin re- lease from cultured pri- mary tumor cells (• )

and metastatic tumor cells

(■). The results are shown as percent changes in

gastrin content in the first 1hr-incubation medium

plus that in the second 1hr-incubation medium against gastrin content in the 24hr basal incuba- tion medium (without NMB orGRP-10). N= 4, mean•}SEM. *P<0.05 vs. control (vehicle of NMB or GRP-10). Vol.37, No.6 GASTRINOMA AND NEUROMEDIN B,GRP-10 863 detectable. the glucagon concentration in the incu- bation medium itself was subtracted (data Effects of neuromedin B and GRP-1O not shown). on hormone release from cultured tumor cells Gastrin release from both primary and Discussion metastatic tumor cells decreased significantly with 0.1nM neuromedin B. 1nM neuro- The hormonal aspects of a case of medin B and 10nM neuromedin B also malignant gastrinoma with a primary tumor inhibited gastrin release from metastatic in the pancreas body and multiple metastatic tumor cells and from primary tumor cells, tumors in the liver has been investigated. respectively. 0.1-10nM GRP-10 tended to Both the primary and metastatic tumor inhibit gastrin release from both cultured cells contained gastrin and glucagon to the primary and metastatic tumor cells (Fig. 2). same extent, whereas the insulin concen- Insulin release from cultured primary tumor tration in the primary tumor cells was cells was stimulated by 0.1 and 1nM GRP- almost 10 times higher than that in the 10 and was not affected by 0.1-10nM metastatic tumor. Gastrin release from both neuromedin B.(Fig. 3). Glucagon release tumor cells was suppressed by neuromedin tended to be suppressed by neuromedin B B and tended to be suppressed by GRP-l0. and GRP-10, but the values were too However, insulin release from pancreatic small to make a statistical analysis when tumor cells was stimulated by GRP-10 and

Fig. 3. Effects of neuro- medin B (NMB) and GRP-10 on insulin release from cultured primary tumor cells. The results are shown as precent changes in insulin content in the first 1hr-incubation medium plus that in the second 1hr-incubation medium against insulin content in the 24hr basal in- cubation medium (with- out NMB or GRP-10). N=4, mean•}SEM. *P< 0.05 vs. control (vehicle of NMB or GRP-10). Endocrinol. Japon. 864 KAWAI et al. December 1990 not by neuromedin B. antral origin (Basso et al., 1981 and Jansen Neuromedin B and GRP-10 are mam- and Lamers 1982). Gastrin release was not malian bombesin-like peptides which stimu- clearly stimulate by the infusion of bombesin late the release of gastrin, insulin and in 4 patients out of 8 with a non-operated glucagon. The potency of GRP-10 was gastrinoma (Jansen and Lamers, 1982), or in found to be over 10 times greater than 2 patients out of 16 with Zollinger-Ellison that of F neuromedin B in our previous syndrome (Basso et al., 1981). In addition, study (Kawai et al., 1988). These effects of the response seems to be weak in patients bombesin-like peptides are considered to be with liver metastases (Basso et al., 1981). result of their binding to receptor sites. Since the effects of bombesin-like peptides However, in this study gastrin release was on gastrointestinal and pancreatic hormone suppressed by neuromedin B, which might be release are thought to be similar, the sup- ascribed to an abnormality in the function pression of gastrin release by neuromedin of the receptor for bombesin-like peptides B in this in vitro study is somewhat in tumor cells. The absence of the sup- different from previous in vivo studies, sug- pression of gastrin release by secretion in gesting poor response of gastrinoma cells to this patient and its small increase after the bombesin. Although the data presented are ingestion of a mixed meal (Table 1) suggest these for only one malignant gastrinoma, it a similar abnormality. The abnormal re- is suggested that neuromedin B might gulation of hormone release in endocrine suppress gastrin release from a maglinant tumors is sometimes obserbed, i. e.; no gastrinoma taking into account of the stimulation of insulin release by glucose in abnormal response to bombesin in patients insulinoma, no suppression of growth with malignant gastrinoma (Basso et al., hormone release by glucose and an abnormal 1981). Therefore, it is possible that the stimulation of it by TRH in acromegaly, infusion of neuromedin B may be useful in although the precise mechanisms of these distinguishing between malignant and benign abnormalities has not yet been understood. gastrinoma. In this patient, we have not However, the response of insulin release examined the response of the plasma from cultured cells to GRP-10 was normal. gastrin level to neuromedin B or GRP-10, The amount of insulin in tumors was about because the use of these peptides was limited 100 times lower than that in the normal to animals in this study. pancreas, and the amount in the metastatic tumor in the liver was about 15% that in the primary tumor in the pancreas. These Acknowledgements results suggest that the majority of insulin This report was supported by a research grant from the tumor in the pancreas is derived for a Special Research Projects on Instinct from from contaminated islet cells as shown in the University of Tsukuba, Grants in Aid for Fig. 1-B, although the tumor was well Scientific Research from the Ministry of Edu- defined from a normal pancreas. The cation, Science and Culture, Japan (No. 59440085 amount of glucagon in the primary tumor and No. 59340034), and a research grant from was also slightly higher than that in the Ajinomoto Co. Ltd, Japan. metastatic tumor. Previous in vivo studies of gastrinoma References demonstrated that the plasma gastrin level of patients increased little in response to an Basso, N., E. Lezoche, A. Materia, E. Passaro intravenous infusion of bombesin, compared Jr and V. Speranza (1981). Studies with with patients with hypergastrinemia of bombesin in the Zollinger-Ellison syndrome. Vol.37, No.6 GASTRINOMA AND NEUROMEDIN B,GRP-10 865

Br. J. Surg. 68, 97-100. McGuigan, J. E. and. M. M. Wolfe (1980). Faloona, G. R. and R. H. Unger (1974). Secretin injection test in the diagnosis of Glucagon. In: Methods of hormone radio- gastrinoma. Gastroenterology 79, 1324-1331. immunoassay (B. M. Jaffe and H. R. Behrman Minamino, N., K. Kangawa and H. Matsuo eds.), Academic Press, New York, pp. 317- (1983). Neuromedin B : A novel bombesin- 330. like peptide identified in porcine spinal cord. Herbert, V., K.-S. Law. C. W. Gottlieb and S. Biochem. Biophys. Res. Commun. 114, 541- J. Bleicher (1965). Coated charcoal immuno- 548. assay of insulin. J. Chin. Endocrinol. Metab. Mukai, H., K. Kawai, Y. Suzuki K. Yamashita 25, 1375-1384. and E. Munekata (1987). Stimulation of dog Jansen, J. B. M. and C. B. H. Lamers (1982). gastropancreatic hormone release by neuro- Serum gastrin responses to bombesin and food medin B and its analogues. Am. J. Physiol. in patients with hypergastrinemia. Dig. Dis. 252, E765-771. Sci. 27, 303-307. Okada, M., S. Kimura, K. Goto, I. Kanazawa, Kawai, K., Y. Chiba, H. Mukai, E. Munekata S. Ohashi, M. Shiraki and E. Munekata and K. Yamashita (1988). Effects of neuro- (1984). Structure-activity relationships of medin B, gastrin-releasing peptide-10 and bombesin like isolated from their fragment peptides on secretion of gastro- porcine spinal cord. In: Peptide chemistry intestinal and pancreatic hormones in dogs. 1983 (E. Munekata ed.). Protein Research Acta. Endocrinol. 117, 205-213. Foundation, Osaka, pp.111-114. Kishimoto, S. and S. Miyoshi (1983). Zollinger- Reeve, J. R., J. H. Walsh, P. Chew, D. Hawke Ellison syndrome. In : Pancreatic diseases, its and J. E. Shively (1983). Amino acid principles and Clinics (S. Naito ed.) Dobun- sequences of three bombesin-like peptides from shoin, Tokyo, pp.835-856.(In Japanese) canine intestine extracts. J. Biol. Chem. 258, Lamers, C. B. H. and J. H. M. Van Tongeren 5582-5588. (1977). Comparative study of the value of Spindel, E.(1986). Mammalian bombesin-like the calcium, secretin, and meal stimulated peptides. Trends Neurosci. 9, 130-133. increase in serum gastrin to the diagnosis of Von Schenck, H.(1977). Production and the Zollinger-Ellison syndrome. Gut 18, 128- characterization of an antiserum against pan- 134. creatic glucagon. Clin. Chim. Acta 80, 455- McGuigan, J. E. and W. L. Trudeau (1969). Im- 463. munochemical measurement of elevated levels Zollinger, R. M. and E. H. Ellison (1955). of gastrin in the serum of patients with pan- Primary peptic ulcerations of the jejunum creatic tumors of the Zollinger-Ellison variety. associated with islet cell tumors of the pan- N. Eng. J. Med. 278, 1308-1313. creas. Ann. Surg. 142, 709-728.