EXPERIMENTAL
CELL
RESEARCH
186,15-21
(1%))
rice for Autocrine Growth Stimulation mor Cells by a Gastrin/Cholecystokin NASEEMAIVI.HOOSEIN,*~~PETER *Bristol-Baylor tPharmaceutica1
Laboratory, Research
A.KIENER,~
ROBERTC.CURRY,TAND
Department of Pharmacology, and Development Division,
MICHAEL~.
Baylor College of Medicine, Houston, Texas 77030; and Bristol-Myers Company, Wallingford, Connecticut 06492
The peptide gastrin has been shown to have growth stimulatory effects on normal as well as malignant gastrointestinal tissue. In this study, we have examined the possibility of autocrine growth-stimulation of cultured colon tumor cells by a gastrin-like peptide. The gastrin/ CCK receptor antagonist dibutyryl cGMP inhibited the proliferation of two human colon carcinoma cell lines HCT 116 (ECbo = 1.3 m&f) and CBS (EC& = 2.5 mM) in a dose-dependent manner. Marked morphological changes were observed in HCT 116 cells after treatment with I mM dibutyryl cGMP. In receptor binding eGMP competed with 1251-labeled assays, dibutyryl gastrin for binding to HCT 116 cells (I(& = 1.8 m&f). Another derivative of cyclic GMP, 8-Bromo cGMP used as control due to its considerably weaker affinity for the gastrin/CCK receptor, did not compete with radiolabeled gastrin for binding to HCT 116 cells and had no effect on the morphology or proliferation in monolayer cultures of HCT 116 or CBS cells at concentrations up to 10 mM. Antigastrin/CCK antisera was also found to have dose-dependent cytostatic effects on HCT 116 and CBS cells adapted to growth in serum-free medium. The antiproliferative effect of the gastrin/CCK receptor antagonist and antigastrin/CCK antibodies suggested that a gastrin-like peptide secreted by these cells may promote growth. Radioimmunoassay of the conditioned medium of these two cell lines indicated the presence of a gastrin-like peptide (lo-50 pg/106 cells/72 h). Northern analysis using an oligonucleotide DNA probe complementary to the nucleotide sequence coding the dodecapeptide aarboxyl terminal of human gastrin showed three transcripts (O-7,3.3, and 3.7 kb) that hybridized with the probe. These data provide, for the first time, evidence for an autocrine growth stimulatory role of a gastrin/CCK-like peptide in cultured colon tumor cells. 0 1990 Academic Press, Inc.
The gastrointestinal peptide gastrin bas been known for over 80 years as a stimulant of gastric acid secretion [ 11. More recently however, considerable interest has developed regarding the ability of gastrin to act as a growth factor [2-g]. Trophic effects of gastrin were first described in rat oxyntic gland and duodenal mucosa [ Subsequently, the growth of colonic mucosa has al been shown to be promoted by gastrin (reviewed in Ref. [3]). Exogenously added pentagastrin increased total coionic mucosal DNA, protein, and NA content in fasting rats 131. Also, hypogastrinemia induced by antrectomy resulted in significant decreases in tissue weight as well as in DNA and RNA content of rat colonie mueosa [4]. These changes in antrec rats were prevented by pentagastrin treatment, ng that endogenous gastrin is an important regulator of normal colonie mucosal growth and maintenance ]2,4]. Besides its effects on normal ~ast~oi~testi gastrin has been shown to promote the growt rine colonic tumor cell line in uivo [S] and. t the formation of chemically induced rat colorectal tumors [6]. Also, the presence of gastrin receptors has been described in rodent [5] and human colon tumor cell lines [7, 81. Recently, Upp et al. [9] found gastrin receptors in 45 of 67 surgical tumor specimens from patients with primary colon cancers. Immunocytochemical studies have shown the presence of gastrin in cosa [IO]. Gastrinlike immunoreacti found i.n the proximal small intestine oxyntic gland mucosa, the ileum, or the colon [ll]. However, receptors for gastrin are present in normal colonic mucosa in humans [9] = Malignant transformation in the colon co&l therefore involve an aberrant response to circulating gastrin or may be associated with ectopic pro&&ion of a gastrinlike peptide. Autocrine produ.ction of often been linked with the transform 141. To explore this possibility with rega have examined in this study the e
1 Supported by NIH Grants CA34432 and CA45967. * Current address: Department of Biochemistry and Molecular Biology, University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, Texas 77030. 3 To whom reprint requests should be addressed.
4 Abbreviations butyryl-guanosine 15
used: CCK, 3’Scyclic
cholecystokinin; monophosphate;
d 8
co14-4827/90$3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
16
FIG. 1. Magnification
HOOSEIN
Morphology X100.
of HCT
116 colon
tumor
cells.
Control
untreated
antagonist dbcGMP and antigastrin/CCK antisera on two colon carcinoma cell lines: poorly differentiated, aggressively growing HCT 116 cells and well-differentiated, relatively indolent CBS cells. We have also examined the expression and secretion of a gastrin/CCK-like peptide by these two cell lines. Our results indicate that the colon carcinoma cells tested synthesize and secrete a gastrin/CCK-like peptide. Blocking the activity of this endogenously produced peptide by a gastrin receptor antagonist or antisera resulted in inhibition of cellular proliferation, suggesting an autocrine growth stimulatory role for a gastrin/CCK-like peptide. MATERIALS
AND
METHODS
Cell culture. The HCT 116 and CBS colon cancer cell lines were established in vitro from separate human tumor specimens as previously described [15]. Both these cell lines have been adapted to grow in serum-free medium. Working cultures of the cell lines were maintained at 37’C in a humidified atmosphere of 5% CO2 in McCoy’s me-
guanosine 3’:5’-cyclic monophosphate; SDS, sodium dodecyl SSPE, 150 mM sodium chloride, 10 n&f sodium dihydrogen phate, 10 mM ethylenediamine tetracetic acid.
ET
sulfate; phos-
AL.
cells
(A) and those
treated
with
1 m&f dbcGMP
for 5 days
03).
dium 5A (Flow Laboratories) and contained twice the normal concentrations of sodium pyruvate, vitamins, and amino acids. Also included in the serum-free growth medium were transferrin (4 pg/ml), insulin (20 ag/ml), and epidermal growth factor (10 rig/ml). For proliferation assays cells were plated and treated with cGMP derivatives (Sigma) or antigastrin antisera (Amersham) as indicated. Binding assay. Radiolabeling of [Le@] gastrin-17 (Bachem) and binding assays were done essentially as described previously [8]. Nonspecific binding determined in the presence of 225 pi’V [Leui’lgastrin17 was approximately 20% of total binding. I& values were determined by computer fitting using the LIGAND program. Gad-in radioimmunoassay. A solid-phase radioimmunoassay procedure was employed to detect gastrin-like immunoreactivity using gastrin/CCK polyclonal antisera (Amersham) directed toward the common COOH-terminal pentapeptide sequence. The assay was performed in 12 X 75mm star tubes (immunoequality, Nunc, Denmark). Tubes were coated with 0.1 ml of conditioned medium in 0.5 ml of 0.1 Mcarbonate buffer, pH 9.6. Adsorption was facilitated by shaking the tubes overnight at room temperature. Tubes were blocked with 3% BSA in phosphate buffered saline for 2-3 h with constant shaking. Tubes were then treated with 0.5 ml of a 1:2500 dilution of the antiserum. Bound antibodies were detected by incubation for 4 h with iz51protein A (ICN). Tubes were then counted in a gamma counter. This assay enables the measurement of gastrin-like immunoreactivity in the range 0.25-5 pg per assay tube (total incubation volume, 500 ~1). Northern the method
analysis. of Chirgwin
Total cellular RNA was extracted from cells by et al. [16]. Total RNA (50 pg/lane) was sepa-
GROWTH
STIMULATION
BY
A GASTRIN/CCK-LIKE
PEPTEDE
FIG. l-Continued rated by electrophoresis in a 1% agarose gel in the presence of 0.66 M formaldehyde. The RNA was then transferred to Nytran nylon membrane (Schleicher & Schuell) and baked for 2 h. The oligonucleotide DNA probe (36-mer) complementary to the nucleotide sequence coding for the COOH-terminal dodecapeptide of human gastrin was custom-synthesized by Research Genetics (Huntsville, AL) in an auto32P end-labeled probe was prepared by incubating mated synthesizer. 0.5 pg oiigonucleotide with 2.5-fillOX kinase buffer (50 mMDTT, 100 mM MgCl,, 500 mM Tris, pH 7.4), 1.5 ~1 T4 polynucleotide kinase (Boehringer-Mannheim), and 20 ~1 T-~‘P-ATP (200 &i, Amersham). After incubation for I h at 37”C, enzyme was inactivated (65°C for 5 min) and unincorporated 3’P-ATP was removed by chromatography on Whatman DE-52 cellulose. Blots were prehybridized for 3 hat 42°C in 6X SSPE/O.l% SDS/lOX Denhardt’s reagent containing 50 pg/ml yeast tRNA and 50 pg/ml denatured salmon sperm DNA. Hybridization was at 42°C for 22 h in 2 X lo6 cpm of end-labeled probe/ml in 6X SSPE/O.l% SDS. Blots were washed in 2X SSPE/O.l% SDS three times at room temperature for 15 min each followed by a final wash at 42°C. Blots were then autoradiograpbed for 72 h.
RESULTS
Effect of Gastrin Receptor Antagonist dbcGMP on Cellular Morphology and Proliferation Gastrin, CCK, and caerulein are members of a family ofpeptides that share a common COOH-terminal pentapeptide amide [ 171and exert their effects on a particular
with the same class of receptarget tissue by interacting tors [IS]. Butyryl derivatives of c have been shown ts of the biologito be reversible, competitive ant cal action of members of this family of structurally and functionally related peptides [18, 191. ined the effects of two derivatives of c and 8Br cGMP (used as control), on and proliferation of cultured colon tumor cells. Treatment of HCT 116 cells with h 5 days resulted in increased cellular si of numerous cytoplasmic extensions morphological changes indicative of a more differentiatedphenotype were not observed in untreated HGT 116 cells (Fig. 1A) nor were they apparent in cells treated with 1 mlM 8-Br cGMP (data not shown). In proliferation assays with HCT 116 cells (Fig. 2A) and CBS ceils (Fig. 2B), dbcGMP at ~o~ce~trati~~s between 6.5 and 10 mM had dose-dependent cytostatic el’fects. Halfmaximal concentration of dbcG inhibition was 1.3 mA4 for HCT 116 cells a CBS cells. At concentrations of up to cGMP had little effect on cellular 116 or CBS cells (Fig. 2). Thus, in both cell lines examined the dibutyryl but not the B-bromo derivative of
18
HOOSEIN
4
HCTW
o dbcGMP 0 8-Br cGMP
1X 10( % 84 7( 6( !?A G74
0 7
3(
.h I( LI
(
GEE
3
kl 2
30(
Ii 8
ET
AL.
tives has been shown to depend on the presence of a butyryl side chain [19]. In agreement with these findings, 8-Br cGMP (up to 8 mM) in our study failed to inhibit the binding of labeled gastrin to HCT 116 cells (Fig. 3) and also had no effect on cellular morphology or proliferation (Fig. 2). The I(&, value of 1.4 pM obtained for gastrin-17 (Fig. 3) agrees with the results of Weinstock and Baldwin [ 71who have reported binding affinities for gastrin in the micromolar range for gastric carcinoma as well as for a wide variety of other human tumor cell lines. Upp et al. have reported that 58% (38/67) of surgical specimens of human colon tumors had a high gastrinbinding affinity (& < 1 nM) whereas 10% (7/67) had low affinity (& > 1 nM) gastrin binding sites [9]. There is some evidence that the gastrin receptor may exist in two conformational states, one with a high nanomolar binding affinity and the other with the low micromolar affinity observed in colon, gastric, and other human tumor cells [7]. Effect of Antigastrin
20(
10(
(
0
1 LCYCLIC
2
3 GMP
4
5
6
DERIVATIVES
7
8
9
IO
1 mM
FIG. 2. Monolayer proliferation of colon carcinoma cells in the presence and absence of dbcGMP and S-Br cGMP. HCT 116 (serum free, 5000 cells/well) and CBS (serum-free, 10,000 cells/well) were plated in quadruplicate 35-mm dishes. After 24 h cells were treated with varying concentrations of dbcGMP and 8-Br cGMP. HCT 116 cells were counted with a hemocytometer after 6 days and CBS cells after 7 days. Mean values + SD are given. Results shown are representative of three separate assays which gave similar results.
cGMP elicited morphological changes and growth inhibition. Inhibition of Radiolabeled Gastrin Binding to Colon Tumor Cells by dbcGMP. The binding of lz51-labeled cholecystokinin to pancreatic acini has been reported to be inhibited by dbcGMP [ 181. In competition assays dbcGMP inhibited the binding of lz51-labeled [Le@] gastrin-17 to HCT 116 cells half-maximally at a concentration of 1.8 mM (Fig. 3). Thus, there was a close correlation between the ability of dbcGMP to compete with gastrin-binding and its ability to inhibit cellular proliferation (EC& = 1.3 mM; Fig. 2). Antagonism of cholecystokinin-mediated amylase secretion in dispersed pancreatic acini by cGMP deriva-
Antiserum on Proliferation
A commercially prepared antiserum to gastrin/CCK, raised in rabbits using CCK-8 (SO,) covalently conjugated to bovine serum albumin, was used in proliferation assays. The antiserum is specific for the C-terminal. CCK-8 (sulfated or nonsulfated), CCK-33, CCK-39, and gastrin (l-17 and l-34) are all equally well recognized. Results shown in Fig. 4 indicate that this antiserum had concentration-related cytostatic effects on HCT 116 and CBS cells. Preincubation of this antiserum with 25 pM gastrin resulted in abrogation of the growth inhibitory effect in both cell lines (Fig. 4). Three other colon tumor cell lines, RKO, MOSER, and GEO, also adapted to growth in serum-free medium, were similarly growth-inhibited by gastrin/CCK antibodies (data not shown). Radioimmunoassay for Gastrin The antiproliferative effects of gastrin/CCK antiserum (Fig. 4) and the receptor antagonist dbcGMP (Fig. 2) on colon tumor cell lines growing in serum-free medium strongly suggested that a gastrin-like peptide secreted by these cells may stimulate cellular proliferation. We therefore used a radioimmunoassay to measure the amount of gastrin in the conditioned medium. Results are shown in Table 1. Assay of serum-free growth medium conditioned by HCT 116 and CBS cells as well as RKO, MOSER, and GE0 cells (Table 1) indicated the presence of gastrin/CCK-like immunoreactivity between 10 and 50 pg per lo6 cells in 72 h in all of the cell lines examined. Northern Analysis To detect expression of mRNA for a gastrin-like peptide in cultured colon tumor cells we performed Northern analysis using an oligonucleotide probe (36-mer)
GROWTH
9 2 100
STIMULATION
l .a--e----e
---------
m
r-.
80
E cc 5
70
A GASTRIN/CCK-LIKE
PEPTIDE
l 8-Br
90
;
BY
cGMP
60
50 L=i L 40 .-I F p
30 20
-
[CYCLIC FIG. standard
3.
-
.
GMP
DERIVATIVES 1 mM
Competition of’251-[Leu15]-gastrin 17 binding to HCT 116 (serum-free) deviation of assay points done in triplicate was less than 5% and experiments
complementary to the nucleotide sequence coding the dodecapeptide carboxyl terminal of human gastrin. Results shown in Fig. 5 indicate that three transcripts of sizes 0.7,3.3, and 3.7 kb hybridized with this probe. The
g
120
? 110 s (-J 100 cl
90
?
80
9 K
7o
5 3
GASTRIN cells by dbcGMP, were repeated
17, uM 8-Br cGMP, at least three
and [Leu”] times with
gastrin-17. The identical results.
mRNA species that hybridized with the probe were similar for HCT 116 and CBS cells (Fig. 5) and also for RK and GE0 cells (not shown). The mRNA coding for gastrin in porcine antrum has been shown to be approximately 600 nucleotides long [2O, 241. The higher molecular weight species detected in colon tumor cells could represent unprocessed forms of the sage or gastrinrelated sequences. The rat CCK A has been reported to be -750 nucleotides long [22]. The hybridization at 0.7 kb observed at low ringency (Fig. 5) could therefore also be to CCK mRN
This is the first report, to our evidence for positive, autoerine tured colon tumor cells by a gastrm
60 50
Secretion of a Gastrin/CCK-Like Peptide by Human Colon Carcinoma Cell Lines as Determined by Radioimmunoassay Cell line
/ CCK
ANTISERUM
DILUTION
FIG. 4. Cytostatic effect of anti-gastrin/CCK antiserum (Amersham) on the proliferation of colon carcinoma cells adapted to grow in defined, serum-free growth medium. HCT 116 and CBS cells were plated in 24-well plates at 5000 cells/well. Twenty-four hours after plating cells were treated with antiserum alone (0, 0) or with antiserum plus 25 FM gastrin-17 (0, q) at the indicated dilution. Cells were counted 5 days later using a hemocytometer. Nonimmune rabbit IgG (up to 2 me/ml) had no effect on the proliferation of HCT 116 cells. Results are shown as percentage growth of untreated cultures. Mean cell numbers ?z SD (n = 3) are indicated. Experiments were repeated twice with similar results.
HCT 116 RKO MOSER GE0 CBS
Gastrin/CCK-like conditioned
immunoreactivity medium (pg/l@ 23.7 26.7 il.9 35.6 30.3
t+ 2 + +
present in the cells/72 h)
2.0 3.1 1.6’ 4.01 3.1
Note. Cells were grown in serum-free medium to 70-80% confluency in 75cm2 flasks. Fresh growth medium was then added. After 72 h conditioned medium was collected and stored frozen at -20°C until assayed. Values shown are means t SD of six assay points from two separate experiments. Serum-free growth medium (not conditioned by cells), used as control, had undetectable gastrin/CCK-like immunoreactivity. Experiments with HCT 116 and CBS cells were repeated three additional times with similar results.
HOOSEIN
-
28s 3.7 kb 3.3 kb
-
18s
0.7 kb
analysis of HCT 116 and CBS cells using a 36FIG. 5. Northern mer oligonucleotide probe complementary to the nucleotide sequence coding for the dodecapeptide COOH-terminal of human gastrin. See text for experimental details. Amount of total RNA loaded per lane was 50 pg.
In addition to HCT 116 and CBS cells three other colon tumor cell lines (RKO, MOSER, and GEO) adapted to grow in serum-free, defined medium were growth-inhibited by gastrin/CCK antiserum and secreted a gastrin/ CCK-like peptide. Exogenous gastrin-17 (up to 20 PLM) did not promote proliferation in HCT 116 or CBS cells as measured by cell counts or [3H] thymidine incorporation. This finding is in agreement with the results of Weinstock and Baldwin [7], who were also unable to demonstrate any increase in cell number or [3H]thymidine uptake upon gastrin-17 treatment of several gastric carcinoma cell lines that displayed cell surface gastrin
ET
AL.
binding. The lack of effect of exogenous gastrin-17 may be due to occupation of gastrin receptors by an endogenous gastrin/CCK-like peptide. Blocking the interaction of this secreted peptide with the cell surface receptor by gastrin/CCK antiserum resulted in cytostasis which could be reversed by preincubation of the antiserum with gastrin-17 (Fig. 4). When several normal tissues from adult pig were examined for expression of the gastrin gene, gastrin mRNA was detected only in antral mucosa and pituitary [23]. Also, immunocytochemical studies have failed to detect gastrin-like immunoreactivity in mammalian large intestine [ 111. Thus, it appears that gastrin expression in colon tumor cells observed in this study reflects ectopic production which may contribute to the etiology of colon cancer. Aberrant production and autocrine growth stimulation by another peptide hormone bombesin-like peptide has been implicated in small-cell lung carcinoma [24,25]. Several studies have compared control of gastrin gene expression in normal and gastrinoma tissues [26-281. Wiborg et al. [27] reported different start points for gastrin mRNA synthesis in the two tissues while Kariya et al. [28] found no differences in transcriptional start points nor did they find genomic rearrangements associated with gastrinoma formation. Similar studies with colon tumor cells should provide more information regarding ectopic production of gastrin in neoplasia. Recently, Palmer-Smith and co-workers have reported that fasting plasma gastrin levels were significantly greater in patients with adenomatous polyps or colon cancer compared to those in controls [29]. None of these patients had identifiable causes for elevated gastrin, such as drugs or prior surgery. Our results, reported here, suggest that elevated gastrin levels in these patients are due to gastrin secretion by neoplastic colon tissue. In this regard it is of interest that human bronchogenic adenocarcinomas, including large-cell, smallcell, and squamous-cell carcinomas, were found to express the gastrin gene [30]. Gastrin binding sites have also been described on a variety of human tumor cell lines [7]. In preliminary studies we have found that the proliferation of several human breast and lung cancer cell lines was inhibited by gastrin/CCK receptor antagonists, suggesting that a gastrin/CCK-like peptide may play a role in the proliferation of other tumor cell types as well. This possibility and the nature of the molecular form of the gastrin/CCK-like peptide secreted by colon tumor cells needs further examination. The IC5,, values reported for dbcGMP interaction with rat gastrin receptors on gastric glands is 1.2 + 0.4 mM and with rat pancreatic CCK receptors it is 87 + 11 PLM [31]. The low affinity of 1.3 mM for dbcGMP binding to colon tumor cells found in this study suggests that the secreted peptide may be gastrin-like rather than CCK-like. REFERENCES 1.
Edkins,
J. S. (1905)
Proc. R. Sot. Land.
96,376.
GROWTH 2.
L. R., Ames,
Johnson,
D., and Yuen,
STIMULATION L. (1969)
Amer.
BY J. Physiol.
A GASTRIN/CCK-LIKE 18.
Hahne, (1981)
19.
Peiken, S. R., Costenbader, Biol. Chem. 254,5321.
20.
Noyes, G., Mevarech, M., Stein, R., and Agarwal, Proc. Natl. Acad. Sci. USA 76, 1970.
21.
Joon-Yoo, O., Powell, C. Acad. Sci. USA 79,1049.
22.
Deschenes, R. J., Lorenz, L. J., K. J., and Dixon, J. E. (1984) Proc.
23.
Powell, C. T., Ney, C., Aran, Acids Res. 13, 7299.
24.
Cuttita, F., Carney, D. N., Mulshine, J., Moody, J., Fischler, A., and Minna, J. D. (1985) Nature 823.
25.
Layton, Cancer
26.
Boel, E., Vuust, J., Norris, and Marcker, K. A. (1983)
27.
Wiborg, O., Berglund, L., Boel, E., Norris, F., Norris, K., Rehfeeld, J. F., Marcker, K. A., and Vuust, J. (i984) Proc. Natl. Acad. Sci. USA 81,iO67.
28.
Kariya, Y., Kato, K., Hayashizaki, Matsubara, K. (1986) Gene 50,345.
217,251. 3.
Johnson, L. R. (1984) Progress in Cancer Research Vol. 29, pp. 327-336, Raven Press, New York.
4.
Dembinski,
A. Is., andJohnson,
L. R. (1979)
and Therapy,
105,
Endocrinology
769. 5.
Singh, P., Walker, J. C. (1986) Cancer
J. P., Townsend, Res. 46,1612.
6.
Karlin, D. A., McBath, M., Jones, E. D., Elwyn, dahl, M. M. (1985) Cancer Lett. 29,73.
7.
Weinstock,
a.
Hoosein, N. M., Kiener, bra, D. K., and Brattain,
9.
Upp, J. R., Singb, P., Townsend, (1989) Cancer Res. 49,488.
J., and Baldwin,
C. M.,
G. S. (1988)
Jr., and Thompson K. E., and Roms-
Cancer
Res. 48,932.
P. A., Curry, R. C., Rovati, L. C., McGilM. G. (1988) Cancer Res. 48,7199. C. M., Jr., and Thompson,
10.
Greider, M. H., Steinberg, enterology 63,572.
11.
Larsson, L. I. (1979) in Pathology Annual (Sommers, Rosen, P. P., Eds.), pp. 293-316, Appelton-Century-Crofts, York.
12.
Sporn, 878.
13.
Goustin, A. S., Leof, E. B., Shipley, (1986) Cancer Res. 48,1015.
M. B., and Todaro,
6. J. (1980)
N. Engl. G. D.,
and
Gastro-
S. C., and New J. Med.
303,
Moses,
H. L.
Gardner,
C. L., and Gardner.
T., andAgarwa1,
J. D.
J. D. (1979)
J.
K. L. (1979)
K. 5. (1982)
Proc. N&l.
Mad. Acad.
P., a
D. B., Soveny,
‘T. W., Fedorko, @ondon) 316,
C., and Morstyn,
6. (1988)
F., Norris, K., Wind, A., Rebfeid, J. F., Proc. Nat!. Acad. Sci. USA W&2866.
Y., Himeno,
S., Tarui,
S., and
Spandidos,
29.
Brattain, M. G., Brattain, 19. E., Fine, W. D., Khaled, F. M., Marks, M. E., Arcolano, L. A., and Danbury, B. H. (1981) Oncodev. Biol. Med. 2, 355.
Palmer-Smith, Dis. Sci. 34,
30.
Rehfeld, 49,284Q.
Chirgwin, J. M., Przybyla, W. J. (1979) Biochemistry
31.
Chang, R. S. L., Lotti, V. J., Monaghan, R. L., Birnbaum, J., Stapley, E. O., Goetz, M. A., Albers-Schonbeq, G., Patchett, A. A., Liesch, J. M., Hensens, 0. D., and Springer, J. P. (1985) Science 230.177.
Dockray, 1X-139,
Received Revised
Res. 5,485.
J. E., Scanlon, Res. 48,4783.
G. F., and
15.
17.
Anticancer
J. E. (1972)
J. C.
W. F., Jensen, R. T., Lemp, Proc. Natl. Acad. Sci. USA 7
14.
16.
D. (1985)
U., andMcGuigan,
21
PEPTIDE
A. E., McDonald,
B. J., and Rutters,
13,5924.
G. J. (1981) in Gut Hormones (Bloom, Churchill Livingstone, New York.
May 9,1989 version received
August
23, 1989
S. R., Ed.),
pp.
J., Wood, 171.
J. F., Bardram,
J. G.; and Solomon, L., and Hilsted,
T. E. (1989)
L. (1989)
Cancer
Dig.
Res.
CELL
RESEARCH
186,15-21
(1%))
rice for Autocrine Growth Stimulation mor Cells by a Gastrin/Cholecystokin NASEEMAIVI.HOOSEIN,*~~PETER *Bristol-Baylor tPharmaceutica1
Laboratory, Research
A.KIENER,~
ROBERTC.CURRY,TAND
Department of Pharmacology, and Development Division,
MICHAEL~.
Baylor College of Medicine, Houston, Texas 77030; and Bristol-Myers Company, Wallingford, Connecticut 06492
The peptide gastrin has been shown to have growth stimulatory effects on normal as well as malignant gastrointestinal tissue. In this study, we have examined the possibility of autocrine growth-stimulation of cultured colon tumor cells by a gastrin-like peptide. The gastrin/ CCK receptor antagonist dibutyryl cGMP inhibited the proliferation of two human colon carcinoma cell lines HCT 116 (ECbo = 1.3 m&f) and CBS (EC& = 2.5 mM) in a dose-dependent manner. Marked morphological changes were observed in HCT 116 cells after treatment with I mM dibutyryl cGMP. In receptor binding eGMP competed with 1251-labeled assays, dibutyryl gastrin for binding to HCT 116 cells (I(& = 1.8 m&f). Another derivative of cyclic GMP, 8-Bromo cGMP used as control due to its considerably weaker affinity for the gastrin/CCK receptor, did not compete with radiolabeled gastrin for binding to HCT 116 cells and had no effect on the morphology or proliferation in monolayer cultures of HCT 116 or CBS cells at concentrations up to 10 mM. Antigastrin/CCK antisera was also found to have dose-dependent cytostatic effects on HCT 116 and CBS cells adapted to growth in serum-free medium. The antiproliferative effect of the gastrin/CCK receptor antagonist and antigastrin/CCK antibodies suggested that a gastrin-like peptide secreted by these cells may promote growth. Radioimmunoassay of the conditioned medium of these two cell lines indicated the presence of a gastrin-like peptide (lo-50 pg/106 cells/72 h). Northern analysis using an oligonucleotide DNA probe complementary to the nucleotide sequence coding the dodecapeptide aarboxyl terminal of human gastrin showed three transcripts (O-7,3.3, and 3.7 kb) that hybridized with the probe. These data provide, for the first time, evidence for an autocrine growth stimulatory role of a gastrin/CCK-like peptide in cultured colon tumor cells. 0 1990 Academic Press, Inc.
The gastrointestinal peptide gastrin bas been known for over 80 years as a stimulant of gastric acid secretion [ 11. More recently however, considerable interest has developed regarding the ability of gastrin to act as a growth factor [2-g]. Trophic effects of gastrin were first described in rat oxyntic gland and duodenal mucosa [ Subsequently, the growth of colonic mucosa has al been shown to be promoted by gastrin (reviewed in Ref. [3]). Exogenously added pentagastrin increased total coionic mucosal DNA, protein, and NA content in fasting rats 131. Also, hypogastrinemia induced by antrectomy resulted in significant decreases in tissue weight as well as in DNA and RNA content of rat colonie mueosa [4]. These changes in antrec rats were prevented by pentagastrin treatment, ng that endogenous gastrin is an important regulator of normal colonie mucosal growth and maintenance ]2,4]. Besides its effects on normal ~ast~oi~testi gastrin has been shown to promote the growt rine colonic tumor cell line in uivo [S] and. t the formation of chemically induced rat colorectal tumors [6]. Also, the presence of gastrin receptors has been described in rodent [5] and human colon tumor cell lines [7, 81. Recently, Upp et al. [9] found gastrin receptors in 45 of 67 surgical tumor specimens from patients with primary colon cancers. Immunocytochemical studies have shown the presence of gastrin in cosa [IO]. Gastrinlike immunoreacti found i.n the proximal small intestine oxyntic gland mucosa, the ileum, or the colon [ll]. However, receptors for gastrin are present in normal colonic mucosa in humans [9] = Malignant transformation in the colon co&l therefore involve an aberrant response to circulating gastrin or may be associated with ectopic pro&&ion of a gastrinlike peptide. Autocrine produ.ction of often been linked with the transform 141. To explore this possibility with rega have examined in this study the e
1 Supported by NIH Grants CA34432 and CA45967. * Current address: Department of Biochemistry and Molecular Biology, University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, Texas 77030. 3 To whom reprint requests should be addressed.
4 Abbreviations butyryl-guanosine 15
used: CCK, 3’Scyclic
cholecystokinin; monophosphate;
d 8
co14-4827/90$3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
16
FIG. 1. Magnification
HOOSEIN
Morphology X100.
of HCT
116 colon
tumor
cells.
Control
untreated
antagonist dbcGMP and antigastrin/CCK antisera on two colon carcinoma cell lines: poorly differentiated, aggressively growing HCT 116 cells and well-differentiated, relatively indolent CBS cells. We have also examined the expression and secretion of a gastrin/CCK-like peptide by these two cell lines. Our results indicate that the colon carcinoma cells tested synthesize and secrete a gastrin/CCK-like peptide. Blocking the activity of this endogenously produced peptide by a gastrin receptor antagonist or antisera resulted in inhibition of cellular proliferation, suggesting an autocrine growth stimulatory role for a gastrin/CCK-like peptide. MATERIALS
AND
METHODS
Cell culture. The HCT 116 and CBS colon cancer cell lines were established in vitro from separate human tumor specimens as previously described [15]. Both these cell lines have been adapted to grow in serum-free medium. Working cultures of the cell lines were maintained at 37’C in a humidified atmosphere of 5% CO2 in McCoy’s me-
guanosine 3’:5’-cyclic monophosphate; SDS, sodium dodecyl SSPE, 150 mM sodium chloride, 10 n&f sodium dihydrogen phate, 10 mM ethylenediamine tetracetic acid.
ET
sulfate; phos-
AL.
cells
(A) and those
treated
with
1 m&f dbcGMP
for 5 days
03).
dium 5A (Flow Laboratories) and contained twice the normal concentrations of sodium pyruvate, vitamins, and amino acids. Also included in the serum-free growth medium were transferrin (4 pg/ml), insulin (20 ag/ml), and epidermal growth factor (10 rig/ml). For proliferation assays cells were plated and treated with cGMP derivatives (Sigma) or antigastrin antisera (Amersham) as indicated. Binding assay. Radiolabeling of [Le@] gastrin-17 (Bachem) and binding assays were done essentially as described previously [8]. Nonspecific binding determined in the presence of 225 pi’V [Leui’lgastrin17 was approximately 20% of total binding. I& values were determined by computer fitting using the LIGAND program. Gad-in radioimmunoassay. A solid-phase radioimmunoassay procedure was employed to detect gastrin-like immunoreactivity using gastrin/CCK polyclonal antisera (Amersham) directed toward the common COOH-terminal pentapeptide sequence. The assay was performed in 12 X 75mm star tubes (immunoequality, Nunc, Denmark). Tubes were coated with 0.1 ml of conditioned medium in 0.5 ml of 0.1 Mcarbonate buffer, pH 9.6. Adsorption was facilitated by shaking the tubes overnight at room temperature. Tubes were blocked with 3% BSA in phosphate buffered saline for 2-3 h with constant shaking. Tubes were then treated with 0.5 ml of a 1:2500 dilution of the antiserum. Bound antibodies were detected by incubation for 4 h with iz51protein A (ICN). Tubes were then counted in a gamma counter. This assay enables the measurement of gastrin-like immunoreactivity in the range 0.25-5 pg per assay tube (total incubation volume, 500 ~1). Northern the method
analysis. of Chirgwin
Total cellular RNA was extracted from cells by et al. [16]. Total RNA (50 pg/lane) was sepa-
GROWTH
STIMULATION
BY
A GASTRIN/CCK-LIKE
PEPTEDE
FIG. l-Continued rated by electrophoresis in a 1% agarose gel in the presence of 0.66 M formaldehyde. The RNA was then transferred to Nytran nylon membrane (Schleicher & Schuell) and baked for 2 h. The oligonucleotide DNA probe (36-mer) complementary to the nucleotide sequence coding for the COOH-terminal dodecapeptide of human gastrin was custom-synthesized by Research Genetics (Huntsville, AL) in an auto32P end-labeled probe was prepared by incubating mated synthesizer. 0.5 pg oiigonucleotide with 2.5-fillOX kinase buffer (50 mMDTT, 100 mM MgCl,, 500 mM Tris, pH 7.4), 1.5 ~1 T4 polynucleotide kinase (Boehringer-Mannheim), and 20 ~1 T-~‘P-ATP (200 &i, Amersham). After incubation for I h at 37”C, enzyme was inactivated (65°C for 5 min) and unincorporated 3’P-ATP was removed by chromatography on Whatman DE-52 cellulose. Blots were prehybridized for 3 hat 42°C in 6X SSPE/O.l% SDS/lOX Denhardt’s reagent containing 50 pg/ml yeast tRNA and 50 pg/ml denatured salmon sperm DNA. Hybridization was at 42°C for 22 h in 2 X lo6 cpm of end-labeled probe/ml in 6X SSPE/O.l% SDS. Blots were washed in 2X SSPE/O.l% SDS three times at room temperature for 15 min each followed by a final wash at 42°C. Blots were then autoradiograpbed for 72 h.
RESULTS
Effect of Gastrin Receptor Antagonist dbcGMP on Cellular Morphology and Proliferation Gastrin, CCK, and caerulein are members of a family ofpeptides that share a common COOH-terminal pentapeptide amide [ 171and exert their effects on a particular
with the same class of receptarget tissue by interacting tors [IS]. Butyryl derivatives of c have been shown ts of the biologito be reversible, competitive ant cal action of members of this family of structurally and functionally related peptides [18, 191. ined the effects of two derivatives of c and 8Br cGMP (used as control), on and proliferation of cultured colon tumor cells. Treatment of HCT 116 cells with h 5 days resulted in increased cellular si of numerous cytoplasmic extensions morphological changes indicative of a more differentiatedphenotype were not observed in untreated HGT 116 cells (Fig. 1A) nor were they apparent in cells treated with 1 mlM 8-Br cGMP (data not shown). In proliferation assays with HCT 116 cells (Fig. 2A) and CBS ceils (Fig. 2B), dbcGMP at ~o~ce~trati~~s between 6.5 and 10 mM had dose-dependent cytostatic el’fects. Halfmaximal concentration of dbcG inhibition was 1.3 mA4 for HCT 116 cells a CBS cells. At concentrations of up to cGMP had little effect on cellular 116 or CBS cells (Fig. 2). Thus, in both cell lines examined the dibutyryl but not the B-bromo derivative of
18
HOOSEIN
4
HCTW
o dbcGMP 0 8-Br cGMP
1X 10( % 84 7( 6( !?A G74
0 7
3(
.h I( LI
(
GEE
3
kl 2
30(
Ii 8
ET
AL.
tives has been shown to depend on the presence of a butyryl side chain [19]. In agreement with these findings, 8-Br cGMP (up to 8 mM) in our study failed to inhibit the binding of labeled gastrin to HCT 116 cells (Fig. 3) and also had no effect on cellular morphology or proliferation (Fig. 2). The I(&, value of 1.4 pM obtained for gastrin-17 (Fig. 3) agrees with the results of Weinstock and Baldwin [ 71who have reported binding affinities for gastrin in the micromolar range for gastric carcinoma as well as for a wide variety of other human tumor cell lines. Upp et al. have reported that 58% (38/67) of surgical specimens of human colon tumors had a high gastrinbinding affinity (& < 1 nM) whereas 10% (7/67) had low affinity (& > 1 nM) gastrin binding sites [9]. There is some evidence that the gastrin receptor may exist in two conformational states, one with a high nanomolar binding affinity and the other with the low micromolar affinity observed in colon, gastric, and other human tumor cells [7]. Effect of Antigastrin
20(
10(
(
0
1 LCYCLIC
2
3 GMP
4
5
6
DERIVATIVES
7
8
9
IO
1 mM
FIG. 2. Monolayer proliferation of colon carcinoma cells in the presence and absence of dbcGMP and S-Br cGMP. HCT 116 (serum free, 5000 cells/well) and CBS (serum-free, 10,000 cells/well) were plated in quadruplicate 35-mm dishes. After 24 h cells were treated with varying concentrations of dbcGMP and 8-Br cGMP. HCT 116 cells were counted with a hemocytometer after 6 days and CBS cells after 7 days. Mean values + SD are given. Results shown are representative of three separate assays which gave similar results.
cGMP elicited morphological changes and growth inhibition. Inhibition of Radiolabeled Gastrin Binding to Colon Tumor Cells by dbcGMP. The binding of lz51-labeled cholecystokinin to pancreatic acini has been reported to be inhibited by dbcGMP [ 181. In competition assays dbcGMP inhibited the binding of lz51-labeled [Le@] gastrin-17 to HCT 116 cells half-maximally at a concentration of 1.8 mM (Fig. 3). Thus, there was a close correlation between the ability of dbcGMP to compete with gastrin-binding and its ability to inhibit cellular proliferation (EC& = 1.3 mM; Fig. 2). Antagonism of cholecystokinin-mediated amylase secretion in dispersed pancreatic acini by cGMP deriva-
Antiserum on Proliferation
A commercially prepared antiserum to gastrin/CCK, raised in rabbits using CCK-8 (SO,) covalently conjugated to bovine serum albumin, was used in proliferation assays. The antiserum is specific for the C-terminal. CCK-8 (sulfated or nonsulfated), CCK-33, CCK-39, and gastrin (l-17 and l-34) are all equally well recognized. Results shown in Fig. 4 indicate that this antiserum had concentration-related cytostatic effects on HCT 116 and CBS cells. Preincubation of this antiserum with 25 pM gastrin resulted in abrogation of the growth inhibitory effect in both cell lines (Fig. 4). Three other colon tumor cell lines, RKO, MOSER, and GEO, also adapted to growth in serum-free medium, were similarly growth-inhibited by gastrin/CCK antibodies (data not shown). Radioimmunoassay for Gastrin The antiproliferative effects of gastrin/CCK antiserum (Fig. 4) and the receptor antagonist dbcGMP (Fig. 2) on colon tumor cell lines growing in serum-free medium strongly suggested that a gastrin-like peptide secreted by these cells may stimulate cellular proliferation. We therefore used a radioimmunoassay to measure the amount of gastrin in the conditioned medium. Results are shown in Table 1. Assay of serum-free growth medium conditioned by HCT 116 and CBS cells as well as RKO, MOSER, and GE0 cells (Table 1) indicated the presence of gastrin/CCK-like immunoreactivity between 10 and 50 pg per lo6 cells in 72 h in all of the cell lines examined. Northern Analysis To detect expression of mRNA for a gastrin-like peptide in cultured colon tumor cells we performed Northern analysis using an oligonucleotide probe (36-mer)
GROWTH
9 2 100
STIMULATION
l .a--e----e
---------
m
r-.
80
E cc 5
70
A GASTRIN/CCK-LIKE
PEPTIDE
l 8-Br
90
;
BY
cGMP
60
50 L=i L 40 .-I F p
30 20
-
[CYCLIC FIG. standard
3.
-
.
GMP
DERIVATIVES 1 mM
Competition of’251-[Leu15]-gastrin 17 binding to HCT 116 (serum-free) deviation of assay points done in triplicate was less than 5% and experiments
complementary to the nucleotide sequence coding the dodecapeptide carboxyl terminal of human gastrin. Results shown in Fig. 5 indicate that three transcripts of sizes 0.7,3.3, and 3.7 kb hybridized with this probe. The
g
120
? 110 s (-J 100 cl
90
?
80
9 K
7o
5 3
GASTRIN cells by dbcGMP, were repeated
17, uM 8-Br cGMP, at least three
and [Leu”] times with
gastrin-17. The identical results.
mRNA species that hybridized with the probe were similar for HCT 116 and CBS cells (Fig. 5) and also for RK and GE0 cells (not shown). The mRNA coding for gastrin in porcine antrum has been shown to be approximately 600 nucleotides long [2O, 241. The higher molecular weight species detected in colon tumor cells could represent unprocessed forms of the sage or gastrinrelated sequences. The rat CCK A has been reported to be -750 nucleotides long [22]. The hybridization at 0.7 kb observed at low ringency (Fig. 5) could therefore also be to CCK mRN
This is the first report, to our evidence for positive, autoerine tured colon tumor cells by a gastrm
60 50
Secretion of a Gastrin/CCK-Like Peptide by Human Colon Carcinoma Cell Lines as Determined by Radioimmunoassay Cell line
/ CCK
ANTISERUM
DILUTION
FIG. 4. Cytostatic effect of anti-gastrin/CCK antiserum (Amersham) on the proliferation of colon carcinoma cells adapted to grow in defined, serum-free growth medium. HCT 116 and CBS cells were plated in 24-well plates at 5000 cells/well. Twenty-four hours after plating cells were treated with antiserum alone (0, 0) or with antiserum plus 25 FM gastrin-17 (0, q) at the indicated dilution. Cells were counted 5 days later using a hemocytometer. Nonimmune rabbit IgG (up to 2 me/ml) had no effect on the proliferation of HCT 116 cells. Results are shown as percentage growth of untreated cultures. Mean cell numbers ?z SD (n = 3) are indicated. Experiments were repeated twice with similar results.
HCT 116 RKO MOSER GE0 CBS
Gastrin/CCK-like conditioned
immunoreactivity medium (pg/l@ 23.7 26.7 il.9 35.6 30.3
t+ 2 + +
present in the cells/72 h)
2.0 3.1 1.6’ 4.01 3.1
Note. Cells were grown in serum-free medium to 70-80% confluency in 75cm2 flasks. Fresh growth medium was then added. After 72 h conditioned medium was collected and stored frozen at -20°C until assayed. Values shown are means t SD of six assay points from two separate experiments. Serum-free growth medium (not conditioned by cells), used as control, had undetectable gastrin/CCK-like immunoreactivity. Experiments with HCT 116 and CBS cells were repeated three additional times with similar results.
HOOSEIN
-
28s 3.7 kb 3.3 kb
-
18s
0.7 kb
analysis of HCT 116 and CBS cells using a 36FIG. 5. Northern mer oligonucleotide probe complementary to the nucleotide sequence coding for the dodecapeptide COOH-terminal of human gastrin. See text for experimental details. Amount of total RNA loaded per lane was 50 pg.
In addition to HCT 116 and CBS cells three other colon tumor cell lines (RKO, MOSER, and GEO) adapted to grow in serum-free, defined medium were growth-inhibited by gastrin/CCK antiserum and secreted a gastrin/ CCK-like peptide. Exogenous gastrin-17 (up to 20 PLM) did not promote proliferation in HCT 116 or CBS cells as measured by cell counts or [3H] thymidine incorporation. This finding is in agreement with the results of Weinstock and Baldwin [7], who were also unable to demonstrate any increase in cell number or [3H]thymidine uptake upon gastrin-17 treatment of several gastric carcinoma cell lines that displayed cell surface gastrin
ET
AL.
binding. The lack of effect of exogenous gastrin-17 may be due to occupation of gastrin receptors by an endogenous gastrin/CCK-like peptide. Blocking the interaction of this secreted peptide with the cell surface receptor by gastrin/CCK antiserum resulted in cytostasis which could be reversed by preincubation of the antiserum with gastrin-17 (Fig. 4). When several normal tissues from adult pig were examined for expression of the gastrin gene, gastrin mRNA was detected only in antral mucosa and pituitary [23]. Also, immunocytochemical studies have failed to detect gastrin-like immunoreactivity in mammalian large intestine [ 111. Thus, it appears that gastrin expression in colon tumor cells observed in this study reflects ectopic production which may contribute to the etiology of colon cancer. Aberrant production and autocrine growth stimulation by another peptide hormone bombesin-like peptide has been implicated in small-cell lung carcinoma [24,25]. Several studies have compared control of gastrin gene expression in normal and gastrinoma tissues [26-281. Wiborg et al. [27] reported different start points for gastrin mRNA synthesis in the two tissues while Kariya et al. [28] found no differences in transcriptional start points nor did they find genomic rearrangements associated with gastrinoma formation. Similar studies with colon tumor cells should provide more information regarding ectopic production of gastrin in neoplasia. Recently, Palmer-Smith and co-workers have reported that fasting plasma gastrin levels were significantly greater in patients with adenomatous polyps or colon cancer compared to those in controls [29]. None of these patients had identifiable causes for elevated gastrin, such as drugs or prior surgery. Our results, reported here, suggest that elevated gastrin levels in these patients are due to gastrin secretion by neoplastic colon tissue. In this regard it is of interest that human bronchogenic adenocarcinomas, including large-cell, smallcell, and squamous-cell carcinomas, were found to express the gastrin gene [30]. Gastrin binding sites have also been described on a variety of human tumor cell lines [7]. In preliminary studies we have found that the proliferation of several human breast and lung cancer cell lines was inhibited by gastrin/CCK receptor antagonists, suggesting that a gastrin/CCK-like peptide may play a role in the proliferation of other tumor cell types as well. This possibility and the nature of the molecular form of the gastrin/CCK-like peptide secreted by colon tumor cells needs further examination. The IC5,, values reported for dbcGMP interaction with rat gastrin receptors on gastric glands is 1.2 + 0.4 mM and with rat pancreatic CCK receptors it is 87 + 11 PLM [31]. The low affinity of 1.3 mM for dbcGMP binding to colon tumor cells found in this study suggests that the secreted peptide may be gastrin-like rather than CCK-like. REFERENCES 1.
Edkins,
J. S. (1905)
Proc. R. Sot. Land.
96,376.
GROWTH 2.
L. R., Ames,
Johnson,
D., and Yuen,
STIMULATION L. (1969)
Amer.
BY J. Physiol.
A GASTRIN/CCK-LIKE 18.
Hahne, (1981)
19.
Peiken, S. R., Costenbader, Biol. Chem. 254,5321.
20.
Noyes, G., Mevarech, M., Stein, R., and Agarwal, Proc. Natl. Acad. Sci. USA 76, 1970.
21.
Joon-Yoo, O., Powell, C. Acad. Sci. USA 79,1049.
22.
Deschenes, R. J., Lorenz, L. J., K. J., and Dixon, J. E. (1984) Proc.
23.
Powell, C. T., Ney, C., Aran, Acids Res. 13, 7299.
24.
Cuttita, F., Carney, D. N., Mulshine, J., Moody, J., Fischler, A., and Minna, J. D. (1985) Nature 823.
25.
Layton, Cancer
26.
Boel, E., Vuust, J., Norris, and Marcker, K. A. (1983)
27.
Wiborg, O., Berglund, L., Boel, E., Norris, F., Norris, K., Rehfeeld, J. F., Marcker, K. A., and Vuust, J. (i984) Proc. Natl. Acad. Sci. USA 81,iO67.
28.
Kariya, Y., Kato, K., Hayashizaki, Matsubara, K. (1986) Gene 50,345.
217,251. 3.
Johnson, L. R. (1984) Progress in Cancer Research Vol. 29, pp. 327-336, Raven Press, New York.
4.
Dembinski,
A. Is., andJohnson,
L. R. (1979)
and Therapy,
105,
Endocrinology
769. 5.
Singh, P., Walker, J. C. (1986) Cancer
J. P., Townsend, Res. 46,1612.
6.
Karlin, D. A., McBath, M., Jones, E. D., Elwyn, dahl, M. M. (1985) Cancer Lett. 29,73.
7.
Weinstock,
a.
Hoosein, N. M., Kiener, bra, D. K., and Brattain,
9.
Upp, J. R., Singb, P., Townsend, (1989) Cancer Res. 49,488.
J., and Baldwin,
C. M.,
G. S. (1988)
Jr., and Thompson K. E., and Roms-
Cancer
Res. 48,932.
P. A., Curry, R. C., Rovati, L. C., McGilM. G. (1988) Cancer Res. 48,7199. C. M., Jr., and Thompson,
10.
Greider, M. H., Steinberg, enterology 63,572.
11.
Larsson, L. I. (1979) in Pathology Annual (Sommers, Rosen, P. P., Eds.), pp. 293-316, Appelton-Century-Crofts, York.
12.
Sporn, 878.
13.
Goustin, A. S., Leof, E. B., Shipley, (1986) Cancer Res. 48,1015.
M. B., and Todaro,
6. J. (1980)
N. Engl. G. D.,
and
Gastro-
S. C., and New J. Med.
303,
Moses,
H. L.
Gardner,
C. L., and Gardner.
T., andAgarwa1,
J. D.
J. D. (1979)
J.
K. L. (1979)
K. 5. (1982)
Proc. N&l.
Mad. Acad.
P., a
D. B., Soveny,
‘T. W., Fedorko, @ondon) 316,
C., and Morstyn,
6. (1988)
F., Norris, K., Wind, A., Rebfeid, J. F., Proc. Nat!. Acad. Sci. USA W&2866.
Y., Himeno,
S., Tarui,
S., and
Spandidos,
29.
Brattain, M. G., Brattain, 19. E., Fine, W. D., Khaled, F. M., Marks, M. E., Arcolano, L. A., and Danbury, B. H. (1981) Oncodev. Biol. Med. 2, 355.
Palmer-Smith, Dis. Sci. 34,
30.
Rehfeld, 49,284Q.
Chirgwin, J. M., Przybyla, W. J. (1979) Biochemistry
31.
Chang, R. S. L., Lotti, V. J., Monaghan, R. L., Birnbaum, J., Stapley, E. O., Goetz, M. A., Albers-Schonbeq, G., Patchett, A. A., Liesch, J. M., Hensens, 0. D., and Springer, J. P. (1985) Science 230.177.
Dockray, 1X-139,
Received Revised
Res. 5,485.
J. E., Scanlon, Res. 48,4783.
G. F., and
15.
17.
Anticancer
J. E. (1972)
J. C.
W. F., Jensen, R. T., Lemp, Proc. Natl. Acad. Sci. USA 7
14.
16.
D. (1985)
U., andMcGuigan,
21
PEPTIDE
A. E., McDonald,
B. J., and Rutters,
13,5924.
G. J. (1981) in Gut Hormones (Bloom, Churchill Livingstone, New York.
May 9,1989 version received
August
23, 1989
S. R., Ed.),
pp.
J., Wood, 171.
J. F., Bardram,
J. G.; and Solomon, L., and Hilsted,
T. E. (1989)
L. (1989)
Cancer
Dig.
Res.
