Identification of a c&Regulatory Element Mediating Somatostatin Inhibition of Epidermal Growth Factor-Stimulated Gastrin Gene Transcription

Dale Bachwich,

Juanita

Gastrointestinal Unit Massachusetts General Boston, Massachusetts Center for the Studv of Massachusetts General Boston, Massachusetts

Merchant*,

and Stephen

J. Brand

Hospital 02114 lnflammatorv Bowel Disease Hospital ’ 02114

(S.J.B.)

Gastrin oligonucleotide constructs lacking the D oligonucleotide (gatcCATATGGCAGGGTA), located at -82 to -89 in the 5’-flanking DNA, were not inhibited by somatostatin, indicating that a somatostatin inhibitory c&element is located between -82 and -89 in the 5’-flanking DNA of the human gastrin promoter. (Molecular Endocrinology 6: 1175-1184, 1992)

Antral gastrin secretion and gene expression is inhibited by the paracrine release of somatostatin from antral D cells. Transforming growth factor-a and epidermal growth factor (EGF) stimulate gastrin reporter gene constructs when transfected into pituitary GH., cells. Somatostatin inhibits EGF stimulation of gastrin gene expression, which is in part mediated at the level of transcriptional regulation as somatostatin inhibits EGF stimulation of gastrin reporter gene constructs. Somatostatin inhibition was abolished by pertussis toxin, indicating somatostatin inhibits transcription through the inhibitory G protein Gi. Somatostatin inhibition was unaffected by vanadate and okadaic acid, implying this inhibitory pathway is mediated neither through phosphotyrosine phosphatases nor serinelthreonine phosphatases, respectively. Gastrin reporter genes containing 82 base pairs of the 5’-flanking DNA were sufficient to confer both EGF responsiveness and inhibition by somatostatin in GH, cells. However, transcription of a gastrin reporter gene construct containing only the EGF response element (GGGGCGGGGTGGGGGG), located at -88 to -53, was stimulated by EGF but was not inhibited by somatostatin. Thus, somatostatin inhibits EGF-stimulated gastrin gene transcription by a mechanism other than by interfering with cell signals elicited by the EGF receptor. Since the 82 GASCAT is inhibited by somatostatin, this result also implies that sequences adjacent to the EGF response element contain a cis-regulatory element mediating transcriptional inhibition by somatostatin. This &-element was located using gastrin reporter genes comprising sequential segments of the human gastrin promoter sequence from the transcriptional start site to -82 in the 5’-flanking DNA. 0888-9809/92/l 175-l 184$03.00/0 Molecular Endocrinology Copyright 0 1992 by The Endocrine

INTRODUCTION

Somatostatin is a widely expressed inhibitory peptide found within the nervous system, endocrine tissues, and the gastrointestinaltract (1). Somatostatin acts as a paracrineinhibitor of hormonereleasefrom the pituitary, pancreaticislets,and stomach(1). In the stomach, gastric acid stimulates the release of somatostatin, which then mediatesa paracrinefeedback inhibitionof gastrin secretion from antral G cells (2). Somatostatin inhibitsactivation of cells by extracellular stimuli, which work through different signaltransduction mechanisms.In rat GH4 pituitary cell lines, somatostatininhibits cell stimulationby TRH and vasoactive intestinal peptide (VIP) (3), which respectively activate the phospholipaseC and CAMP-signalingpathways. In both cases, the somatostatin inhibition is blocked by pertussistoxin (4, 5) consistent with the somatostatininhibitionbeing mediatedby an inhibitory GTP bindingprotein, Gi (5). In primaryculturesof canine G cells, somatostatininhibitsgastrinreleasethrough Gi since pertussistoxin blocks this inhibition(6). In addition to inhibiting hormone secretion, in vivo administrationof somatostatin inhibits the stimulated gastrin gene expressionthat results from inhibitionof gastric acid secretion (7). Suppressionof gastrin gene expressionby somatostatinis mediated,at leastin part,

Society

1175

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.

MOL 1176

ENDO.

1992

by inhibition of gastrin transcription because in vitro somatostatin inhibits stimulation of gastrin reporter genes stably transfected into GH4 cells (8). Somatostatin had little effect on the basal gastrin transcription in GH, cells but inhibited gastrin transcription stimulated by epidermal growth factor (EGF), VIP, or TRH (8). Somatostatin selectively inhibits gastrin gene transcription, since somatostatin does not inhibit stimulation of GH gene transcription in pituitary cells (9). The present study analyzes the intracellular mechanism through which somatostatin inhibits EGF activation of gastrin gene transcription in GH., cells. This rat pituitary cell line is used because the gastrin gene is normally expressed in the anterior pituitary at levels l10% of the level expressed in the antrum (lo), and no antral G cell line exists. While the GH,, cell line does not express the endogenous rat gastrin gene, both the rat and human gastrin genes are expressed when they are transfected into the cell line at levels comparable to the endogenous rat GH promoter (8). The GH4 cell model has been used successfully to model other aspects of gastrin gene expression (8, 11) and has been used extensively to study the mechanism of action of somatostatin (12). While primary G cell cultures have been used successfully to study regulation of gastrin secretion, these primary G cell cultures are insufficiently pure and difficult to transfect, rendering them unsuitable for transfection studies. The regulatory pathways in GH4 cells are highly relevant to those found in antral G cells. Both the antral G cell and the GHI cell are regulated by the neuropeptides gastrin releasing peptide and somatostatin (6, 8, 13, 14). Somatostatin exerts its inhibition in both antral G cells and GH4 cells through G, (5, 6). Transforming growth factor-a (TGF-a), an EGF homolog, is a paracrine regulatory peptide in the gastric mucosa (15). EGF stimulates gastrin gene expression in gastric carcinoma cell lines (our unpublished observations) and in GH4 cells transfected with the gastrin gene (8). EGF stimulation of gastrin transcription in GH4 cells is mediated by a gastrin EGF response element (gERE) lying between -54 and -68 in the gastrin promoter, which binds a novel DNA binding protein (11). In this study GH, cells have been used to analyze somatostatin inhibition of EGF-stimulated gastrin gene transcription. Minimal EGF-responsive gastrin reporter genes were constructed to analyze the mechanism through which somatostatin inhibits gastrin transcription. Somatostatin does not inhibit gastrin gene transcription by blocking EGF receptor transmembrane signaling, since the minimal EGF-responsive gastrin reporter gene was not inhibited by somatostatin. Somatostatin inhibition of gastrin transcription required a cis-regulatory element located immediately 5’ to the gERE.

RESULTS A previous study demonstrated in vivo that somatostatin inhibits stimulation of antral gastrin mRNA in-

Vol6

No. 8

duced by achlorhydria. However, this previous in vivo study could not determine if somatostatin inhibits G cells simply by an indirect mechanism, thorough inhibiting release of G cell stimuli in the antral mucosa (7). A direct inhibitory action of somatostatin on gastrin gene expression in isolated cells was demonstrated by transfecting gastrin gene constructs into GH4 cells. To determine whether somatostatin inhibits stimulation of gastrin mRNA levels, inhibition of gastrin gene expression by somatostatin was studied in GH4 cells stably transfected with a human gastrin minigene. The gastrin minigene construct contains the human gastrin gene with 1.3 kilobases (kb) 5’-flanking DNA and 2 kb 3’flanking DNA. This construct is highly expressed in GH4 cells, and gastrin mRNA levels are stimulated by EGF (11). Gastrin-expressing GH, cells were incubated with various combinations of somatostatin analog octreotide SMS 201-995 (SMS) (500 nM) and EGF (10 nM) for 24 h. Somatostatin reduced basal gastrin mRNA levels and inhibited the stimulation of gastrin mRNA levels by EGF (Fig. 1). These changes were selective, since there was no decrease in glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels. Somatostatin inhibition of gastrin gene expression in GHI cells is mediated at least partly at the level of gene transcription. GH, cells were stably transfected with gastrin chloramphenicol acetyltransferase (CAT) reporter genes 1300 GASCAT and 82 GASCAT which contained, respectively, 1300 or 82 base pairs (bp) 5’flanking DNA sequence and the first exon human gastrin gene ligated 5’ to the CAT reporter gene in the pOCAT plasmid. Both these constructs initiate mRNA transcripts from the authentic transcriptional initiation site of the gastrin promoter after transfection into GH, cells (11). Figure 2 shows that EGF stimulates gastrin reporter gene activity of both 82 GASCAT and 1300 GASCAT constructs, confirming previous observations (8,ll). Somatostatin did not inhibit reproducibly unstimulated basal reporter gene activity of either the 1300 GASCAT or 82 GASCAT constructs. However, somatostatin inhibited EGF-stimulated reporter gene activity of both the 1300 GASCAT and 82 GASCAT constructs. Since most inhibitory effects of somatostatin are mediated by inhibitory G,s (12), 82 GASCAT-expressing GH4 cells were incubated with pertussis toxin to determine whether G, mediates somatostatin’s inhibition of transcription. Pertussis toxin did not significantly decrease basal or EGF stimulation of gastrin reporter gene activity (Fig. 3). However, pertussis toxin abolished somatostatin inhibition of EGF-stimulated gastrin transcription from the 82 GASCAT construct, demonstrating that G, mediates inhibition of gastrin transcription by somatostatin. Somatostatin activates phosphotyrosine phosphatases, resulting in the dephosphorylation of tyrosine residues on the EGF receptor (16) and sodium vanadate inhibits these phosphotyrosine phosphatases (17). Incubation of the 82 GASCAT-expressing GH, cells in sodium vanadate did not affect somatostatin inhibition of gastrin gene transcription (Fig. 3) indicating that

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.

A Somatostatin Inhibitory cis-Element

A

LANE

I

2

3

1177

4

5 \ \\

\ \ 26s

d

B

Optical

6 \ \\

:

\’

\\

\

\\

\

: \

_,

3.M

Density

5.43

LANE

1

Density

13.60

2

5.12

15.66

34

SMS

7.64

56

76

6.21

10.52 6.49

BPM,

7

“’

-:

16s“‘

Optical

6

9.54

EGF

4.71 6.45

6.63

SMS+EGF

Fig. 1. Somatostatin inhibits EGF-Stimulated Gastrin Gene Expression A, Northern blot analysis of human gastrin mRNA expression in GH4 cells. GHI cells stably transfected with the human gastrin gene were left at basal conditions (lanes 1 and 2) or treated with either somatostatin analog SMS 201-995 (SMS) at a concentration of 500 nM (lanes 3 and 4) EGF at a concentration of 10 nM (lanes 5 and 6) or both SMS and EGF (lanes 7 and 8). After 24 h of treatment, total RNA was harvested, 15 Kg/lane resolved on a 1% denaturing agarose gel, and transferred to a nitrocellulose sheet. This nitrocellulose was hybridized with a 32P-labeled riboprobe complementary to the human gastrin exon 3 mRNA, indicating the presence of gastrin mRNA. Optical density measurements are shown below the respective bands. B, Northern blot analysis of GAPDH mRNA. The nitrocellulose was rehybridized with a 32P-labeled riboprobe complementary to the rat GAPDH mRNA. Optical density measurements are shown below the respective bands. C, Northern blot analysis of human gastrin mRNA expression in GH., cells after normalization by GAPDH values. Optical density values of gastrin mRNA signals shown in A were normalized by the respective optical density values of GAPDH signals shown in B.

somatostatin’seffect on gastrin gene transcription is not mediatedthrough a phosphotyrosinephosphatase. Somatostatin activation of Gi releases an as yet unknown diffusible agent which increases potassium

channel conductance (18). The target of this second messengerprobably includesspecificprotein phosphatases, since okadaic acid, a specific serine/threonine phosphataseinhibitor, blocks K’ channelactivation by somatostatin (18). Incubation of the 82 GASCAT-expressing GH4 cells in okadaic acid had no effect on somatostatin inhibition of gastrin gene transcription (Fig. 3) indicatingsomatostatin’seffect on gastrin gene transcription is also not mediated through a serine/ threoninephosphatase. Somatostatincould inhibit EGF stimulationof gastrin transcription essentially by two possiblemechanisms. Somatostatin could interfere with membrane signal transduction of the EGF receptor, analogousto somatostatin’s inhibitionof VIP receptor-mediatedactivation of adenylate cyclase (3). Alternatively, somatostatin could initiate a series of intracellular events, which antagonize somedistal intracellularevent stimulatedby EGF receptor activation, but not block allcellularactions elicitedby EGF receptor stimulation.In the specificcase of the selective inhibition of gastrin transcription, this inhibition could be mediated by a transcription factor which recognizesdifferent cis-regulatory elementsfrom those which mediate EGF responsiveness.To examine this question, somatostatin inhibition of the minimal EGF-responsive gastrin reporter gene construct (2xgERE-ACAT) was measuredafter stable and transient transfection into GH, cells. The 2xgERE-A.CAT construct comprisestwo copies of a gERE ligated 5’ to a basal gastrin CAT reporter gene (A-CAT). The ACAT construct has significant basalactivity after transfection into GH, ceils but is unresponsiveto EGF stimulation (11). Two copies of the gERE conferred EGF responsivenessto the A-CAT reporter gene. The gERE element(GGGGCGGGGTGGGGGG)comprisesa novel EGF responseelementlying between -68 and -53 of the humangastrin gene (11). Although 2xgERE-A.CAT reporter gene transcription increased markedly after EGF stimulation of GH, cells, somatostatin did not inhibit EGF stimulation of 2xgERE-ACAT activity in contrast to 82 GASCAT transcription, in which EGF transcription was inhibited by somatostatin (Fig. 4A). Nevertheless, the GH4 cells expressing the 2xgEREACAT constructs still had functional somatostatinreceptors, sincesomatostatinstill inhibitedVIP-stimulated CAMP accumulationin these cells (Fig. 46). This result impliesthat somatostatin does not act principally to inhibit EGF receptor stimulation of the transcription factors that bind to the gERE. Thus, somatostatin inhibitionof EGF-stimulatedgastrin transcription is not mediated by somatostatin inhibition of EGF receptor membranesignaltransduction. Since the EGF responseof the 82 GASCAT (-82 to +60) construct is inhibited by somatostatin, it implies that somatostatin inhibition is mediated through DNA sequencesbetween -82 to +60 that flank the gERE element (-68 to -53). To identify the cis-regulatory elementsmediatingsomatostatininhibition,oligonucleotide sequencesbetween -82 and -69 (oligo D) and -53 and -29 (oligo B) were ligated into reporter gene

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.

MOL 1178

ENDO.

1992

Vol6

No. 8

n Basal q SMS EGF E?I EGF+SMS

1300 GASCAT

Fig. 2. Somatostatin

82 GASCAT

Inhibits EGF-Stimulated Gastrin Transcription Gastrin-CAT reporter genes were stably transfected into GHI cells. Pooled 82 GASCAT constructs were left at basal conditions or treated with either 500 24 h of treatment, the cells were harvested and assayed for CAT activity. CAT untreated control for each plasmid construct. Experiments were performed at all experiments were then pooled and analyzed by Student’s r test. Error bars

stable clones incorporating the 1300 GASCAT and nM SMS, 10 nM EGF, or both SMS and EGF. After activity is expressed as fold increase relative to the least twice in duplicate or triplicate. The data from represent SEM of the pooled data. l *, P < 0.01.

r = Basal q SMS EGF H EGF+SMS

Control

PT

Van.

O.A.

Fig. 3. Pertussis

Toxin Abolishes Somatostatin’s Inhibition of EGF-Stimulated Gastrin Transcription GH, cells stably transfected with the 82 GASCAT construct were treated with either 100 rig/ml pertussis toxin, 30 PM sodium vanadate, or 10 nM okadaic acid. In the experiments with pertussis toxin, the cells were pretreated for 4 h before the addition of EGF and SMS. In the experiments with vanadate and okadaic acid there was no pretreatment period. Cells were then either left at basal conditions, treated with 500 nM SMS, treated with 10 nM EGF, or a combination of both EGF and SMS for 24 h. The cells were then assayed for CAT activity. CAT activity is expressed as fold increase over basal CAT activity. Experiments were performed at least twice in triplicate. The data from all experiments were then pooled and analyzed by Student’s t test. Error bars represent SEM of the pooled data. *‘, P < 0.01.

constructs 5’ to the basal gastrin promoter construct A-CAT in combination with oligonucleotide C, comprising -68 to -53, containing the gERE. These constructs were transiently transfected into GH, cells and somatostatin inhibition of the EGF-stimulated reporter gene activity determined (Fig. 5). Somatostatin inhibited EGF stimulation of the DCBA CAT oligonucleotide construct, which reconstitutes the DNA sequence in the 5’ flanking

DNA sequence of 82 GASCAT. Furthermore, the EGF response of DCA CAT construct was inhibited by somatostatin. However, the EGF response of the CBA CAT construct was not inhibited by somatostatin. These results imply that sequences within the D oligonucleotide contain cis-regulatory elements that mediate somatostatin inhibition of transcription.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.

A Somatostatin

Inhibitory

cis-Element

1179

20000

n

Basal i&l SMS EGF H EGF+SMS

10000

0

82 GASCAT 40

v) = 8

30

.-s =.E i?

2o

2xgERE-A.CAT

I

?i 5 g

‘0

0

Basal

SMS

VIP

VIP+SMS

Does Not Inhibit EGF-Stimulated Gene Transcription of the 2xgE E-A.CAT Reporter Gene Construct Pooled stable GH., clones incorporating the 82 GASCAT and 2xgERE-A.CAT constructs were left at basal conditions or treated with either 500 nM SMS, 10 nM EGF, or both SMS and EGF. After 24 h of treatment, the cells were harvested and assayed for Fig. 4. A, Somatostatin

CAT activity. CAT activity is expressed as counts per min. Experiments were performed at least twice in duplicate or triplicate. The data from all experiments were then pooled and analyzed by Student’s t test. Error bars represent SEM of the pooled data. l *,

P < 0.01; l , P < 0.05. 8, Somatostatin inhibits VIP-stimulated CAMP accumulation in GH, cells stably transfected with the 2xgEREA.CAT reporter gene construct. Regulation of the 2xgERE-A.CAT Stable GH4 clones by VIP and SMS. Pooled stable GH, clones incorporating the 2xgERE-A.CAT construct were suspended in Ham’s F-10 media with 5 mg/ml lactalbumin and 0.1 mM isobutylmethylxanthine. Aliquots of suspended cells were incubated for 20 min with either 500 nM SMS, 100 nM VIP, both SMS and VIP, or left at basal conditions. The cells were pelleted and the supernatant assayed for CAMP by RIA; n = 6 for each treatment. Error

bars represent

SEM.

**, P
6 for each data point. Error bars represent SEM of the pooled data. **, P < 0.01.

cell lines (22). Since somatostatin did not inhibit the minimal EGF-responsive construct (2xgERE-CAT), somatostatin does not inhibit EGF stimulation by blocking membrane signal transduction of the EGF receptor. It thus seems likely that somatostatin inhibits by an independent pathway, eliciting transcriptional effects distinct

from

those

activated

by

EGF

stimulation.

This

inference is supported by earlier observations that somatostatin also inhibits stimulation of gastrin transcription

by VIP

and

TRH,

peptides

ferent signal transduction Somatostatin inhibitory studied in GH4 pituitary inhibits different signaling anisms (12). Somatostatin lation of GH, cells, which

which

act through

dif-

pathways (8). mechanisms have been well cell lines, and somatostatin pathways by different mechinhibits TRH and VIP stimuare mediated by phospholi-

pase C and CAMP signaling pathways, respectively. Somatostatin inhibits VIP-mediated adenylate cyclase activation, which reduces the CAMP second messenger response to hormonal stimulation (3). By contrast, somatostatin inhibition of TRH does not decrease phospholipid turnover (5) but is mediated through membrane hyperpolarization, which decreases intracellular calcium influx (23, 24). Potentially, this could result in selective inhibition of those TRH responses that are mediated by changes in intracellular calcium. By contrast, signaling events activated by diacylglycerol such as c kinase activation may not be inhibited. This would provide a mechanism for somatostatin to selectively inhibit some responses to TRH stimulation. As seen with somatostatin’s inhibition of gastrin secretion from primary G cells (6) inhibition of gastrin

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.

A Somatostatin

inhibitory

c&-Element

transcription is also blocked by pertussis toxin, indicating mediation by G, proteins. G, proteins activated by somatostatin inhibit different signal transduction pathways by different mechanisms. Gi inhibits proximally the CAMP pathway activated by VIP stimulation through antagonizing adenylate cyclase activation by G, (25). Thus, somatostatin reduces the increase in CAMP produced in response to VIP stimulation. Somatostatin inhibits the phospholipase C pathway by acting distally to the generation of the intracellular messengers inositol trisphosphate and diacylglycerol (5). Since protein phosphorylation/dephosphorylation is emerging as a major mechanism through which intracellular signaling regulates gene transcription, effects of somatostatin on specific protein phosphorylation could include changes in specific gene transcription. Although transcriptional regulation by protein kinases has received most attention, specific dephosphorylation of transcription factors is a mechanism of transcriptional regulation (26). While the experiments presented here imply that neither phosphotyrosine phosphatases nor serine/threonine phosphatases mediate the inhibitory effect of somatostatin on gene transcription, the limitations of this cell culture system prevent a definitive answer to this question. Somatostatin inhibition of the 1300 GASCAT construct was more complete than the inhibition of the 82 GASCAT construct (Fig. 2). This may mean that there are additional somatostatin &-elements between -82 and -1300 in the gastrin gene. However, the inhibition of the human gastrin gene expression (Fig. 1) was similar to the 82 GASCAT construct. Further experiments will be required to determine whether additional somatostatin response elements are present in the gastrin gene. Somatostatin inhibition of EGF-stimulated gastrin transcription was dependent on a short cis-regulatory element lying between -82 to -68 of the gastrin promoter. From other examples of signal-regulated gene transcription (27) this DNA element probably binds a specific transcription factor whose function is modulated by cellular responses elicited by somatostatin. However, the -82 to -68 sequence does not inhibit basal activity of the A-CAT basal gastrin reporter gene transfected into GH, cells (11). This suggests that the transcription factors that interact with this sequence are not simple transcriptional repressors. This inference is also supported by the observation that somatostatin inhibits stimulation of ga$trin transcription by EGF but had little effect on basal gastrin transcription. Thus, somatostatin inhibition differs from transcriptional repression mediated by APl and glucocorticoids, which suppress basal as well as stimulated transcription (28). Although the -82 to -68 sequence is necessary for somatostatin inhibition, further studies are required to show that it is sufficient. In contrast to the EGF response where the gERE alone is sufficient (11) somatostatin inhibition may require additional 3’ cis-regulatory sequences in the gERE element. Since somatostatin inhibition has only been examined in the context

1181

of EGF stimulation, all constructs studied contained the gERE element 3’ to the -82 to -68 D oligo. Thus, it is possible that the cis-regulatory elements mediating somatostatin inhibition include additional 3’-sequences that overlap the gERE. If this is the case, it suggests a competitive model (29) whereby somatostatin-modulated factors compete for a DNA binding site which overlaps the gERE element bound by EGF-activated factors. An example of a competitive inhibitory mechanism is seen with APl inhibition of retinoic acid stimulation of osteocalcin gene transcription (30). Alternatively, somatostatin inhibition may be mediated by a transcription factor which binds entirely within -82 to -68 sequence, which interferes with DNA binding or transcriptional activation of EGF-modulated factors to the adjacent gERE. Further experiments analyzing constructs in which the spacing and orientation of the -82 to -68 sequence relative to the gERE are varied should elucidate the mechanism of transcriptional inhibition mediated by somatostatin. The -82 to -68 sequence necessary for somatostatin inhibition contains the sequence CATATGG which comprises an E box motif (31). The E box motif, defined by the consensus sequence CANNTG, binds basic helix-loop-helix (bHLH) transcription factors. Although bHLH transcription factors determine cell-specific gene expression in many different tissues, bHLH proteins are expressed in all mammalian cells studied (32). Interestingly, GH pituitary cells express an E box binding protein that forms a specific DNA/protein complex identical to that seen in islet cells (33). Each cell type has multiple bHLH DNA binding complexes (34) formed through heterodimerization of cell-specific HLH proteins with ubiquitous HLH proteins. These different HLH heterodimer proteins have distinct functional effects on gene transcription. The E box sequence in the rat insulin enhancer binds the E12/PAN 1 HLH protein which activates insulin transcription in islet cells (34, 35) but represses insulin transcription in nonislet cell lines (36). The -82 to -68 E box sequence of the human gastrin gene also binds HLH protein in vitro (our unpublished observations) and interacts with the same islet nuclear proteins that bind to the insulin E box element (37). The function of this -82 to -68 gastrin E box sequence depends on the cell type. The gastrin E box activates basal transcription in islet cell lines but not in GH4 cells (11, 37). Although somatostatin inhibits EGF-stimulated transcription of gastrin reporter genes in GH4 cells, in HIT islet cells somatostatin does not inhibit EGF-stimulated transcription of these constructs (our unpublished data). However, somatostatin does inhibit insulin gene expression in these HIT islet cells through stimulating insulin mRNA degradation (38). Thus, it appears that somatostatin inhibition of gastrin gene transcription is cell type specific. Further studies aimed at determining whether the E box motif in the -82 to -68 element mediates somatostatin inhibition in GH4 cells may reveal novel mechanisms of transcriptional regulation by paracrine regulatory peptides.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.

MOL 1182

ENDO.

MATERIALS Gastrin-CAT

Vol6

AND METHODS Constructs

An Ndel-fstl restriction fragment of the human gastrin promoter consisting of the first 82 bp of the 5’-flanking sequence and the entire first exon (noncoding) was ligated upstream of a promoterless CAT gene in the expression vector pOCAT, hereafter named 82 GASCAT, as previously described (8). The oligonucleotide reporter genes were constructed by synthesizing four oligonucleotides, A to D, comprising sequential segments of the human gastrin promoter sequence from the transcriptional start site to -82 in the 5’-flanking DNA. Oligonucleotides A, B, C, and D encode gastrin sequences from +9 to -28, -29 to -53, -54 to -68, and -69 to -82, respectively. All oligonucleotides included BamHl and Bglll sites at their 5’ and 3’ ends, respectively. An A-CAT reporter gene was first constructed by ligating into the BamHl site of the DOCAT oliaonucleotide A, a 37-bo svnthetic oliaonucleotide’containing-28 bp of 5’-flanking sequence and rhe first 9 bp of exon 1, which includes the TATA box (-26) and the mRNA initiation site of the human gastrin gene. Oligonucleotide reporter genes used in these experiments were then constructed by ligating oligonucleotides B, C, and D either individually or in combination 5’ to oligonucleotide A in the ACAT plasmid as previously described (11). Plasmid DNA used for transfections was purified twice by equilibrium cesium chloride-ethidium bromide density gradient centrifugation. Previous RNase protection analysis with these constructs in the GH., cell line has shown that transcription initiates from the gastrin promoter (11). Cell Culture

and DNA Transfection

GH4 cells were cultured in Ham’s F-l 0 medium (GIBCO-BRL, Gaithersburg, MD) with 12.5% horse serum and 2.5% fetal calf serum, penicillin at 100 fig/ml, and streptomycin at 100 pg/ml in a humidified atmosphere of 5% C02-95% air. For transient transfections, subconfluent GH4 cells were transfected with a mixture containinq DNA at 15 fig/ml precipitated in HEPES-buffered saline (140mM NaCI, 5 mh KCI, 0.75 mM Na7HPOd-2H?O. 6 mM dextrose, and 25 mM HEPES. DH 7.05) and 125 mM- &Cl, for 4 h at ‘37 C and were then’ glycerol shocked for 2 min. The transfected cells were then replated in 60-mm culture plates and incubated for 24 h. Triplicate plates were then either left at basal conditions, treated with SMS 201-995 (Sandoz, East Hanover, NJ) at a concentration of 500 nM alone, EGF at a concentration of 10 nM alone, or a combination of EGF and SMS for 24 h. The gastrin reporter gene activity was measured by assaying CAT enzyme activity (3% Stable transformants were obtained by selecting cells resistant to the aminoglycoside G418 (400 fig/ml; Sigma, St. Louis, MO) after cotransfecting CAT reporter gene constructs with the Rous sarcoma virus-neomycin resistance gene using the calcium phosphate precipitate method. GH, cells stably transfected with each reporter gene construct were grown to 90% confluence in 60-mm plates under conditions described above. Triplicate plates were then either left at basal conditions, treated with SMS 201-995 at a concentration of 500 nM alone, EGF at a concentration of 10 nM alone, or a combination of EGF and SMS for 24 h. The gastrin reporter gene activity was measured by assaying CAT enzyme activity. Northern

Blot

RNA Analysis

Cloned genomic DNA containing the entire human gastrin gene together with 1.3 kb 5’-flanking DNA and 2.5 kb 3’-flanking DNA was stably transfected into GH, cells using the calcium phosphate method and cotransfecting with the Rous sarcoma virus-neomycin resistance gene. Stable transformants were selected as described above. GHI cells stably transfected were

No. 8

grown to 90% confluence in loo-mm plates under conditions described above. Plates were then either left at basal conditions, treated with SMS 201-995 at a concentration of 500 nM alone, EGF at a concentration of 10 nM, or a combination of EGF and SMS for 24 h. Total RNA was extracted from GH, cells by lysis of the cells with guanidine monothiocyanate followed by precipitation of the RNA with 4 M LiCl (40). Fifteen micrograms of total RNA were resolved on a 1% agarose denaturing gel and transferred to nitrocellulose. RNA blots were hybridized with a 32P-labeled probe complementary to exon Ill of the human gastrin gene as previously described (11). VIP Stimulation

of Stably

Transfected

Cells

Pooled stable GH4 clones incorporating the 2xgERE-A.CAT construct were suspended in Ham’s F-l 0 media with 5 mg/ml lactalbumin and 0.1 mM isobutylmethylxanthine. Aliquots of suspended cells were incubated for 20 min with either 500 nM SMS, 100 nM VIP, both SMS and VIP, or left at basal conditions The cells were pelleted and the supernatent assayed for CAMP using an RIA kit (Biotechnologies, Inc. Staunton, MA). Pertussis

Toxin

Treatment

Pooled stable GHI clones incorporating the 82 GASCAT construct were grown to 90% confluence in 60-mm plates as described above. The media was then replaced using media containing pertussis toxin (Sigma) at a concentration of 100 rig/ml, and the cells were incubated for 4 h. Triplicate plates were then either left at basal conditions, treated with SMS 201-995 at a concentration of 500 nM alone, EGF at a concentration of 10 nM alone, or a combination of EGF and SMS for 24 h. The gastrin reporter gene activity was measured by assaying CAT enzyme activity. Sodium

Vanadate

Treatment

Pooled stable GH4 clones incorporating the 82 GASCAT construct were treated for 24 h with varying doses of sodium vanadate (Fisher Scientific, Springfield, NJ) to determine the maximum dose that was not lethal to these cells, as determined by trypan blue exclusion. Vanadate above a concentration of 30 WM killed these cells within 24 h of incubation (data not shown). Pooled stable GHI clones incorporating the 82 GASCAT construct were grown to 90% confluence in 60-mm plates as described above. Triplicate plates were then either left at basal conditions, treated with SMS 201-995 at a concentration of 500 nM alone, EGF at a concentration of 10 nM alone, or a combination of EGF and SMS, all in the presence of 30 PM sodium vanadate, for 24 h. The gastrin reporter gene activity was measured by assaying CAT enzyme activity. Okadaic

Acid

Treatment

A toxicity curve with okadaic acid was performed using pooled stable GHI clones incorporating the 82 GASCAT construct similar to the toxicity curve performed for vanadate as described above. Okadaic acid (Moana Bioproducts, Honolulu, HA) in concentrations above 10 nM were lethal to these cells within 24 h (data not shown). Pooled stable GH4 clones incorporating the 82 GASCAT construct were grown to 90% confluence in 60-mm plates as described above. Triplicate plates were then either left at basal conditions, treated with SMS 201-995 at a concentration of 500 nM alone, EGF at a concentration of 10 nM alone, or a combination of EGF and SMS, all in the presence of 10 nM okadaic acid, for 24 h. The gastrin reporter gene activity was measured by assaying CAT enzyme activity. Statistical

Analysis

Experiments performed

were performed in duplicate or in triplicate and at least twice as shown in the figure legends.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.

A Somatostatin

Inhibitory

c&Element

Statistical significance was assessed by performing t test on data pooled from two or three experiments, n for each result was at least 6.

1183

Student’s and the 16.

Acknowledgments 17. Received December 2, 1991. Revision received May 4, 1992. Accepted May 4, 1992. Address requests for reprints to: Dr. Stephen J. Brand, Gastrointestinal Unit, Jackson 7, Massachusetts General Hospital, Fruit Street, Boston, Massachusetts 02114. This research was supported by NIH Grants R29 DK-40543, ROl DK-42147, and F32 DK-08405, and Center for the Study of Inflammatory Bowel Disease Grant P30 DK-43351. Present address: Gastroenterology Division, University of Michigan Medical Center, Ann Arbor, Michigan 48109.

18.

19.

20.

l

21.

REFERENCES 1. Reichlin S 1986 Somatostatin: historical aspects. Stand J Gastroenterol 21 :(Suppl 119):1-l 0 2. Saffouri B, Weir GC, Bitar K, Makhlouf G 1980 Gastrin and somatostatin secretion by perfused rat stomach: functional linkage of antral peptides. Am J Physiol 238:G495-G501 3. Dorflinger LJ, Schonbrunn A 1983 Somatostatin inhibits vasoactive intestinal peptide-stimulated cyclic adenosine monophosphate accumulation in GH pituitary cells. Endocrinology 113:1541-l 550 4. Koch BD, Schonbrunn A 1984 The somatostatin receptor is directly coupled to adenylate cyclase in GH4Cl pituitary cell membranes. Endocrinology 114:1784-l 790 5. Yajima Y, Akita Y, Saito T 1986 Pertussis toxin blocks the inhibitory effects of somatostatin on CAMP-dependent vasoactive intestinal peptide and CAMP-independent thyrotropin releasing hormone-stimulated prolactin secretion in GH3 cells. J Biol Chem 261:2684-2689 6. Delvalle J, Yamada T 1990 Amino acids and amines stimulate gastrin release from canine antral G-cells via different pathways. J Clin Invest 85:139-143 7. Brand SJ, Stone D 1988 Reciprocal regulation of antral gastrin and somatostatin gene expression by omeprazoleinduced achlorhydria. J Clin Invest 82:1059-l 066 8. Godley JM, Brand SJ 1989 Regulation of the gastrin promoter by epidermal growth factor and neuropeptides. Proc Natl Acad Sci USA 86:3036-3040 9. Barinaga M, Bilezikijan LM, Vale WW, Rosenfeld MG, Evans RM 1985 Independent effects of growth hormone releasing factor on growth hormone release and gene transcription. Nature 314:279-281 10. Rehfeld J 1985 Accumulation of nonamidated preproqastrin and preprocholecystokinin products in porcine $uitan, corticotrophs. J Biol Chem 261:5841-5847 11. Merchant JL, ‘Demediuk B, Brand SJ 1991 A GC-rich element confers epidermal growth factor responsiveness to transcription from the gastrin promoter. Mol Cell Biol 11:2686-2696 12. Schonbrunn A 1990 Somatostatin action in pituitary cells involves two independent transduction mechanisms. Metabolism 39:96-l 00 13. Sugano K, Park J, Soll A, Yamada T 1987 Stimulation of gastrin release by bombesin and gastrin releasing peptide. J Clin Invest 79:935-942 14. Giraud AS, Soll AH, Cuttitta F, Walsh JH 1987 Bombesin stimulation of gastrin release from canine gastrin cells in primary culture. Am J Physiol 252:G413-G420 15. Beauchamp RD, Barnard JA, McCutchen CM, Cherner JA, Coffey Jr RJ 1989 Localization of transforming growth

22.

23.

24.

25.

26.

27.

28.

29. 30.

31.

32.

33.

34.

35.

factor alpha and its receptor in gastric mucosal cells. J Clin Invest 84:1017-l 023 Hierowski MT, Liebow C, du Sapin K, Schally AV 1984 Stimulation by somatostatin of dephosphorylation of membrane proteins in pancreatic cancer MIA PaCa-2 cell line. FEBS Lett 179:252-256 Leis JF, Kaplan NO 1982 An acid phosphatase in the plasma membranes of human astrocytoma showing marked specificity toward phosphotyrosine protein. Proc Natl Acad Sci USA 79:6507-6511 White RE, Schonbrunn A, Armstrong DL 1991 Somatostatin stimulates Ca2+-activated K+ channels through protein dephosphorylation. Nature 351:570-573 Rentoumis A, Chatterjee V, Madison L, Datta S, Gallagher G, DeGroot L, Jameson JL 1990 Negative and positive transcriptional regulation by thyroid hormone receptor isoforms. Mol Endocrinol 4:1522-l 531 Delidow BC, Billis WM, Agarwal P, White B 1991 Inhibition of prolactin gene transcription by transforming growth factor beta in GH3 cells. Mol Endocrinol 5:1716-1722 Davis JRE, Wilson EM, Vidal ME, Johnson AP, Lynch SS, Sheppard MC 1989 Regulation of growth hormone secretion and messenger ribonucleic acid accumulation in human somatotropinoma cells in vitro. J Clin Endocrinol Metab 69:704-708 Gick GG, Bancroft C 1987 Glucocorticoid stimulation of growth hormone messenger ribonucleic acid levels in GH3 cells is inhibited by calcium but not by somatostatin. Endocrinology 120:1986-l 990 Koch BD, Schonbrunn A 1988 Characterization of the cyclic AMP-independent actions of somatostatin in GH cells II. J Biol Chem 263:226-234 Koch BD, Blalock JB, Schonbrunn A 1988 Characterization of the cyclic AMP-independent actions of somatostatin in GH cells I. J Biol Chem 263:216-225 Koch BD. Dorflinaer LJ. Schonbrunn A 1985 Pertussis toxin blocks both-cyclic’ AMP-mediated and cyclic AMPindependent actions of somatostatin. J Biol Chem 260:13138-13145 Boyle WJ, Smeal T, Deflze LHK, Angel P, Woodgett JR, Karin M, Hunter T 1991 Activation of protein kinase C decreases phosphorylation of c-Jun at sites that negatively regulate Its DNA-binding activity. Cell 64:573-584 Montminy MR, Bilezikjian LM 1987 Binding of a nuclear protein to the cyclic-AMP response element of the somatostatin gene. Nature 328:175-l 78 Carsten J, Rahmsdorf HJ, Park K, Cato A, Gebel B, Ponta H, Herrlich P 1990 Antitumor promotion and antiinflammation: down-modulation of AP-1 (Fos/Jun) activity by glucocorticoid hormone. Cell 62:1189-l 204 Levine M, Manley JL 1989 Transcriptional repression of eukaryotic promoters. Cell 59:405-408 Schule R, Umesono K, Mangelsdorf DJ, Bolado J, Pike JW, Evans RM 1990 Jun-Fos and receptors for vitamins A and D recognize a common response element in the human osteocalcin gene. Cell 61:497-504 Church GM, Ephrussi A, Gilbert W, Tonegawa S 1985 Cell type specific contacts to immunoglobulin enhancers in nuclei. Nature 313:798-801 Murre C, McCaw PS, Vaessin H, Caudy M, Jan LY, Jan YN, Cabrera CV, Buskin JN, Hauschka SD, Lassar AB, Weintraub H, Baltimore D 1989 Interactions between heterologous helix-loop-helix proteins generate complexes that bind specifically to a common DNA sequence. Cell 581537-544 Aronheim A, Ohlsson H, Park CW, Edlund T, Walker MD 1991 Distribution and characterization of helix-loop-helix enhancer-binding proteins from pancreatic beta cells and lymphocytes. Nucleic Acids Res 19:3893-3899 Moss LG, Moss JB, Rutter WJ 1988 Systematic binding analysis of the insulin gene transcription control region: insulin and immunoglobulin enhancer utilize similar transactivators. Mol Cell Biol 8:2620-2627 Nir U, Walker MD, Rutter WJ 1986 Regulation of rat

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.

MOL 1184

ENDO.

Vol6

1992

insulin I gene expression: evidence for negative regulation in non-pancreatic cells. Proc Natl Acad Sci USA 83:318031 84 36. Whelan J, Poon D, Weil PA, Stein R 1989 Pancreatic beta cell type-specific expression of the rat insulin II gene is controlled by positive and negative cellular elements. Mol Cell Biol 9:3253-3259 37. Wang TC, Brand SJ 1990 Islet cell-specific regulatory domain in the gastrin promoter contains adjacent positive and negative DNA elements. J Biol Chem 2658908-8914

No. 8

38.

Philippe J 1991 Somatostatin regulates insulin gene expression at a post-transcriptional level. Program of the 73rd Annual Meeting of The Endocrine Society, Washington, DC, 1991, p 264 (Abstract) 39. Neumann JR, Morency CA, Russian KO 1987 A novel rapid assay for chloramphenicol acetyltransferase gene exoression. Biotechniaues 5:444 40. Cathala G, Savouret J, Mendez B, West B, Karin M, Martial JA. Baxter JD 1983 A method for isolation of intact, translationally active ribonucleic acid. DNA 2:329335

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 22 November 2015. at 08:40 For personal use only. No other uses without permission. . All rights reserved.