GASTROENTEROLOGY
1991;101:303-311
Establishment and Characterization of a Human Carcinoid in Nude Mice and Effect of Various Agents on Tumor Growth B. MARK EVERS, COURTNEY M. TOWNSEND, Jr., J. ROBERT ERICK ALLEN, STEPHEN C. HURLBUT, SUN WHE KIM, SRINIVASAN RAJARAMAN, POMILA SINGH, JEAN CLAUDE REUBI, and JAMES C. THOMPSON
UPP,
Departments and Sandoz
Texas:
of Surgery and Pathology, Research Institute, Berne,
University Switzerland
The authors have established a long-term tissue culture cell line (BON) derived from a metastatic human carcinoid tumor of the pancreas. The cells have been in continuous passage for 46 months. Tissue culture cells produce tumors in a dosedependent fashion after SC inoculation of cell suspensions in athymic nude mice. BON tumors, grown in nude mice, are histologically identical to the original tumor; they possess gastrin and somatostatin receptors, synthesize serotonin and chromog ranin A, and have a doubling time of approximately 13 days. The antiproliferative effects of the longacting somatostatin analogue, SMS 201-995 (300 @kg, t.i.d.), and 2% ac-difluoromethylornithine on BON xenografts in nude mice were examined. Tumor size was significantly decreased by day 14 of treatment with either agent and at all points of analysis thereafter until the animals were killed (day 33). In addition, tumor weight, DNA, RNA, and protein contents were significantly decreased in treated mice compared with controls. Establishment of this human carcinoid xenograft line, BON, provides an excellent model to study further the biological behavior of carcinoid tumors and the in vivo effect of chemotherapeutic agents on tumor growth.
Cfrom neuroectodermal (1,2). Primarily found in
arcinoid tumors are endocrine neoplasms derived cells of the neural crest the gastrointestinal (GI) tract, these tumors are uncommon, but not rare, with a reported incidence in autopsy series ranging from 0.1% to 1.4% (1). Originally called “Karzinoide” by Oberndorfer (3) to describe its apparent benign na-
of Texas
Medical
Branch,
Galveston,
ture, modern immunohistochemical techniques have allowed identification and isolation of the vasoactive substance serotonin [ti-hydroxytryptamine (5-HT)] and a host of biogenic amines and hormones produced by carcinoid tumors (1,2). Treatment of patients with carcinoid tumors has been hampered by the lack of an animal model. In the majority of patients, treatment of metastatic carcinoid tumors has been directed largely toward relief of the debilitating sequelae of the carcinoid syndrome by means of blocking agents [for example, methysergide (4), cyproheptadine (4,5), and ketanserin (6,7)] or the inhibitory hormone, somatostatin (8-16). Antiproliferative agents that are commonly used to slow tumor growth include combinations of streptozotocin, 5-fluorouracil, doxorubicin, or cyclophosphamide (5,1719). However, most carcinoid tumors respond poorly to cytotoxic therapy with these agents (1,4,18-19). In this study, we report the first in vivo establishment of a long-term xenograft cell line of a human pancreatic carcinoid tumor. We have characterized tumor growth rate in vivo, the status of gastrin and somatostatin receptors, the morphology, and the production of peptides and amines. In addition, we have examined possible antiproliferative effects of the longacting somatostatin analogue, SMS 201-995, and ol-difluoromethylornithine (DFMO) (either alone or in combination) on growth of the carcinoid tumor. Abbreviations usedin thispaper: DFMO, a-difluoromethylornithine; DMEM, Dulbecco’s modified Eagle medium: FCS, fetal calf serum; S-I-IT,5-hydroxytriptamine. o 1991 by the American Gastroenterological Association 9036-5085/91/$3.09
304
GASTROENTEROLOGY
EVERS ET AL.
Materials
and Methods
Surgical Specimen Collection and Maintenance of Tumor Line The operative specimen of a peripancreatic lymph node was sterilely obtained in April 1986 from a 28-year-old man with metastatic carcinoid tumor of the pancreas who had come into the hospital with symptoms of obstructive jaundice and diarrhea. Before exploratory laparotomy the patient had not received any interventional treatment. Final diagnosis of a metastatic carcinoid tumor was made by histological examination of the lymph node and positive staining for neuron-specific enolase and 5-HT. A portion of the lymph node was washed in saline, minced, and tumor fragments placed in Dulbecco’s modified Eagle medium (DMEM; Gibco, Grand Island, NY) and F12K (Gibco) in a 1:l ratio supplemented with,lO% (vol/vol) fetal calf serum (FCS; Hyclone Laboratories, Logan, UT) and 1% gentamicin. The cells were grown at 37°C in an atmosphere of 95% air and 5% CO,. Cells were routinely passed by removing the medium and overlaying the cell monolayer with 0.25% trypsin:O.l% ethylenediaminetetraacetic acid (EDTA). The elimination of fibroblasts from the stroma of the tumor tissue was accomplished by brief exposures to 0.06% trypsin:0.02% EDTA. This procedure was repeated until no further fibroblast growth was observed. Tumor cells from passage 5 were frozen after successful in vitro adaptation and removal of all fibroblasts. The cell line is presently maintained in DMEM and F12K growth medium supplemented with 10% FCS; it is passed at a 1:2 ratio when cells reach 80% confluence. Cell cultures are routinely monitored for mycoplasma contamination, and no mycoplasma growth has been detected.
Animals Male athymic nude mice (Balb/c, 20-25 g, 3-4 weeks of age; Life Science, St. Petersburg, FL) were housed under specific pathogen-free conditions in a temperaturecontrolled isolation unit with IX-hour light-dark cycles in accordance with the National Research Council’s Guide for the Care and Use of the Nude Mouse in Biomedical Research (20). The mice were fed a standard chow (Autoclavable Rodent Chow no. 5010; Ralston Purina, St. Louis, MO) and sterile water, both given ad libitum.
Solid Tumors in Nude Mice and In Vitro Growth Rate BON cells (passage 8) were harvested from subconfluent cultures by a l-minute treatment with 0.25% trypsin and 0.1% EDTA. Single-cell suspensions (5 X 106, 1 X lo’, and 2 x 10’) in RPM1 1640 without serum (total volume, 0.1 mL) were injected SC at a single site in the dermis of nude mice (n = 10 mice/group). Tumor size (longest perpendicular diameter) was measured biweekly with Vernier calipers (Mitutoyo Corp., Tokyo, Japan) accurate to 0.5 mm. Surface areas were calculated as the product of the two greatest perpendicular tumor diameters and were expressed as
Vol. 101, No. 2
square millimeters. Tumor areas were plotted on semilogarithmic paper, and the tumor doubling time (defined as time in days required for mean tumor area to double during logarithmic growth) was determined directly from the graph. On day 48, mice were killed and tumors from passage 8 were removed for light and electron microscopy, immunohistochemistry, and measurement of gastrin receptors.
Light and Electron Microscopy For light microscopy, tumor tissue blocks measuring 1 x 1 x 0.5 cm were fixed in 10% neutral buffered formalin for 6-8 hours, processed routinely, and embedded in paraffin Sections (4 pm thick) were stained with H&E and examined. For transmission electron microscopy, tissue blocks (1 mm3) were fixed in half-strength Karnovsky’s fixative for 4 to 6 hours, postfixed in osmium tetroxide (l%), and embedded in Epon. Sections (70 nm thick) were stained with uranyl acetate and Reynold’s lead citrate and examined with a Philips 410 electron microscope (Philips Medical Systems Inc., Shelton, CT).
Zmmunohistochemical
Studies
Formalin-fixed, paraffin-embedded tissue samples were studied by means of a three-layer immunohistochemical method. The primary antibodies used were murine monoclonal antibody to chromogranin A (Boehringer Mannheim, Indianapolis, IN) and monospecific polyclonal antibodies to serotonin, substance P, pancreatic polypeptide, vasoactive intestinal peptide, glucagon, gastrin and bombesin (all from Dako Corp., Santa Barbara, CA). Tissue sections (4 Km thick) were sequentially incubated with the appropriate primary antibodies, swine anti-rabbit immunoglobulin G (IgG) or rabbit anti-mouse IgG (l:lOO-1:400), followed by rabbit or mouse peroxidase-antiperoxidase complexes (1:lOO) (Dako Corp.), with frequent washes in phosphate-buffered saline between incubations. The peroxidase reaction product was visualized by incubating with diaminobenzidine (0.05%) and hydrogen peroxide (0.01%). Controls included omission of primary antibodies and substitution of primary antibodies with nonimmune sera from the same species.
Gastrin BindingAssay Tumors were quickly removed and washed with cold buffer A (Tris, 10 mmol/L; KCI, 2 mmol/L; MgCl,, 2.5 mmol/L; and sucrose, 0.25 mol/L; pH 7.4) containing BSB [I% bovine serum albumin, fraction V (Sigma Chemical Co., St. Louis, MO)], 0.1% soybean trypsin inhibitor (Worthington Biochemical Corp., Freehold, NJ), and 0.1% bacitracin (Sigma Chemical Co.). Tumors were then stored at - 70°C in an ultradeep freeze until further analysis. Specific binding sites for gastrin were measured on cell membranes prepared from tumor samples by our previously published methods (21,22).
HUMAN CARCINOID IN NUDE MICE
August 1991
Somatostatin ReceptorAssay Somatostatin receptors were measured on BON tumors by autoradiography on tissue sections (10 km thick) as described before in detail for various tumors, including hormone-producing gastroenteropancreatic tumors (23). The iodinated Tyr3 analogue of SMS 201-995 (code-named 204-090) was used as the radioligand.
Administration
of SMS 201-995 and DFMO
Initially, to establish tumors, dispersed BON cells (passage 8, 1 x 10’ cells) were inoculated SC in nude mice. When tumors became approximately 10 cm’ in area, the mice were killed and tumors minced into s-mm’ pieces that were then implanted bilaterally into the flanks of 20 nude mice. The mice were randomly allocated to receive either saline (0.1 mL, IP, t.i.d.), 2% (wt/vol) DFMO (a gift from W. J. Hudak, Ph.D., Manager of Research Information at the Merrell Research Center, Cincinnati, OH) in drinking water, SMS 201-995 (300 p&kg, IP, t.i.d.), or a combination of DFMO and SMS 201-995 beginning the day of tumor implantation and continuing until they were killed. Water bottles were covered to prevent light degradation. Drinking water was renewed every 2 days. SMS 201-995 (a gift of Sandoz Research Institute, Hanover, NJ) was diluted to the required concentration with saline. Mice were weighed weekly, and tumors were measured twice weekly by the same observer. The surface areas of the tumors were calculated as described above. Mice were killed on day 33, and tumors were removed, weighed, and frozen at -70°C until assayed for DNA, RNA, and protein content.
305
dependent fashion (Figure 1).After 2 x 107BON cells were injected, tumors were visible by 7 days in all mice. When 1 x 10' cells were injected, tumors grew in all mice: however, they were not visible until day 10. After injection of 5 X 10" cells,80% of tumors grew and were visible by day 17. Once the tumors began to grow, the growth rates were similar regardless of initial inoculum, and tumor-doubling time was approximately 13 days. All tumors grew as discrete encapsulated masses, without evidence of local invasion or distant metastasis. Tumor xenografts were sent to the American Type Culture Collection (Rockville, MD) where isoenzyme determination was performed by electrophoresis, confirming that this was human tissue and not a spontaneously occurring mouse tumor.
Morphological
Studies
Light microscopy. The original patient’s tumor was composed of masses of ovoid cells arranged in a nonorganoid fashion in a delicate, vascular stroma. Cellular and nuclear pleomorphism was minimal (Figure 2A). Tumors from the nude mouse exhibited a similar histopathologic pattern with moderate nuclear pleomorphism of the cells. (Figure 2B).
200 -
BON (2 X 10’)
* 100:
Protein, DNA, and RNAAnalysis Tumors were thawed, homogenized (Polytron; kinematica GmbH, Kriens-Luzern, Switzerland], and extracted by the method of Ogur and Rosen (24). Protein content was determined by the method of Lowry et al. (25), with bovine serum albumin as standard. DNA content was measured by the Burton (26) modification of the diphenylamine procedure with calf thymus DNA used as the standard. RNA content was measured by means of the orcinol procedure with yeast RNA as the standard (27).
Statistics Results are expressed as the mean * SEM. Antiproliferative effects of SMS 201-995 and DFMO were analyzed by two-way analysis of variance. A value of P < 0.05was considered significant.
80 60 T sw
“4”030 20 -
Tumor Growth in Nude Mice
When single-cell suspensions were injected SC into nude mice, tumors were produced in a dose-
106)
0
!Y v) 8 5 Z
1;: t4321
Results
10’)
I 0
1
I
I
I
I
10
1
20
30
40
50
60
DAYS Figure 1. Logarithmic growth curves of BON tumors injected SC as single-cell suspensions (5 x lo’, 1 x lo’, 2 x 10’ cells) in the dermis of nude mice (n = 10 mice/group).
306
GASTROENTEROLOGY
EVERS ET AL.
Vol.
101,No.2
Gas&in Binding
Specific binding sites for gastrin in the range of lo-18 fmol/mg protein with a high binding affinity (K, = N 0.134 nmol/L) were measured from a Scatch-
ard plot of the specific binding data on tumor membranes. The binding sites were specific for binding of gastrin and did not show any significant binding affinity for cholecystokinin or other unrelated peptides tested (bombesin, vasoactive intestinal peptide, or insulin) (data not shown).
Somatostatin Binding
Somatostatin receptors were localized on tumor tissue exclusively; however, nonhomogenous distribution can be observed, with tumor regions having higher receptor density than nontumor area. Binding of ““I-[Ty?]-SMS 201-995 was shown to be displaced in the nanomolar range using increasing concentrations of SMS 201-995 in successive tissue sections (Figure 6).The binding was specific because unrelated peptides (e.g., luteinizing hormone-releasing hormone) did not compete with the ligand.
Effect of a-Difluoromethylornithine 202-995 on BON Growth
Figure 2. Light micrograph of the patient’s tumor [A)and the nude mouse tumor (B) demonstrating similar histopathologic characteristics (H&E; original magnification x 500).
and SMS
There were no significant differences in final body weight in treated mice compared with the control group, and mice receiving SMS 201-995 and DFMO showed no ill effects. Food and water intake was monitored, and there were no differences in
Electron microscopy. BON tumors from nude mice were composed of polygonal cells arranged in masses in a vascular stroma. A majority of cells exhibited an abundance of cytoplasmic secretory granules that were membrane-bound, several mitochondria, and profiles of endoplasmic reticulum. Sheaves of tonofilaments (cytokeratin) were also found in several cells, the features being characteristic of epithelial and endocrine differentiation (Figure 3). Immunohistochemical
Studies
BON xenografts stain positive for serotonin (Figure 4) and chromogranin A granules (Figure 5). The staining was appreciable in approximately 50%60% of the BON tumor cell population exhibiting both serotonin and chromogranin A immunoreactivity. Substance P, pancreatic polypeptide, vasoactive intestinal polypeptide, glucagon, gastrin, and bombesin were not detected in the BON tumor by immunohistochemical staining.
Figure 3. Electron micrograph of BON tumor demonstrating abundant dense secretory granules interspersed with sheaves of cytokeratin (arrows) (original magnification x 7100).
HUMAN CARCINOID IN NUDE MICE
August 1991
307
agent alone. Tumor weight of the combination-treated group was 23% protein content was 18% and DNA and RNA contents were 28% of the values in the control group. Discussion We have successfully established a line of human carcinoid tumor cells (BON) in athymic nude mice. When dispersed cells are injected SC in the mice, encapsulated tumors are produced that are histologically similar to the original patient tumor. BON has a stable tumor-doubling time, it stains positive for 5-HT and chromogranin A, and it has retained high-affinity gastrin receptors on the cell membrane. In addition, we have shown somatostatin receptors by autoradiography. These characteristics make BON a useful model in which to study various treatment regimens on growth of carcinoid tumors and is, to our knowledge, the first long-term human carcinoid xenograft cell line to be established in nude mice. Also, BON cells grown in tissue culture exhibit
Figure 4. Light micrograph nin (arrows) (counterstained cation X 1000).
of the BON tumor stained for semtowith hematoxylin; original magnifi-
consumption between groups. Two mice (one control and one SMS 201-995-treated) died during the course of the experiment because of traumatic injections (confirmed by necropsy) and were excluded from analysis. Both SMS 201-995 (300 t~&g, IP, t.i.d.) and 2% DFMO, administered either as single agents or in combination, significantly inhibited tumor area by day 14; inhibition continued to the time of killing (day 33) (Figure i’]. Tumor weight, DNA, RNA, and protein contents were similarly inhibited after treatment (Figure 8). The mean tumor weight of the DFMO-treated group was 42% of that of the control group, protein content was 40%, and DNA and RNA contents were 53% of controls. Mean tumor weight in the SMS 201-995-treated group was 38%, protein content was 39%, DNA content was 34%, and RNA content was 44% of the values in the control group. Although mean values of tumor area, weight, and biochemical determinations were the lowest in mice treated with combination therapy, these values were not significantly different from those in mice treated with either
Figure 5. Light micrograph of the BON tumor stained for chromogranin A (arrows) (counterstained with hematoxylin; original magnification x 1000).
308 EVERS ET AL.
GASTROENTEROLOGY
Vol. 101, No. 2
models of human carcinoid tumors involve transplantation of tumor chunks into the anterior chamber of the eye of rats immunosuppressed with cyclosporine (35,36). Using this model, tumors could be maintained only for short periods (up to 5 weeks], thus preventing any definitive assessment of growth characteristics or of effects of anticancer agents on tumor regression. These human carcinoid xenograft tumors in nude mice will provide a useful model to study various aspects of endocrine tumor biology and will, we hope, allow an accurate assessment of effects of various chemotherapeutic regimens. However, we realize that we are limited to a single cell line, and agents that inhibit growth of BON may not suppress the growth of other carcinoid tumors. Ideally, additional cell lines from other patients with carcinoid tumors should be established and various treatment regimens evaluated using these different tumor lines. Recent studies have shown that exogenous administration of pentagastrin may be useful as a provocative diagnostic test for patients with carcinoid tumors (7,37). In this study, we found that BON cells possess high-affinity gastrin receptors. We have not measured 5-HT release in vivo following pentagastrin administration to mice bearing BON tumors. However, preliminary results from our laboratory have shown a dosedependent release of 5-HT from BON cells in vitro following addition of either pentagastrin or gastrin-17 (38), suggesting a specific receptor-mediated effect of gastrin on 5-HT release from human carcinoid cells. Studies using established in vivo models would be useful to further examine receptor-mediated amine release from human carcinoid tumors. We found that chronic IP administration of SMS inhibited 201-995 (300 pg/kg, t.i.d.) significantly
Figure 6. Somatostatin receptors in BON: H&E-stained tumor section (A),autoradiogram showing total binding of “‘I-[Tyr3]SMS 201-995 (B), and autoradiogram showing nonspecific binding (in presence of lo-’ moI/‘L unlabeled SMS 201-995) (C) (bar = 1 mm).
similar phenotypic changes as noted by the presence of secretory granules by electron microscopy. In tissue culture, BON cells secrete 5-HT (28), pancreastatin (29), and chromogranin A (30). Other investigators have reported establishment of human carcinoid tumor cells in vitro (31-33) and the successful transplantation, in vivo, of a gastric carcinoid originating from African rodents (hlastomys natalensis) (34). Although valuable in studying production and release of various amines, these models contribute little in clarifying aspects of in vivo growth and behavior of human carcinoids. Previous in vivo
BON
28
DAYS Figure 7. BON tumor area (mm”) in relation to time from implantation comparing 2% DFMO treatment (single-lratched bars; n = 10 tumors), SMS 201-995 (300 &kg, IP, t.i.d.; doublehatched bars; n = 6 tumors), and the combination of SMS 201995 + 2% DFMO (closed bars; n = 10 tumors) to control group (open bars; n = 6 tumors). *P < 0.05, two-way analysis of variance.
HUMAN CARCINOID IN NUDE MICE
August 1991
Figure 8. Tumor weight (g) and protein, DNA, and RNA content (mg) of BON tumors in control group (open bars; n = 81, DFMO group (single-hatched bars; n = lo), SMS 201-995 group (doublehatched bars; n = 8), combination of SMS 201-995 and DFMO (closed bars; n = 10) at sacrifice. *P < 0.05 by two-way analysis of variance.
1.0
309
s1
s E
-
0.5
0.0I
growth of BON beginning on day 14 of treatment and continuing until the mice were killed. The exact mechanism for this antitumor effect is not known. In this study, we have found that BON possesses somatostatin receptors, and one possibility may be a direct antiproliferative effect that is mediated through these specific receptors. Another possibility is that the effect of somatostatin may be indirect, by means of inhibiting release of gastrointestinal hormones, growth hormones, or other growth factors (i.e., epidermal growth factor, somatomedin C) known to stimulate growth of certain tumors (39). A combination of effects could also be possible. Our study clearly shows an antiproliferative effect of somatostatin on experimental carcinoid tumor growth in nude mice and is consistent with others that have shown inhibition of tumor growth in experimental models of nonendocrine solid cancers (4044) and endocrine tumors (45,46); however, the clinical effect of somatostatin in patients with carcinoid syndrome has been more controversial. Reports on the antiproliferative effect of somatostatin analogues have usually been based on a small number of patients with advanced carcinoid tumors and liver metastases who were initially started on somatostatin treatment for symptomatic control. The doses of somatostatin and the criteria for treatment response vary, and the documentation of response is usually based on interpretation of serial computed tomography or liver scans. Stockmann et al. (47) reported no effect on the size of liver metastases in five patients with carcinoid tumors who were treated long-term with SMS 201-995 (150 &d). Kvols et al. (15) reported on their experience with SMS 201-995 (450 kg/d) in the treatment of 25 patients with carcinoid
syndrome. Symptoms were relieved in 22 of the patients; however, only 3 of 13 evaluable patients had apparent improvement of perfusion defects noted on serial liver scans. Interpretation of the antiproliferative effect of somatostatin in some reports is further complicated by multiple therapeutic interventions, including various chemotherapeutic agents and embolization of the hepatic artery (48,49). A derivative of pentanoic acid, DFMO, is an irreversible inhibitor of ornithine decarboxylase, which in turn catalyzes the first and rate-limiting step in polyamine synthesis (50). Polyamines play an important role in cell growth and differentiation in rapidly dividing tissues and depletion of intracellular polyamines by various agents can slow and eventually stop cell growth (51). Studies from our laboratory have shown that DFMO inhibits the growth of some gut tumors, including a hamster pancreatic adenocarcinema (H2T) (52), mouse (MC-26) and human colon cancers (53,54), and humangastric cancers (55,56). In our present study, DFMO, given as a single agent in drinking water, reduced tumor area by 42% and likewise inhibited tumor weight and protein, DNA, and RNA content after 33 days of treatment, suggesting that polyamine synthesis is important for carcinoid tumor growth. We have previously shown an additive inhibitory effect on human colon cancer xenografts by combining DFMO with SMS 201-995 (54). In our present study, the combination of DFMO and SMS 201-995 produced a further decrease in mean values of BON growth, but these values were not significantly different from those obtained with either treatment alone. However, our results suggest that DFMO or SMS 201-995 may be useful in treating patients with
310
EVERS ET AL.
carcinoid tumors and represent relatively nontoxic, novel approaches to traditional cancer chemotherapy worthy of further experimentation to determine optimal dose and duration of treatment. In conclusion, we have established a transplantable human carcinoid cell line (BON) in nude mice. BON cells contain 5-HT and chromogranin A, have a predictable pattern of growth following tumor inoculation, and are histologically similar to the original tumor from the patient. SMS 201-995 and DFMO, given as single or combination therapy, inhibited growth of BON in nude mice. The tumor line, BON, will be a useful experimental model to improve our understanding of the biological aspects of human carcinoids and to study effects of various therapeutic strategies on growth of carcinoid tumors. References 1. Jager RM, Polk HC Jr. Carcinoid apudomas. Curr Prob Cancer 1977;1:1-53. 2. Creutzfeldt W, Stockmann F. Carcinoids and carcinoid syndrome. Am J Med 1987;82:4-16. 3. Oberndorfer S. Karzinoide tumoren des dunndarms. Frankf Z Path01 1907;1:426-429. 4. Moertel CG. Treatment of the carcinoid tumor and the malignant carcinoid syndrome. J Clin Oncol1963;1:727-740. 5. Harris AL, Smith IE. Regression of carcinoid tumour with cyproheptadine. Br Med J 1982;285:475. 6. Gustafsen J, Lendorf A, Raskov H, Boesby S. Ketanserin versus placebo in carcinoid syndrome. A clinical controlled trial. Stand J Gastroentero11986;21:816-ala. 7. Ahlman H, Dahlstrom A, Gronstad K, Tisell LE, Oberg K, Zinner MJ, Jaffe BM. The pentagastrin test in the diagnosis of the carcinoid syndrome. Blockade of gastrointestinal symptoms by ketanserin. Ann Surg 1985;201:81-86. a. Richter G, Stockmann F, Lembcke B, Conlon JM, Creutzfeldt W. Short-term administration of the somatostatin analogue SMS 201-995 in patients with carcinoid turnours. Stand J Gastroenterol1986;21:193-198. 9. Anderson JV, Bloom SR. Neuroendocrine tumours of the gut: long-term therapy with the somatostatin analogue SMS ZOl995. Stand J Gastroentero11986;21:115-128. 10. Tsai ST, Lewis E, Vinik A. The use of a somatostatin analogue (SMS 201-995) in the management of the flushing syndrome. Stand J Gastroenterol1966;21:267-274. 11. Oberg K, Norhei I, Lundqvist G, Wide L. Treatment of the carcinoid syndrome with SMS 201-995, a somatostatin analogue. Stand J Gastroenterol1986;21:191-192. 12. Deghenghi R. Somatostatin analogues in the treatment of the carcinoid syndrome. Biomed Pharmacother 19aa;42:585-588. 13. Roy RC, Carter RF, Wright PD. Somatostatin, anaesthesia, and the carcinoid syndrome. Peri-operative administration of a somatostatin analogue to suppress carcinoid tumour activity. Anaesthesia 1967;42:627-632. 14. Van Houten AA, Nortier JWR, Vendrik CPJ. Successful symptomatic treatment of malignant carcinoid syndrome with the somatostatin analogue SMS 201-995. Neth J Med 1966;32:194198. 15. Kvols LK, Moertel CG, O’Connell MJ, Schutt AJ, Rubin J, Hahn RG. Treatment of the malignant carcinoid syndrome. Evaluation of a long-acting somatostatin analogue. N Engl J Med 1966;315:663-666.
GASTROENTEROLOGY
Vol. 101, No. 2
16. Vinik A, Moattari AR. Use of somatostatin analog in management of carcinoid syndrome. Dig Dis Sci 1989;34:145-27s. 17. Engstrom PF, Lavin PT, Moertel CG, Folsch E, Douglass HO Jr. Streptozocin plus fluorouracil versus doxorubicin therapy for metastatic carcinoid tumor. J Clin Oncoli984;2:1255-1259. ia. Moertel CG, Hanley JA. Combination chemotherapy trials in metastatic carcinoid tumor and the malignant carcinoid syndrome. Cancer Clin Trials 1979;2:327-334. 19. Kvols LK, Buck M. Chemotherapy of metastatic carcinoid and islet cell tumors. A review. Am J Med 1987;62:77-83. 20. Committee on Care and Use of the “Nude” Mouse. Guide for the care and use of the nude (thymus-deficient) mouse in biomedical research. BAR News 1976;19:Ml-MZO. 21. Singh P, Rae-Venter B, Townsend CM Jr, Khalil T, Thompson
22.
23.
24.
25.
26.
27. 28.
29.
30.
31.
32.
33.
34.
35.
JC. Gastrin receptors in normal and malignant gastro-intestinal mucosa: age-associated changes. Am J Physiol1965;249:G761G769. Singh P, Le S, Beauchamp RD, Townsend CM Jr, Thompson JC. Inhibition of pentagastrin-stimulated up-regulation of gastrin receptors and growth of mouse colon tumor in vivo by proglumide, a gastrin receptor antagonist. Cancer Res 1987;47:5ooo5004. Reubi JC, Haecki WH, Lamberts SWJ. Hormone-producing gastrointestinal tumors contain a high density of somatostatin receptors. J Clin Endocrinol Metab 1967;65:1127-1134. Ogur M, Rosen G. The nucleic acids of plant tissues. I. The extraction and estimation of desoxypentose nucleic acid and pentose nucleic acid. Arch Biochem 1950;25:262-276. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem 1951;193:265-275. Burton K. A study of the conditions and mechanisms of the diphenylamine reaction for the calorimetric estimation of deoxyribonucleic acid. Biochemistry 1956;62:315-323. Dische Z. Qualitative and quantitative calorimetric determination of heptoses. J Biol Chem 1953;204:963-997. Ishizuka J, Beauchamp RD, Greeley GH Jr, Townsend CM Jr, Thompson JC. 5-Hydroxytryptamine is an autocrine factor for carcinoid cells (abstr). Gastroenterology 1990;9a:A287. Jeng Y-J, Townsend CM Jr, Nagasawa S, Chuo S, Kern K, Yanaihara N, Ferrar RS, Thompson JC, Greeley GH Jr, Regulation of pancreastatin release from a human pancreatic carcinoid cell line in vitro. Endocrinology 1991;12a:220-225. Chuo S, Townsend CM Jr, Farrar S, Kern K, Yanaihara N, Thompson JC, Greeley GH Jr. Coordinate pancreastatin and chromogranin A secretion from a human pancreatic carcinoid: evidence of chromogranin processing to pancreastatin (abstr). Gastroenterology 1990;96:A653. Norheim I, Wilander E, Oberg K, Theodorsson-Norheim E, Lundqvist M-L, Lindgren P, Bergh J. Tachykinin production by carcinoid tumours in culture. Eur J Cancer Clin Oncol1967;23: 669-695. Ahlman H, Ahlund L, Nilsson 0, Skolnik G, Theodorsson E, Dahlstrom A. Carcinoid tumour cells in long-term culture: release of serotonin but not of tachykinins on stimulation with adrenoceptor agonists. Int J Cancer 1988;42:506-510. Debons-Guillemin M-C, Launay J-M, Roseto A, Peries J. Serotonin and histamine production by human carcinoid cells in culture. Cancer Res 1962;42:1513-1516. Hosoda S, Suzuki K, Sudo K, Yoshida N, Tanaka C. Transplantable argyrophilic gastric carcinoid of mastomys natalensis secreting both histamine and serotonin. J Nat1 Cancer Inst 1979;63:1447-1449. Nilsson 0, Ahlman H, Ericson LE, Skolnik G, Dahlstrom A. Release of serotonin from human carcinoid tumor cells in vitro and grown in the anterior eye chamber of the rat. Cancer 1986;58:676-684.
August
1991
36. Nilsson 0, Dahlstrom A, Gronstad K-O, Rosengren L, Briving C, Skolnik G, Ahlman H. Successful transplantation of a human midgut carcinoid tumour to the anterior eye chamber of the rat. Acta Physiol Stand 1984;120:317-319. 37. Gronstad K-O, Nilsson 0, Hedman I, Skolnik G, Dahlstrom A, Ahlman H. On the mode of action of the pentagastrin test in the carcinoid syndrome. Stand J Gastroenterol 1985;20:508-511. 38. Lawrence JP, Haber B, Ishizuka J, Thompson JC. The effect of various gastrin forms on secretion on a carcinoid tumor in vitro (abstr). Gastroenterology 1989;96:A291. 39. Evers BM, Parekh D, Townsend CM Jr, Thompson JC. Somatostatin and analogues in the treatment of cancer. A review. Ann Surg 1991;213:190-198. 40. Setyono-Han B, Henkelman MS, Foekens JA, Klijn JGM. Direct inhibitory effects of somatostatin (analogues) on the growth of human breast cancer cells. Cancer Res 1987;47:1566-1570. 41. Weber C, Merriam L, Koschitzky T, Karp F, Benson M, Forde K, LoGerfo P. Inhibition of growth of human breast carcinomas in viva by somatostatin analogue, SMS #20X-995, treatment of nude mouse xenografts. Surgery 1989;106:416-422. 42. Scambia G, Panici PB, Baiocchi G, Perrone L, Iacobelli S, Mancuso S. Antiproliferative effects of somatostatin and the somatostatin analog SMS 201-995 on three human breast cancer cell lines. J Cancer Res Clin Oncol1987;114:306-308. 43. Schally AV, Redding TW. Somatostatin analogs as adjuncts to agonists of luteinizing hormone-releasing hormone in the treatment of experimental prostate cancer. Proc Nat1 Acad Sci USA 1987;84:7275-7279. 44. Upp JR Jr, Olson D, Poston GJ, Alexander RW. Townsend CM Jr, Thompson JC. Inhibition of growth of two human pancreatic in vim by somatostatin analog SMS 201-995. adenocarcinomas Am J Surg 1988;155:29-35. 45. Milhoan RA. Trudel JL, Lawrence JP, Townsend CM Jr, Thompson JC. Somatostatin inhibits growth of human small cell lung carcinoma in vim. Surg Forum 1988;39:438-439. 46. Reubi JC. A somatostatin analogue inhibits chondrosarcoma and insulinoma tumourgrowth. ActaEndocrinol1985;109:108114. 47. Stockmann R, Richter G, Lembcke B, Conlon JM, Creutzfeldt W. Long-term treatment of patients with endocrine gastrointestinal tumours with the somatostatin analogue SMS 201-995. Stand J Gastroenterol 1986;21:230-237. 48. Souquet J-C, Sassolas G, Forichon J, Champetier P, Partensky C, Chayvialle J-A. Clinical and hormonal effects of a long-
HUMAN
CARCINOID
IN NUDE MICE
311
acting somatostatin analogue in pancreatic endocrine tumors and in carcinoid syndrome. Cancer 1987;59:1654-1660. 49. Ahlman H, Ahlund L, Dahlstrom A, Nilsson 0, Skolnik, Tisell L-E, Tylen U. Use of a somatostatin analogue in association with surgery and hepatic arterial embolisation in the treatment of the carcinoid syndrome. Br J Cancer 1987;56:840-842, 50. Sjoerdsma A, Schechter PJ. Chemotherapeutic implications of polyamine biosynthesis inhibition. Clin Pharmacol Ther 1984; 35:287-300. 51. Heby 0, Janne J. Polyamine antimetabolites: biochemistry, specificity and biological effects of inhibitors of polyamine synthesis. In: Morris DR, Marton JL, eds. Polyamines in biology and medicine. New York: Marcel Dekker, 1981:243-310. 52. Saydjari R, Townsend CM Jr, Barranco SC, James E, Thompson JC. Effects of cyclosporin A and a-difluoromethylornithine on the growth of hamster pancreatic cancer in vitro. J Nat1 Cancer Inst 1986;77:1087-1092. 53. Marx M, Glass EJ, Townsend CM Jr, Barranco SC, Thompson JC. Effects of o-difluoromethylornithine (DFMO) on mouse colon cancer in vitro and in viva (abstr) Gastroenterology 1983;84:A1242. 54. Milhoan RA, Milhoan LH, Trudel JL, Lawrence JP, Townsend CM Jr, Thompson JC. Somatostatin and a-difluoromethylornithine (DFMO) have additive inhibitory effects on growth of human colon carcinoma (HCC) in viva. Biomed Res 1988;9:A30. 55. Barranco SC, Ford PJ, Townsend CM Jr. Heterogeneous survival responses of human gastric cancer clones to a-difluoromethyl-ornithine in vitro. Invest New Drugs 1986;4:337-345. 56. Upp JR Jr, Beauchamp RD, Townsend CM Jr, Barranco SC, Singh P, Rajaraman S, James E, Thompson JC. Inhibition of human gastric adenocarcinoma xenograft growth in nude mice by oc-difluoromethylornithine. Cancer Res 1988;48:3265-3269.
Received August 17,199O. Accepted December 7, 1990. Address requests for reprints to: Courtney M. Townsend, Jr., M.D., Department of Surgery, The University of Texas Medical Branch, Galveston, Texas 77550. Dr. Kim is a visiting scientist from the Department of Surgery, College of Medicine, Seoul National University, Seoul, Korea. Supported by grants from the National Institutes of Health (5R37 DK 15241, PO1 DK 35608, CA 38561), from the American Cancer Society [PDT 220) and in conjunction with the Walls Medical Research Foundation.
1991;101:303-311
Establishment and Characterization of a Human Carcinoid in Nude Mice and Effect of Various Agents on Tumor Growth B. MARK EVERS, COURTNEY M. TOWNSEND, Jr., J. ROBERT ERICK ALLEN, STEPHEN C. HURLBUT, SUN WHE KIM, SRINIVASAN RAJARAMAN, POMILA SINGH, JEAN CLAUDE REUBI, and JAMES C. THOMPSON
UPP,
Departments and Sandoz
Texas:
of Surgery and Pathology, Research Institute, Berne,
University Switzerland
The authors have established a long-term tissue culture cell line (BON) derived from a metastatic human carcinoid tumor of the pancreas. The cells have been in continuous passage for 46 months. Tissue culture cells produce tumors in a dosedependent fashion after SC inoculation of cell suspensions in athymic nude mice. BON tumors, grown in nude mice, are histologically identical to the original tumor; they possess gastrin and somatostatin receptors, synthesize serotonin and chromog ranin A, and have a doubling time of approximately 13 days. The antiproliferative effects of the longacting somatostatin analogue, SMS 201-995 (300 @kg, t.i.d.), and 2% ac-difluoromethylornithine on BON xenografts in nude mice were examined. Tumor size was significantly decreased by day 14 of treatment with either agent and at all points of analysis thereafter until the animals were killed (day 33). In addition, tumor weight, DNA, RNA, and protein contents were significantly decreased in treated mice compared with controls. Establishment of this human carcinoid xenograft line, BON, provides an excellent model to study further the biological behavior of carcinoid tumors and the in vivo effect of chemotherapeutic agents on tumor growth.
Cfrom neuroectodermal (1,2). Primarily found in
arcinoid tumors are endocrine neoplasms derived cells of the neural crest the gastrointestinal (GI) tract, these tumors are uncommon, but not rare, with a reported incidence in autopsy series ranging from 0.1% to 1.4% (1). Originally called “Karzinoide” by Oberndorfer (3) to describe its apparent benign na-
of Texas
Medical
Branch,
Galveston,
ture, modern immunohistochemical techniques have allowed identification and isolation of the vasoactive substance serotonin [ti-hydroxytryptamine (5-HT)] and a host of biogenic amines and hormones produced by carcinoid tumors (1,2). Treatment of patients with carcinoid tumors has been hampered by the lack of an animal model. In the majority of patients, treatment of metastatic carcinoid tumors has been directed largely toward relief of the debilitating sequelae of the carcinoid syndrome by means of blocking agents [for example, methysergide (4), cyproheptadine (4,5), and ketanserin (6,7)] or the inhibitory hormone, somatostatin (8-16). Antiproliferative agents that are commonly used to slow tumor growth include combinations of streptozotocin, 5-fluorouracil, doxorubicin, or cyclophosphamide (5,1719). However, most carcinoid tumors respond poorly to cytotoxic therapy with these agents (1,4,18-19). In this study, we report the first in vivo establishment of a long-term xenograft cell line of a human pancreatic carcinoid tumor. We have characterized tumor growth rate in vivo, the status of gastrin and somatostatin receptors, the morphology, and the production of peptides and amines. In addition, we have examined possible antiproliferative effects of the longacting somatostatin analogue, SMS 201-995, and ol-difluoromethylornithine (DFMO) (either alone or in combination) on growth of the carcinoid tumor. Abbreviations usedin thispaper: DFMO, a-difluoromethylornithine; DMEM, Dulbecco’s modified Eagle medium: FCS, fetal calf serum; S-I-IT,5-hydroxytriptamine. o 1991 by the American Gastroenterological Association 9036-5085/91/$3.09
304
GASTROENTEROLOGY
EVERS ET AL.
Materials
and Methods
Surgical Specimen Collection and Maintenance of Tumor Line The operative specimen of a peripancreatic lymph node was sterilely obtained in April 1986 from a 28-year-old man with metastatic carcinoid tumor of the pancreas who had come into the hospital with symptoms of obstructive jaundice and diarrhea. Before exploratory laparotomy the patient had not received any interventional treatment. Final diagnosis of a metastatic carcinoid tumor was made by histological examination of the lymph node and positive staining for neuron-specific enolase and 5-HT. A portion of the lymph node was washed in saline, minced, and tumor fragments placed in Dulbecco’s modified Eagle medium (DMEM; Gibco, Grand Island, NY) and F12K (Gibco) in a 1:l ratio supplemented with,lO% (vol/vol) fetal calf serum (FCS; Hyclone Laboratories, Logan, UT) and 1% gentamicin. The cells were grown at 37°C in an atmosphere of 95% air and 5% CO,. Cells were routinely passed by removing the medium and overlaying the cell monolayer with 0.25% trypsin:O.l% ethylenediaminetetraacetic acid (EDTA). The elimination of fibroblasts from the stroma of the tumor tissue was accomplished by brief exposures to 0.06% trypsin:0.02% EDTA. This procedure was repeated until no further fibroblast growth was observed. Tumor cells from passage 5 were frozen after successful in vitro adaptation and removal of all fibroblasts. The cell line is presently maintained in DMEM and F12K growth medium supplemented with 10% FCS; it is passed at a 1:2 ratio when cells reach 80% confluence. Cell cultures are routinely monitored for mycoplasma contamination, and no mycoplasma growth has been detected.
Animals Male athymic nude mice (Balb/c, 20-25 g, 3-4 weeks of age; Life Science, St. Petersburg, FL) were housed under specific pathogen-free conditions in a temperaturecontrolled isolation unit with IX-hour light-dark cycles in accordance with the National Research Council’s Guide for the Care and Use of the Nude Mouse in Biomedical Research (20). The mice were fed a standard chow (Autoclavable Rodent Chow no. 5010; Ralston Purina, St. Louis, MO) and sterile water, both given ad libitum.
Solid Tumors in Nude Mice and In Vitro Growth Rate BON cells (passage 8) were harvested from subconfluent cultures by a l-minute treatment with 0.25% trypsin and 0.1% EDTA. Single-cell suspensions (5 X 106, 1 X lo’, and 2 x 10’) in RPM1 1640 without serum (total volume, 0.1 mL) were injected SC at a single site in the dermis of nude mice (n = 10 mice/group). Tumor size (longest perpendicular diameter) was measured biweekly with Vernier calipers (Mitutoyo Corp., Tokyo, Japan) accurate to 0.5 mm. Surface areas were calculated as the product of the two greatest perpendicular tumor diameters and were expressed as
Vol. 101, No. 2
square millimeters. Tumor areas were plotted on semilogarithmic paper, and the tumor doubling time (defined as time in days required for mean tumor area to double during logarithmic growth) was determined directly from the graph. On day 48, mice were killed and tumors from passage 8 were removed for light and electron microscopy, immunohistochemistry, and measurement of gastrin receptors.
Light and Electron Microscopy For light microscopy, tumor tissue blocks measuring 1 x 1 x 0.5 cm were fixed in 10% neutral buffered formalin for 6-8 hours, processed routinely, and embedded in paraffin Sections (4 pm thick) were stained with H&E and examined. For transmission electron microscopy, tissue blocks (1 mm3) were fixed in half-strength Karnovsky’s fixative for 4 to 6 hours, postfixed in osmium tetroxide (l%), and embedded in Epon. Sections (70 nm thick) were stained with uranyl acetate and Reynold’s lead citrate and examined with a Philips 410 electron microscope (Philips Medical Systems Inc., Shelton, CT).
Zmmunohistochemical
Studies
Formalin-fixed, paraffin-embedded tissue samples were studied by means of a three-layer immunohistochemical method. The primary antibodies used were murine monoclonal antibody to chromogranin A (Boehringer Mannheim, Indianapolis, IN) and monospecific polyclonal antibodies to serotonin, substance P, pancreatic polypeptide, vasoactive intestinal peptide, glucagon, gastrin and bombesin (all from Dako Corp., Santa Barbara, CA). Tissue sections (4 Km thick) were sequentially incubated with the appropriate primary antibodies, swine anti-rabbit immunoglobulin G (IgG) or rabbit anti-mouse IgG (l:lOO-1:400), followed by rabbit or mouse peroxidase-antiperoxidase complexes (1:lOO) (Dako Corp.), with frequent washes in phosphate-buffered saline between incubations. The peroxidase reaction product was visualized by incubating with diaminobenzidine (0.05%) and hydrogen peroxide (0.01%). Controls included omission of primary antibodies and substitution of primary antibodies with nonimmune sera from the same species.
Gastrin BindingAssay Tumors were quickly removed and washed with cold buffer A (Tris, 10 mmol/L; KCI, 2 mmol/L; MgCl,, 2.5 mmol/L; and sucrose, 0.25 mol/L; pH 7.4) containing BSB [I% bovine serum albumin, fraction V (Sigma Chemical Co., St. Louis, MO)], 0.1% soybean trypsin inhibitor (Worthington Biochemical Corp., Freehold, NJ), and 0.1% bacitracin (Sigma Chemical Co.). Tumors were then stored at - 70°C in an ultradeep freeze until further analysis. Specific binding sites for gastrin were measured on cell membranes prepared from tumor samples by our previously published methods (21,22).
HUMAN CARCINOID IN NUDE MICE
August 1991
Somatostatin ReceptorAssay Somatostatin receptors were measured on BON tumors by autoradiography on tissue sections (10 km thick) as described before in detail for various tumors, including hormone-producing gastroenteropancreatic tumors (23). The iodinated Tyr3 analogue of SMS 201-995 (code-named 204-090) was used as the radioligand.
Administration
of SMS 201-995 and DFMO
Initially, to establish tumors, dispersed BON cells (passage 8, 1 x 10’ cells) were inoculated SC in nude mice. When tumors became approximately 10 cm’ in area, the mice were killed and tumors minced into s-mm’ pieces that were then implanted bilaterally into the flanks of 20 nude mice. The mice were randomly allocated to receive either saline (0.1 mL, IP, t.i.d.), 2% (wt/vol) DFMO (a gift from W. J. Hudak, Ph.D., Manager of Research Information at the Merrell Research Center, Cincinnati, OH) in drinking water, SMS 201-995 (300 p&kg, IP, t.i.d.), or a combination of DFMO and SMS 201-995 beginning the day of tumor implantation and continuing until they were killed. Water bottles were covered to prevent light degradation. Drinking water was renewed every 2 days. SMS 201-995 (a gift of Sandoz Research Institute, Hanover, NJ) was diluted to the required concentration with saline. Mice were weighed weekly, and tumors were measured twice weekly by the same observer. The surface areas of the tumors were calculated as described above. Mice were killed on day 33, and tumors were removed, weighed, and frozen at -70°C until assayed for DNA, RNA, and protein content.
305
dependent fashion (Figure 1).After 2 x 107BON cells were injected, tumors were visible by 7 days in all mice. When 1 x 10' cells were injected, tumors grew in all mice: however, they were not visible until day 10. After injection of 5 X 10" cells,80% of tumors grew and were visible by day 17. Once the tumors began to grow, the growth rates were similar regardless of initial inoculum, and tumor-doubling time was approximately 13 days. All tumors grew as discrete encapsulated masses, without evidence of local invasion or distant metastasis. Tumor xenografts were sent to the American Type Culture Collection (Rockville, MD) where isoenzyme determination was performed by electrophoresis, confirming that this was human tissue and not a spontaneously occurring mouse tumor.
Morphological
Studies
Light microscopy. The original patient’s tumor was composed of masses of ovoid cells arranged in a nonorganoid fashion in a delicate, vascular stroma. Cellular and nuclear pleomorphism was minimal (Figure 2A). Tumors from the nude mouse exhibited a similar histopathologic pattern with moderate nuclear pleomorphism of the cells. (Figure 2B).
200 -
BON (2 X 10’)
* 100:
Protein, DNA, and RNAAnalysis Tumors were thawed, homogenized (Polytron; kinematica GmbH, Kriens-Luzern, Switzerland], and extracted by the method of Ogur and Rosen (24). Protein content was determined by the method of Lowry et al. (25), with bovine serum albumin as standard. DNA content was measured by the Burton (26) modification of the diphenylamine procedure with calf thymus DNA used as the standard. RNA content was measured by means of the orcinol procedure with yeast RNA as the standard (27).
Statistics Results are expressed as the mean * SEM. Antiproliferative effects of SMS 201-995 and DFMO were analyzed by two-way analysis of variance. A value of P < 0.05was considered significant.
80 60 T sw
“4”030 20 -
Tumor Growth in Nude Mice
When single-cell suspensions were injected SC into nude mice, tumors were produced in a dose-
106)
0
!Y v) 8 5 Z
1;: t4321
Results
10’)
I 0
1
I
I
I
I
10
1
20
30
40
50
60
DAYS Figure 1. Logarithmic growth curves of BON tumors injected SC as single-cell suspensions (5 x lo’, 1 x lo’, 2 x 10’ cells) in the dermis of nude mice (n = 10 mice/group).
306
GASTROENTEROLOGY
EVERS ET AL.
Vol.
101,No.2
Gas&in Binding
Specific binding sites for gastrin in the range of lo-18 fmol/mg protein with a high binding affinity (K, = N 0.134 nmol/L) were measured from a Scatch-
ard plot of the specific binding data on tumor membranes. The binding sites were specific for binding of gastrin and did not show any significant binding affinity for cholecystokinin or other unrelated peptides tested (bombesin, vasoactive intestinal peptide, or insulin) (data not shown).
Somatostatin Binding
Somatostatin receptors were localized on tumor tissue exclusively; however, nonhomogenous distribution can be observed, with tumor regions having higher receptor density than nontumor area. Binding of ““I-[Ty?]-SMS 201-995 was shown to be displaced in the nanomolar range using increasing concentrations of SMS 201-995 in successive tissue sections (Figure 6).The binding was specific because unrelated peptides (e.g., luteinizing hormone-releasing hormone) did not compete with the ligand.
Effect of a-Difluoromethylornithine 202-995 on BON Growth
Figure 2. Light micrograph of the patient’s tumor [A)and the nude mouse tumor (B) demonstrating similar histopathologic characteristics (H&E; original magnification x 500).
and SMS
There were no significant differences in final body weight in treated mice compared with the control group, and mice receiving SMS 201-995 and DFMO showed no ill effects. Food and water intake was monitored, and there were no differences in
Electron microscopy. BON tumors from nude mice were composed of polygonal cells arranged in masses in a vascular stroma. A majority of cells exhibited an abundance of cytoplasmic secretory granules that were membrane-bound, several mitochondria, and profiles of endoplasmic reticulum. Sheaves of tonofilaments (cytokeratin) were also found in several cells, the features being characteristic of epithelial and endocrine differentiation (Figure 3). Immunohistochemical
Studies
BON xenografts stain positive for serotonin (Figure 4) and chromogranin A granules (Figure 5). The staining was appreciable in approximately 50%60% of the BON tumor cell population exhibiting both serotonin and chromogranin A immunoreactivity. Substance P, pancreatic polypeptide, vasoactive intestinal polypeptide, glucagon, gastrin, and bombesin were not detected in the BON tumor by immunohistochemical staining.
Figure 3. Electron micrograph of BON tumor demonstrating abundant dense secretory granules interspersed with sheaves of cytokeratin (arrows) (original magnification x 7100).
HUMAN CARCINOID IN NUDE MICE
August 1991
307
agent alone. Tumor weight of the combination-treated group was 23% protein content was 18% and DNA and RNA contents were 28% of the values in the control group. Discussion We have successfully established a line of human carcinoid tumor cells (BON) in athymic nude mice. When dispersed cells are injected SC in the mice, encapsulated tumors are produced that are histologically similar to the original patient tumor. BON has a stable tumor-doubling time, it stains positive for 5-HT and chromogranin A, and it has retained high-affinity gastrin receptors on the cell membrane. In addition, we have shown somatostatin receptors by autoradiography. These characteristics make BON a useful model in which to study various treatment regimens on growth of carcinoid tumors and is, to our knowledge, the first long-term human carcinoid xenograft cell line to be established in nude mice. Also, BON cells grown in tissue culture exhibit
Figure 4. Light micrograph nin (arrows) (counterstained cation X 1000).
of the BON tumor stained for semtowith hematoxylin; original magnifi-
consumption between groups. Two mice (one control and one SMS 201-995-treated) died during the course of the experiment because of traumatic injections (confirmed by necropsy) and were excluded from analysis. Both SMS 201-995 (300 t~&g, IP, t.i.d.) and 2% DFMO, administered either as single agents or in combination, significantly inhibited tumor area by day 14; inhibition continued to the time of killing (day 33) (Figure i’]. Tumor weight, DNA, RNA, and protein contents were similarly inhibited after treatment (Figure 8). The mean tumor weight of the DFMO-treated group was 42% of that of the control group, protein content was 40%, and DNA and RNA contents were 53% of controls. Mean tumor weight in the SMS 201-995-treated group was 38%, protein content was 39%, DNA content was 34%, and RNA content was 44% of the values in the control group. Although mean values of tumor area, weight, and biochemical determinations were the lowest in mice treated with combination therapy, these values were not significantly different from those in mice treated with either
Figure 5. Light micrograph of the BON tumor stained for chromogranin A (arrows) (counterstained with hematoxylin; original magnification x 1000).
308 EVERS ET AL.
GASTROENTEROLOGY
Vol. 101, No. 2
models of human carcinoid tumors involve transplantation of tumor chunks into the anterior chamber of the eye of rats immunosuppressed with cyclosporine (35,36). Using this model, tumors could be maintained only for short periods (up to 5 weeks], thus preventing any definitive assessment of growth characteristics or of effects of anticancer agents on tumor regression. These human carcinoid xenograft tumors in nude mice will provide a useful model to study various aspects of endocrine tumor biology and will, we hope, allow an accurate assessment of effects of various chemotherapeutic regimens. However, we realize that we are limited to a single cell line, and agents that inhibit growth of BON may not suppress the growth of other carcinoid tumors. Ideally, additional cell lines from other patients with carcinoid tumors should be established and various treatment regimens evaluated using these different tumor lines. Recent studies have shown that exogenous administration of pentagastrin may be useful as a provocative diagnostic test for patients with carcinoid tumors (7,37). In this study, we found that BON cells possess high-affinity gastrin receptors. We have not measured 5-HT release in vivo following pentagastrin administration to mice bearing BON tumors. However, preliminary results from our laboratory have shown a dosedependent release of 5-HT from BON cells in vitro following addition of either pentagastrin or gastrin-17 (38), suggesting a specific receptor-mediated effect of gastrin on 5-HT release from human carcinoid cells. Studies using established in vivo models would be useful to further examine receptor-mediated amine release from human carcinoid tumors. We found that chronic IP administration of SMS inhibited 201-995 (300 pg/kg, t.i.d.) significantly
Figure 6. Somatostatin receptors in BON: H&E-stained tumor section (A),autoradiogram showing total binding of “‘I-[Tyr3]SMS 201-995 (B), and autoradiogram showing nonspecific binding (in presence of lo-’ moI/‘L unlabeled SMS 201-995) (C) (bar = 1 mm).
similar phenotypic changes as noted by the presence of secretory granules by electron microscopy. In tissue culture, BON cells secrete 5-HT (28), pancreastatin (29), and chromogranin A (30). Other investigators have reported establishment of human carcinoid tumor cells in vitro (31-33) and the successful transplantation, in vivo, of a gastric carcinoid originating from African rodents (hlastomys natalensis) (34). Although valuable in studying production and release of various amines, these models contribute little in clarifying aspects of in vivo growth and behavior of human carcinoids. Previous in vivo
BON
28
DAYS Figure 7. BON tumor area (mm”) in relation to time from implantation comparing 2% DFMO treatment (single-lratched bars; n = 10 tumors), SMS 201-995 (300 &kg, IP, t.i.d.; doublehatched bars; n = 6 tumors), and the combination of SMS 201995 + 2% DFMO (closed bars; n = 10 tumors) to control group (open bars; n = 6 tumors). *P < 0.05, two-way analysis of variance.
HUMAN CARCINOID IN NUDE MICE
August 1991
Figure 8. Tumor weight (g) and protein, DNA, and RNA content (mg) of BON tumors in control group (open bars; n = 81, DFMO group (single-hatched bars; n = lo), SMS 201-995 group (doublehatched bars; n = 8), combination of SMS 201-995 and DFMO (closed bars; n = 10) at sacrifice. *P < 0.05 by two-way analysis of variance.
1.0
309
s1
s E
-
0.5
0.0I
growth of BON beginning on day 14 of treatment and continuing until the mice were killed. The exact mechanism for this antitumor effect is not known. In this study, we have found that BON possesses somatostatin receptors, and one possibility may be a direct antiproliferative effect that is mediated through these specific receptors. Another possibility is that the effect of somatostatin may be indirect, by means of inhibiting release of gastrointestinal hormones, growth hormones, or other growth factors (i.e., epidermal growth factor, somatomedin C) known to stimulate growth of certain tumors (39). A combination of effects could also be possible. Our study clearly shows an antiproliferative effect of somatostatin on experimental carcinoid tumor growth in nude mice and is consistent with others that have shown inhibition of tumor growth in experimental models of nonendocrine solid cancers (4044) and endocrine tumors (45,46); however, the clinical effect of somatostatin in patients with carcinoid syndrome has been more controversial. Reports on the antiproliferative effect of somatostatin analogues have usually been based on a small number of patients with advanced carcinoid tumors and liver metastases who were initially started on somatostatin treatment for symptomatic control. The doses of somatostatin and the criteria for treatment response vary, and the documentation of response is usually based on interpretation of serial computed tomography or liver scans. Stockmann et al. (47) reported no effect on the size of liver metastases in five patients with carcinoid tumors who were treated long-term with SMS 201-995 (150 &d). Kvols et al. (15) reported on their experience with SMS 201-995 (450 kg/d) in the treatment of 25 patients with carcinoid
syndrome. Symptoms were relieved in 22 of the patients; however, only 3 of 13 evaluable patients had apparent improvement of perfusion defects noted on serial liver scans. Interpretation of the antiproliferative effect of somatostatin in some reports is further complicated by multiple therapeutic interventions, including various chemotherapeutic agents and embolization of the hepatic artery (48,49). A derivative of pentanoic acid, DFMO, is an irreversible inhibitor of ornithine decarboxylase, which in turn catalyzes the first and rate-limiting step in polyamine synthesis (50). Polyamines play an important role in cell growth and differentiation in rapidly dividing tissues and depletion of intracellular polyamines by various agents can slow and eventually stop cell growth (51). Studies from our laboratory have shown that DFMO inhibits the growth of some gut tumors, including a hamster pancreatic adenocarcinema (H2T) (52), mouse (MC-26) and human colon cancers (53,54), and humangastric cancers (55,56). In our present study, DFMO, given as a single agent in drinking water, reduced tumor area by 42% and likewise inhibited tumor weight and protein, DNA, and RNA content after 33 days of treatment, suggesting that polyamine synthesis is important for carcinoid tumor growth. We have previously shown an additive inhibitory effect on human colon cancer xenografts by combining DFMO with SMS 201-995 (54). In our present study, the combination of DFMO and SMS 201-995 produced a further decrease in mean values of BON growth, but these values were not significantly different from those obtained with either treatment alone. However, our results suggest that DFMO or SMS 201-995 may be useful in treating patients with
310
EVERS ET AL.
carcinoid tumors and represent relatively nontoxic, novel approaches to traditional cancer chemotherapy worthy of further experimentation to determine optimal dose and duration of treatment. In conclusion, we have established a transplantable human carcinoid cell line (BON) in nude mice. BON cells contain 5-HT and chromogranin A, have a predictable pattern of growth following tumor inoculation, and are histologically similar to the original tumor from the patient. SMS 201-995 and DFMO, given as single or combination therapy, inhibited growth of BON in nude mice. The tumor line, BON, will be a useful experimental model to improve our understanding of the biological aspects of human carcinoids and to study effects of various therapeutic strategies on growth of carcinoid tumors. References 1. Jager RM, Polk HC Jr. Carcinoid apudomas. Curr Prob Cancer 1977;1:1-53. 2. Creutzfeldt W, Stockmann F. Carcinoids and carcinoid syndrome. Am J Med 1987;82:4-16. 3. Oberndorfer S. Karzinoide tumoren des dunndarms. Frankf Z Path01 1907;1:426-429. 4. Moertel CG. Treatment of the carcinoid tumor and the malignant carcinoid syndrome. J Clin Oncol1963;1:727-740. 5. Harris AL, Smith IE. Regression of carcinoid tumour with cyproheptadine. Br Med J 1982;285:475. 6. Gustafsen J, Lendorf A, Raskov H, Boesby S. Ketanserin versus placebo in carcinoid syndrome. A clinical controlled trial. Stand J Gastroentero11986;21:816-ala. 7. Ahlman H, Dahlstrom A, Gronstad K, Tisell LE, Oberg K, Zinner MJ, Jaffe BM. The pentagastrin test in the diagnosis of the carcinoid syndrome. Blockade of gastrointestinal symptoms by ketanserin. Ann Surg 1985;201:81-86. a. Richter G, Stockmann F, Lembcke B, Conlon JM, Creutzfeldt W. Short-term administration of the somatostatin analogue SMS 201-995 in patients with carcinoid turnours. Stand J Gastroenterol1986;21:193-198. 9. Anderson JV, Bloom SR. Neuroendocrine tumours of the gut: long-term therapy with the somatostatin analogue SMS ZOl995. Stand J Gastroentero11986;21:115-128. 10. Tsai ST, Lewis E, Vinik A. The use of a somatostatin analogue (SMS 201-995) in the management of the flushing syndrome. Stand J Gastroenterol1966;21:267-274. 11. Oberg K, Norhei I, Lundqvist G, Wide L. Treatment of the carcinoid syndrome with SMS 201-995, a somatostatin analogue. Stand J Gastroenterol1986;21:191-192. 12. Deghenghi R. Somatostatin analogues in the treatment of the carcinoid syndrome. Biomed Pharmacother 19aa;42:585-588. 13. Roy RC, Carter RF, Wright PD. Somatostatin, anaesthesia, and the carcinoid syndrome. Peri-operative administration of a somatostatin analogue to suppress carcinoid tumour activity. Anaesthesia 1967;42:627-632. 14. Van Houten AA, Nortier JWR, Vendrik CPJ. Successful symptomatic treatment of malignant carcinoid syndrome with the somatostatin analogue SMS 201-995. Neth J Med 1966;32:194198. 15. Kvols LK, Moertel CG, O’Connell MJ, Schutt AJ, Rubin J, Hahn RG. Treatment of the malignant carcinoid syndrome. Evaluation of a long-acting somatostatin analogue. N Engl J Med 1966;315:663-666.
GASTROENTEROLOGY
Vol. 101, No. 2
16. Vinik A, Moattari AR. Use of somatostatin analog in management of carcinoid syndrome. Dig Dis Sci 1989;34:145-27s. 17. Engstrom PF, Lavin PT, Moertel CG, Folsch E, Douglass HO Jr. Streptozocin plus fluorouracil versus doxorubicin therapy for metastatic carcinoid tumor. J Clin Oncoli984;2:1255-1259. ia. Moertel CG, Hanley JA. Combination chemotherapy trials in metastatic carcinoid tumor and the malignant carcinoid syndrome. Cancer Clin Trials 1979;2:327-334. 19. Kvols LK, Buck M. Chemotherapy of metastatic carcinoid and islet cell tumors. A review. Am J Med 1987;62:77-83. 20. Committee on Care and Use of the “Nude” Mouse. Guide for the care and use of the nude (thymus-deficient) mouse in biomedical research. BAR News 1976;19:Ml-MZO. 21. Singh P, Rae-Venter B, Townsend CM Jr, Khalil T, Thompson
22.
23.
24.
25.
26.
27. 28.
29.
30.
31.
32.
33.
34.
35.
JC. Gastrin receptors in normal and malignant gastro-intestinal mucosa: age-associated changes. Am J Physiol1965;249:G761G769. Singh P, Le S, Beauchamp RD, Townsend CM Jr, Thompson JC. Inhibition of pentagastrin-stimulated up-regulation of gastrin receptors and growth of mouse colon tumor in vivo by proglumide, a gastrin receptor antagonist. Cancer Res 1987;47:5ooo5004. Reubi JC, Haecki WH, Lamberts SWJ. Hormone-producing gastrointestinal tumors contain a high density of somatostatin receptors. J Clin Endocrinol Metab 1967;65:1127-1134. Ogur M, Rosen G. The nucleic acids of plant tissues. I. The extraction and estimation of desoxypentose nucleic acid and pentose nucleic acid. Arch Biochem 1950;25:262-276. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem 1951;193:265-275. Burton K. A study of the conditions and mechanisms of the diphenylamine reaction for the calorimetric estimation of deoxyribonucleic acid. Biochemistry 1956;62:315-323. Dische Z. Qualitative and quantitative calorimetric determination of heptoses. J Biol Chem 1953;204:963-997. Ishizuka J, Beauchamp RD, Greeley GH Jr, Townsend CM Jr, Thompson JC. 5-Hydroxytryptamine is an autocrine factor for carcinoid cells (abstr). Gastroenterology 1990;9a:A287. Jeng Y-J, Townsend CM Jr, Nagasawa S, Chuo S, Kern K, Yanaihara N, Ferrar RS, Thompson JC, Greeley GH Jr, Regulation of pancreastatin release from a human pancreatic carcinoid cell line in vitro. Endocrinology 1991;12a:220-225. Chuo S, Townsend CM Jr, Farrar S, Kern K, Yanaihara N, Thompson JC, Greeley GH Jr. Coordinate pancreastatin and chromogranin A secretion from a human pancreatic carcinoid: evidence of chromogranin processing to pancreastatin (abstr). Gastroenterology 1990;96:A653. Norheim I, Wilander E, Oberg K, Theodorsson-Norheim E, Lundqvist M-L, Lindgren P, Bergh J. Tachykinin production by carcinoid tumours in culture. Eur J Cancer Clin Oncol1967;23: 669-695. Ahlman H, Ahlund L, Nilsson 0, Skolnik G, Theodorsson E, Dahlstrom A. Carcinoid tumour cells in long-term culture: release of serotonin but not of tachykinins on stimulation with adrenoceptor agonists. Int J Cancer 1988;42:506-510. Debons-Guillemin M-C, Launay J-M, Roseto A, Peries J. Serotonin and histamine production by human carcinoid cells in culture. Cancer Res 1962;42:1513-1516. Hosoda S, Suzuki K, Sudo K, Yoshida N, Tanaka C. Transplantable argyrophilic gastric carcinoid of mastomys natalensis secreting both histamine and serotonin. J Nat1 Cancer Inst 1979;63:1447-1449. Nilsson 0, Ahlman H, Ericson LE, Skolnik G, Dahlstrom A. Release of serotonin from human carcinoid tumor cells in vitro and grown in the anterior eye chamber of the rat. Cancer 1986;58:676-684.
August
1991
36. Nilsson 0, Dahlstrom A, Gronstad K-O, Rosengren L, Briving C, Skolnik G, Ahlman H. Successful transplantation of a human midgut carcinoid tumour to the anterior eye chamber of the rat. Acta Physiol Stand 1984;120:317-319. 37. Gronstad K-O, Nilsson 0, Hedman I, Skolnik G, Dahlstrom A, Ahlman H. On the mode of action of the pentagastrin test in the carcinoid syndrome. Stand J Gastroenterol 1985;20:508-511. 38. Lawrence JP, Haber B, Ishizuka J, Thompson JC. The effect of various gastrin forms on secretion on a carcinoid tumor in vitro (abstr). Gastroenterology 1989;96:A291. 39. Evers BM, Parekh D, Townsend CM Jr, Thompson JC. Somatostatin and analogues in the treatment of cancer. A review. Ann Surg 1991;213:190-198. 40. Setyono-Han B, Henkelman MS, Foekens JA, Klijn JGM. Direct inhibitory effects of somatostatin (analogues) on the growth of human breast cancer cells. Cancer Res 1987;47:1566-1570. 41. Weber C, Merriam L, Koschitzky T, Karp F, Benson M, Forde K, LoGerfo P. Inhibition of growth of human breast carcinomas in viva by somatostatin analogue, SMS #20X-995, treatment of nude mouse xenografts. Surgery 1989;106:416-422. 42. Scambia G, Panici PB, Baiocchi G, Perrone L, Iacobelli S, Mancuso S. Antiproliferative effects of somatostatin and the somatostatin analog SMS 201-995 on three human breast cancer cell lines. J Cancer Res Clin Oncol1987;114:306-308. 43. Schally AV, Redding TW. Somatostatin analogs as adjuncts to agonists of luteinizing hormone-releasing hormone in the treatment of experimental prostate cancer. Proc Nat1 Acad Sci USA 1987;84:7275-7279. 44. Upp JR Jr, Olson D, Poston GJ, Alexander RW. Townsend CM Jr, Thompson JC. Inhibition of growth of two human pancreatic in vim by somatostatin analog SMS 201-995. adenocarcinomas Am J Surg 1988;155:29-35. 45. Milhoan RA. Trudel JL, Lawrence JP, Townsend CM Jr, Thompson JC. Somatostatin inhibits growth of human small cell lung carcinoma in vim. Surg Forum 1988;39:438-439. 46. Reubi JC. A somatostatin analogue inhibits chondrosarcoma and insulinoma tumourgrowth. ActaEndocrinol1985;109:108114. 47. Stockmann R, Richter G, Lembcke B, Conlon JM, Creutzfeldt W. Long-term treatment of patients with endocrine gastrointestinal tumours with the somatostatin analogue SMS 201-995. Stand J Gastroenterol 1986;21:230-237. 48. Souquet J-C, Sassolas G, Forichon J, Champetier P, Partensky C, Chayvialle J-A. Clinical and hormonal effects of a long-
HUMAN
CARCINOID
IN NUDE MICE
311
acting somatostatin analogue in pancreatic endocrine tumors and in carcinoid syndrome. Cancer 1987;59:1654-1660. 49. Ahlman H, Ahlund L, Dahlstrom A, Nilsson 0, Skolnik, Tisell L-E, Tylen U. Use of a somatostatin analogue in association with surgery and hepatic arterial embolisation in the treatment of the carcinoid syndrome. Br J Cancer 1987;56:840-842, 50. Sjoerdsma A, Schechter PJ. Chemotherapeutic implications of polyamine biosynthesis inhibition. Clin Pharmacol Ther 1984; 35:287-300. 51. Heby 0, Janne J. Polyamine antimetabolites: biochemistry, specificity and biological effects of inhibitors of polyamine synthesis. In: Morris DR, Marton JL, eds. Polyamines in biology and medicine. New York: Marcel Dekker, 1981:243-310. 52. Saydjari R, Townsend CM Jr, Barranco SC, James E, Thompson JC. Effects of cyclosporin A and a-difluoromethylornithine on the growth of hamster pancreatic cancer in vitro. J Nat1 Cancer Inst 1986;77:1087-1092. 53. Marx M, Glass EJ, Townsend CM Jr, Barranco SC, Thompson JC. Effects of o-difluoromethylornithine (DFMO) on mouse colon cancer in vitro and in viva (abstr) Gastroenterology 1983;84:A1242. 54. Milhoan RA, Milhoan LH, Trudel JL, Lawrence JP, Townsend CM Jr, Thompson JC. Somatostatin and a-difluoromethylornithine (DFMO) have additive inhibitory effects on growth of human colon carcinoma (HCC) in viva. Biomed Res 1988;9:A30. 55. Barranco SC, Ford PJ, Townsend CM Jr. Heterogeneous survival responses of human gastric cancer clones to a-difluoromethyl-ornithine in vitro. Invest New Drugs 1986;4:337-345. 56. Upp JR Jr, Beauchamp RD, Townsend CM Jr, Barranco SC, Singh P, Rajaraman S, James E, Thompson JC. Inhibition of human gastric adenocarcinoma xenograft growth in nude mice by oc-difluoromethylornithine. Cancer Res 1988;48:3265-3269.
Received August 17,199O. Accepted December 7, 1990. Address requests for reprints to: Courtney M. Townsend, Jr., M.D., Department of Surgery, The University of Texas Medical Branch, Galveston, Texas 77550. Dr. Kim is a visiting scientist from the Department of Surgery, College of Medicine, Seoul National University, Seoul, Korea. Supported by grants from the National Institutes of Health (5R37 DK 15241, PO1 DK 35608, CA 38561), from the American Cancer Society [PDT 220) and in conjunction with the Walls Medical Research Foundation.
