Institute of Pharmacology and Toxicology (G.J.M., K.D., M.B.,
M.G.), University of Bonn, Bonn, Germany,
Byk Gulden Lomberg GmbH
(W.A.S.), Konstanz, Germany and
Evangelisches Krankenhaus (D.W.S.),
Bonn, Germany
Radioligand binding experiments were carried out to identify and
characterize nonadrenoceptor [3H]idazoxan binding sites
and [3H](1,2-di-(2-tolyl)guanidine) binding sites in the
rat and human stomach. Furthermore, we examined two selected aspects of
their potential functional significance. Binding of
[3H]idazoxan (Kd = 11.1 nM and
12.4 nM, respectively) and [3H]DTG
(Kd = 932 nM and 242 nM, respectively) to cell
membranes from rat and human stomach was rapid, reversible, specific
and saturable. In rat stomach, binding of the radioligands was
inhibited by imidazolines and by nonimidazoline 
-site ligands,
respectively, at different rank orders of affinity, which suggests the
existence of I2-imidazoline binding sites as well as
2-sites. In two functional models, the direct effects of
I2-site ligands and 2-site ligands on
gastric smooth muscle and glands were investigated. (1) Cirazoline, clonidine and 4-chloro-2-(2-imidazolin-2-ylamino)-isoindoline (BDF
6143) failed to contract the longitudinal muscle of the rat stomach
fundus; BDF 6143 also failed to induce relaxation of this preparation
when it was precontracted with 30 mM KCl. (2) Clonidine, idazoxan, BDF
6143, 1,2-di-(2-tolyl)guanidine, agmatine and
(R)-3-(3-hydroxyphenyl)-N-propylpiperidine up to 100 µM did not
induce acid secretion from rabbit isolated gastric glands. Our data
provide evidence that the rat stomach is endowed with 2
sites and I2 binding sites in addition to the previously
identified non-I1/non-I2
[3H]clonidine binding sites. Our experiments also offer
basic evidence of the existence of I2 and binding sites
in the human stomach. Neither the I2 and
[3H]clonidine binding sites nor the sites in rat
stomach are directly related to a postsynaptic effect on gastric smooth
muscle or to acid release from isolated gastric glands.
 |
Introduction |
Nonadrenoceptor
IBS are recognized with high to moderate affinity by imidazolines and
related compounds, but not by catecholamines. At least two classes of
IBS exist, I1-IBS and I2-IBS, which can be
labeled by [3H]clonidine and [3H]idazoxan,
respectively (for review, see Regunathan and Reis, 1996
; Molderings,
1997). Binding experiments with [3H]clonidine and
[3H]idazoxan in membranes from guinea pig (Houi et
al., 1987), rabbit (Tesson et al., 1992) and rat
gastric tissue (Molderings et al., 1995) provided basic
evidence that IBS are also present in the stomach. Interestingly, it
has been shown that in the rat stomach, -receptor ligands exhibited
a remarkably high affinity for nonadrenoceptor [3H]clonidine binding sites (Molderings et
al., 1995). Moreover, -like sites were recently identified in
the porcine gastric mucosa (Harada et al., 1994), but not in
rat and human gastric tissue.
On the basis of these findings, the first aim of the present study was
to identify and characterize [3H]idazoxan and
[3H]DTG binding sites in the rat and human stomach and to
investigate whether a relationship exists between nonadrenoceptor
[3H]clonidine and [3H]idazoxan binding
sites on the one hand and binding sites on the other. Therefore, we
determined and compared the affinity of key ligands for IBS and sites in rat stomach membranes labeled with
[3H]clonidine, [3H]idazoxan or
[3H]DTG (a radioligand for sites; Weber et
al., 1986).
In previous in vivo studies, imidazolines such as clonidine
exerted a dual action on gastric acid secretion; at low concentrations, these compounds reduced acid secretion (Del Tacca et al.,
1982; Bhandare et al., 1991; Blandizzi et al.,
1995; Carlisle et al., 1995; Glavin and Smyth, 1995). This
inhibitory effect on acid secretion was prevented by alpha-2
adrenoceptor antagonists (Del Tacca et al., 1982; Bhandare
et al., 1991; Blandizzi et al., 1995), which
suggests that it is mediated mainly by activating presynaptic alpha-2 adrenoceptors on cholinergic nerves innervating the
stomach. Additionally, evidence has been presented that peripheral
I1-imidazoline receptors might also contribute to the
antisecretory and antiulcer effects of imidazoline derivatives
(Carlisle et al., 1995; Glavin and Smyth, 1995). At higher
concentrations, several imidazolines and agmatine stimulated acid
secretion in vitro and in vivo (Medgett and
McCulloch, 1979; Del Tacca et al., 1982; Houi et
al., 1987; Bhandare et al., 1991; Glavin et
al., 1995). The stimulatory effect was not due to activation of
alpha-2 adrenoceptors, because it was not mimicked by the
alpha-2 adrenoceptor agonist 
-methylnoradrenaline and it
was not counteracted by yohimbine (Houi et al., 1987). Some
have speculated that the stimulatory effect of the imidazolines may be
due to activation of imidazoline receptors in stomach tissue (Houi
et al., 1987; Bhandare et al., 1991; Glavin
et al., 1995). Therefore, the second aim of this study was
to investigate whether ligands with high affinity for IBS and/or sites could induce acid secretion from isolated rabbit gastric glands,
which is the standard preparation for investigating acid secretion
in vitro.
Finally, it was demonstrated that CDS, a putative endogenous ligand at
IBS (Regunathan and Reis, 1996) induced a contraction of
gastric smooth muscle via an unknown mechanism of action
(Felsen et al., 1987). It has been proposed that imidazoline
receptors on gastric smooth muscle may be involved. Hence the third aim of the present study was to examine whether other ligands at
imidazoline binding sites mimic the effect of CDS and, if so, whether
binding of drugs at [3H]clonidine,
[3H]idazoxan and [3H]DTG binding sites
directly contract rat stomach smooth muscle cells in a vagal
nerve-independent manner. Parts of this study have been presented at
scientific meetings.
| |
Materials and Methods |
Membrane preparation.
Fresh stomachs were obtained from
Wistar Kyoto rats immediately after killing. Segments of
macroscopically normal human stomach were obtained from male or female
patients undergoing gastric surgery. The study was approved in all
respects by the local ethics committee. All steps of the preparation
procedure were performed on ice. The rat glandular stomach and the
segments of the mucosal layer from human stomach were prepared and cut
into small fragments that were placed in 40 ml of buffer solution
containing sucrose 270 mM, ascorbic acid 0.6 mM and Tris-sulfate 10 mM
(pH 7.4), minced by means of an Ultraturax (five times for 20 s
each) and homogenized using a glass-Teflon homogenizer (three times for 30 s each). The homogenates were centrifuged (5 min, 1200 × g, 4°C). The supernatant was filtered through four layers
of gauze, diluted to 420 ml with HEPES buffer (HEPES-Na+ 5 mM, EGTA 0.1 mM, PMSF 0.3 mM, pH 7.4; buffer I) and recentrifuged (20 min, 40,000 × g, 4°C). The pellet was washed twice
and then resuspended in buffer I, homogenized, diluted to give a
protein concentration of about 2 mg/ml and stored at 
80°C until
use. Before use, the membranes were centrifuged (20 min, 40,000 × g, 4°C), resuspended in the incubation buffer
(HEPES-Na+ 5 mM, EGTA 0.5 mM, MgCl2 0.5 mM,
ascorbic acid 0.1 mM, pH 7.4; buffer II), homogenized by ultrasonics
and diluted to a final protein concentration of about 0.6 mg/ml.
Binding assay.
A 400-µl aliquot of membranes was incubated
for 55 min with [3H]idazoxan or [3H]DTG (25 µl) at 4°C in a final volume of 0.5 ml. The reaction was stopped by
rapid vacuum filtration with a Brandel cell harvester through Whatman
GF/C glass-fiber filters presoaked with polyethylenimine 0.5 M and
clonidine 0.1 mM (to reduce filter binding), followed by rapid washing
(within about 5 s) of the incubation tubes and filters with 10 ml
ice-cold buffer II. Filters were placed in 6 ml of scintillation fluid
and shaken overnight, and the radioactivity was determined by liquid
scintillation counting at 44% efficiency. Nonspecific binding was
defined as radioligand binding in the presence of BDF 6143 100 µM
([3H]idazoxan binding) or in the presence of (+)-3-PPP
100 µM ([3H]DTG binding), and it accounted for 5% and
21% of the total radioactivity retained in the filters when
[3H]idazoxan and [3H]DTG 10 nM were used in
the competition experiments, respectively. Adrenaline 10 µM, which
has no affinity for imidazoline binding sites (Molderings et
al., 1993, 1994) was added to the assay to prevent the
radioligands from binding to alpha-2 adrenoceptors.
In rat stomach, saturation studies with [3H]idazoxan were
performed with radioligand concentrations ranging from 0.1 to 46 nM to
determine receptor number and affinity. Because it has been shown that
the affinity of [3H]DTG for peripheral sites is in
the high nanomolar range i.e., a concentration of the
radioligand at which it cannot be used because the cost is too high),
equilibrium-saturation binding of [3H]DTG was performed
by incubating the membranes with 10 nM [3H]DTG and
increasing the concentrations of unlabeled DTG for 55 min at 4°C in
both rat and human stomach. A similar experimental protocol has been
used to determine the affinity and the density of
[3H]idazoxan binding sites in human stomach. This
competitive protocol not only has the advantage over saturation
protocols that less radioligand is consumed, but the contribution of
nonspecific binding is also minimized because the radioligand is used
at a low concentration (DeBlasi et al., 1989). Competition
studies were done using 10 nM [3H]idazoxan or
[3H]DTG, respectively, and 13 different concentrations,
ranging from 0.1 nM to 100 µM, of the unlabeled ligand under
investigation. All experiments were carried out in triplicate.
Determination of gastric acid release.
Gastric glands were
prepared from anesthetized New Zealand rabbits by high-pressure
perfusion of the stomach followed by collagenase digestion of pieces of
fundic mucosa. Gastric glands were suspended in Krebs-Henseleit
solution (NaCl 132.5 mM, KCl 5.4 mM, Na2HPO4 5 mM, NaH2PO4 5 mM, MgSO4 1 mM,
CaCl2 1.2 mM) containing 2 mg/ml glucose and 0.125 µM
[dimethylamine-14C]aminopyrine; pH 7.4. Glands were
incubated at 37°C in the presence of the drugs at the concentrations
indicated. After 30 min, the glands were sedimented by rapid
centrifugation, and the aminopyrine accumulation ratio was determined
from the relation of radioactivity between the glands and the
supernatant (for details, see Simon et al., 1990).
Contraction of gastric smooth muscle.
Gastric longitudinal
muscle strips (3 × 10 mm) were obtained from Wistar-Kyoto rats.
The stomach, opened along the greater curvature, was stripped free of
all mucosal tissue, and longitudinal strips were cut at right angles to
the visible circular muscle bundles. In an organ bath, tissue strips
were pre-equilibrated for about 1 h at 37°C in oxygenated
physiological salt solution of the following composition (mM): NaCl
118, KCl 4.7, CaCl2 2.5, MgCl2 1.2, NaHCO3 25, KH2PO4 1.2, glucose 10;
pH 7.4. This solution contained 1 µM atropine, 1 µM prazosin and 3 µM rauwolscine throughout the experiments to mask
alpha-adrenoceptors and to eliminate potential vagal nerve
influences. Cumulative concentration-response curves were determined
for the drugs under study. Time-matched control experiments were
carried out in parallel. Contractions were monitored isometrically
using Statham force transducers under a load of 1 g. After the
last concentration of the test compound, the preparation was washed
twice, and then KCl was added to the organ bath (final concentration 85 mM) to induce maximum contraction of the smooth muscles. The responses
of the test compounds are expressed as the ratio of the contraction
evoked by the test drugs to that caused by 85 mM KCl.
Protein assay.
Protein concentration was measured by the
method of Bradford (1976) using bovine serum albumin as the standard.
Data analysis.
Data from the saturation and competition
experiments were analyzed using the least-squares fitting program
GraphPADinPlot (GraphPad Software Inc., San Diego, CA). The
significance of the improvement of fit obtained by the two-site
equation over the fit obtained by the one-site equation was analyzed by
the F statistics (partial F test; De Lean
et al., 1982). Receptor number and affinity from homologous
competitive binding experiments were calculated by formulas (3) and (5)
of DeBlasi et al. (1989). Results are expressed as mean
values ± S.E. The statistical significance of differences was
analyzed by Student's t test for unpaired data.
Drugs.
The following drugs were used in this study:
[14C]Aminopyrine (specific activity 96.6 Ci/mmol),
[3H]DTG (specific activity 35.2 Ci/mmol; NEN DuPont,
Dreieich, FRG), [3H]idazoxan (specific activity 45 Ci/mmol; Amersham, England), adrenaline base (Hoechst, Frankfurt,
FRG), bovine serum albumin, agmatine, naphazoline hydrochloride,
haloperidol, (±)pentazocine hydrochloride (Sigma, Munich, FRG), U46619
(Biosigma, Munich, FRG), moxonidine, BDF 6143 hydrochloride
(Beiersdorf, Hamburg, FRG); cirazoline hydrochloride (Synthélabo,
Paris, France), clonidine hydrochloride (Boehringer, Ingelheim, FRG),
(+)-3-PPP hydrochloride, (±)-ifenprodil tartrate, dizocilpine maleate,
DTG (RBI, Natick, MA); histamine dihydrochloride (Merck, Darmstadt,
FRG), (±)idazoxan hydrochloride (Reckitt and Colman, Hull, UK),
rauwolscine hydrochloride (Roth, Karlsruhe, FRG) and ranitidine
hydrochloride (Glaxo, Greeford, UK). Drugs were dissolved in the
respective buffer with the following exceptions. Stock solutions of
adrenaline and haloperidol were prepared in HCl 0.01 M, and DTG was
dissolved in methanol; they were then further diluted in the respective
buffer. The vehicles did not change radioligand binding.
| |
Results |
[3H]idazoxan binding.
The specific binding of
[3H]idazoxan to rat and human stomach membranes was
saturable. Nonlinear regression analysis revealed a reaction of
[3H]idazoxan with one binding site
(Kd = 11.1 ± 2.2 nM and 12.4 ± 4.7 nM, respectively; Bmax = 139 ± 4 fmol/mg
protein and 68 ± 14 fmol/mg protein, respectively; figs.
1 and 2).
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1.
Saturation curve for specific
[3H]idazoxan binding. Rat stomach membranes were
incubated for 55 min at 4°C with increasing concentrations of
[3H]idazoxan. The graph shows one representative
experiment out of four performed in triplicate. At 10 nM
[3H]idazoxan, nonspecific binding accounted for about 5%
of the total radioactivity retained in the filters.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 2.
Competition of unlabeled idazoxan with
[3H]idazoxan (upper panel) and of unlabeled DTG with
[3H]DTG (lower panel) for its specific binding sites in
human stomach membranes. Each point is the mean of five experiments
performed in triplicate. In some cases, S.E. was smaller than the
symbols. At 10 nM [3H]idazoxan and at 10 nM
[3H]DTG, nonspecific binding accounted for about 5% and
12%, respectively, of the total radioactivity retained in the
filters.
|
|
In competition experiments, most of the compounds listed in table
1 inhibited specific binding of 10 nM
[3H]idazoxan to rat stomach membranes in a
concentration-dependent manner; at this radioligand concentration,
specific binding amounted to 1188 ± 106 dpm (corresponding to
about 95% of total binding; n = 48). Competition of
BDF 6143 with [3H]idazoxan 10 nM revealed an inhibition
curve with a slope factor nH of less than 1. Accordingly, the competition curve for this compound was significantly
better fitted to a two-site than to a one-site model (table 1; fig.
3). With the other competitors, monophasic displacement curves were obtained (nH
was not significantly different from 1.0; fig. 3). The
Ki values at the high-affinity site or the
single site for all drugs investigated ranged from 11 to 265,800 nM
(table 1) with the following rank order of affinities: idazoxan > cirazoline = BDF 6143 > naphazoline 
dicozilpine > (+)-3-PPP > (±)-ifenprodil = clonidine (±)-pentazocine > agmatine. Histamine, rauwolscine and
ranitidine at concentrations up to 100 µM inhibited binding by less
than 50% (table 1).
View this table:
[in this window]
[in a new window]
|
TABLE 1
Affinities (Ki values) of imidazolines and other
compounds for [3H]idazoxan binding sites in rat stomach
membranes
Results from computer analysis of competition curvesa obtained
by adding various concentrations of a competing ligand and a fixed
concentration (10 nM) of [3H]idazoxan. The percentages of
high (%high) and low (%low) affinity sites are given
for BDF 6143, which was best resolved (partial F test, last
column) into a two-site fit. In parenthesis is the number of
experiments performed in triplicate.
|
|
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 3.
Competition of imidazoline derivatives (upper panel)
and of -site ligands (lower panel) with [3H]idazoxan
for its specific binding sites in rat stomach membranes.  idazoxan,
 BDF 6143,  clonidine,  (+)-3-PPP,  (±)-pentazocine,  (±)-ifenprodil. Each point is the mean of 3 to 7 experiments performed
in triplicate. S.E. amounted to up to 11% of the respective mean.
|
|
[3H]DTG binding.
The specific binding of
[3H]DTG to rat and human stomach membranes was saturable.
Homologous displacement experiments with unlabeled DTG and
[3H]DTG revealed a reaction of [3H]DTG with
one binding site with Kd values of 932 ± 319 nM and 242 ± 90 nM, respectively, and
Bmax values of 7087 ± 4024 fmol/mg protein
and 3592 ± 553 fmol/mg protein, respectively (figs. 2 and
4).
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 4.
Competition of unlabeled DTG with
[3H]DTG for its specific binding sites in rat stomach
membranes. Each point is the mean of five experiments performed in
triplicate. In some cases, S.E. was smaller than the symbols. At 10 nM
[3H]DTG, nonspecific binding accounted for about 16% of
the total radioactivity retained in the filters.
|
|
In competition experiments, the compounds listed in table
2 inhibited specific binding of
[3H]DTG 10 nM to rat stomach membranes in a
concentration-dependent manner; at this radioligand concentration,
specific binding amounted to 1206 ± 75 dpm (corresponding to 79%
of total binding; n = 56). The drugs investigated as
competitors of [3H]DTG revealed monophasic displacement
curves (nH was not significantly different from
1; fig. 5). The Ki
value of the compounds ranged from 0.5 to 668 µM (Table 2). The rank
order of affinities was as follows (table 2): haloperidol 
(±)-ifenprodil DTG > BDF 6143 clonidine > (+)-3-PPP > cirazoline naphazoline > idazoxan > moxonidine > agmatine.
View this table:
[in this window]
[in a new window]
|
TABLE 2
Affinities (Ki values) of imidazolines and other
compounds for [3H]DTG binding sites in rat stomach membranes
Results from computer analysis of competition curvesa obtained
by adding various concentrations of a competing ligand and a fixed
concentration (10 nM) of [3H]DTG. In parentheses is the
number of experiments performed in triplicate.
|
|
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 5.
Competition of imidazoline derivatives (upper panel)
and -site ligands (lower panel) with [3H]DTG for its
specific binding sites in rat stomach membranes. idazoxan, BDF
6143, clonidine, (±)-ifenprodil,  haloperidol,  DTG.
Each point is the mean of 4 to 7 experiments performed in triplicate.
S.E. amounted to up to 22% of the respective mean.
|
|
Contraction of rat gastric fundus strips.
In time-matched
control experiments with the vehicle applied, the tension of the
preparation slightly decreased by about 2% to 14% (not shown). The
thromboxane analog U46619 induced a concentration-dependent contraction
of the rat gastric longitidinal muscle strips (fig. 6). In contrast, the imidazolines
cirazoline, clonidine and BDF 6143 up to 100 µM (the highest
concentration investigated, fig. 6) failed to elicit a contractile
response. When the smooth muscle preparation was precontracted with 30 mM KCl (a 75% effective concentration) BDF 6143 up to 100 µM failed
to induce a relaxation of the smooth muscles, whereas verapamil (chosen
as the control compound) completely relaxed the preparation
(n = 5; results not shown).
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 6.
Contractile response to the prostanoid U46619 ()
and to the imidazoline derivatives cirazoline (), BDF 6143 () and
clonidine () in the presence of 1 µM prazosin, 1 µM atropine and
3 µM rauwolscine. S.E. amounted to 2% to 12% of the respective
means (n = 5-7).
|
|
[14C]aminopyrine accumulation in rabbit isolated
gastric glands.
In intact gastric glands from rabbits, histamine
significantly increased [14C]aminopyrine accumulation (an
index of acidification within the glands) in a concentration-dependent
manner (fig. 7). However, the
imidazolines idazoxan and BDF 6143, the guanidine agmatine and the
-site ligands (+)-3-PPP and DTG were without effect on the
[14C]aminopyrine accumulation at concentrations up to 0.1 mM (fig. 7). At this high concentration, clonidine increased
[14C]aminopyrine accumulation in 2 of 4 experiments, but
because of the scatter, the mean value did not reach the level of
significance. However, the compound was without effect on
[14C]aminopyrine accumulation at lower concentrations
(fig. 7). Finally, BDF 6143 up to 100 µM did not alter
histamine-induced [14C]aminopyrine accumulation
(n = 3; results not shown).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 7.
Influence of histamine (), of the imidazoline
derivatives idazoxan (), clonidine () and BDF 6143 () and of
the guanidine agmatine (), as well as of the -site ligands DTG
() and (+)-3-PPP () on [14C]aminopyrine
accumulation in rabbit gastric glands. Each point is the mean of 3 to 7 experiments determined in duplicate. S.E. amounted to up to 10%, with
the exception of 0.1 mM clonidine, for which S.E. was 17%. ** P < .005, *** P < .001 (compared with the corresponding
controls).
|
|
| |
Discussion |
[3H]idazoxan binding.
Binding of
[3H]idazoxan to rat and human stomach membranes was
reversible (not shown), specific, saturable and of high affinity; i.e., the criteria for the identification of a specific
binding site are fulfilled. The Kd values of
[3H]idazoxan (11 nM and 12 nM, respectively) were
identical to the Ki value obtained in the
competition experiments with unlabeled idazoxan in rat stomach. The
values are in the range reported in the literature for
I2-imidazoline binding sites (Wikberg et al.,
1992; Regunathan et al., 1993; Miralles et al.,
1993; Molderings et al., 1994).
In the competition experiments, a shallow displacement curve
(nH < 1.0) was obtained with BDF 6143, but not
with the other compounds. The curve of BDF 6143 was fitted best to two
binding sites by computer modeling. Shallow inhibition curves with Hill coefficients significantly different from unity have also been found in
previous studies on bovine chromaffin cells (Molderings et
al., 1993; Regunathan et al., 1993) and on guinea pig
kidney (Wikberg et al., 1992). As an explanation, it has
been proposed that the IBS may exist in two interconvertible forms
(Wikberg et al., 1992; Li et al., 1994).
Alternatively, the data might be explained by assuming monomer-dimer
equilibria (Strange, 1994), in which BDF 6143 would bind with different
preference to the monomer or dimer of the binding site and hence would
represent a case of negative cooperativity. Finally, it is conceivable
that [3H]idazoxan also labeled an I1-like
site and that this accounts for the low Hill slope of BDF 6143. In this
case, however, cirazoline and naphazoline, which also possess high
affinity for the I1-like sites, should also have yielded a
biphasic displacement curve. That did not occur, so this explanation
appears unlikely.
The present data exclude the possibility that
[3H]idazoxan labeled a histamine receptor, in particular
an H2 receptor or an alpha-2 adrenoceptor,
because histamine and ranitidine failed to compete with high affinity
for the [3H]idazoxan site, and the alpha-2
adrenoceptor antagonist rauwolscine exhibited only very low affinity
for [3H]idazoxan sites. In agreement with this, the rank
order of potency of the competing drugs (table 1) differs from the rank
order expected for binding to one of the alpha-2
adrenoceptor subtypes (Bylund et al., 1994). Moreover,
()-adrenaline in the high concentration of 10 µM was included in
the incubation assay to mask the alpha-2 adrenoceptor
population in the membrane preparation.
The pharmacological properties of the idazoxan binding sites in the rat
stomach were compared with those of I1-IBS and
I2-IBS in bovine adrenomedullary membranes (Molderings
et al., 1993, 1994) and of the nonadrenoceptor
[3H]clonidine binding sites in the same tissue
(Molderings et al., 1995). The rank order of affinity of the
competing drugs for the I1-IBS, i.e.,
naphazoline > clonidine = cirazoline > BDF 6143 > idazoxan > phentolamine (Molderings et al., 1993),
and in particular the affinity of clonidine, are clearly different from
the findings in the present experiments (table 1). Therefore, it is
unlikely that the [3H]idazoxan binding site represents an
I1-IBS. The [3H]idazoxan binding sites are
also not identical to the nonadrenoceptor [3H]clonidine
binding sites previously described in the rat stomach (Molderings
et al., 1995) for two reasons: (1) The affinities of the
imidazolines for the [3H]idazoxan binding sites were not
correlated with their affinities for the [3H]clonidine
binding sites in the same tissue (fig.
8A). (2) -Site ligands, which
exhibited high to moderate affinity for the [3H]clonidine
binding sites in the rat stomach, possess only weak affinity for the
[3H]idazoxan binding sites. In contrast, the rank order
of affinity of competing imidazolines and guanidines at
I2-IBS in bovine adrenomedullary membranes was
cirazoline idazoxan > guanabenz > tolazoline > rilmenidine > clonidine > phentolamine (Molderings
et al., 1994) and hence conforms to the rank order found in
our experiments. Accordingly, the [3H]idazoxan binding
sites identified here may be characterized as I2-IBS.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 8.
(Upper panel) Lack of correlation of the
pKi values of imidazolines determined for the
high-affinity [3H]clonidine binding sites in rat stomach
membranes (from Molderings et al., 1995) with their
pKi values determined for the high-affinity
[3H]idazoxan binding sites in the same tissue (present
study). (Lower panel) Highly significant correlation of the
pKi values of imidazolines and -site ligands
determined for the [3H]DTG binding sites in rat stomach
membranes (present study) with their pKi values determined
for the [3H]DTG binding sites (2 sites) on
N1E-115 cells (Molderings et al., 1996). Numbers in the
graph: 1, idazoxan; 2, cirazoline; 3, BDF 6143; 4, naphazoline; 5, agmatine; 6, clonidine; 7, (+)-3-PPP; 8, haloperidol; 9, (±)-ifenprodil; 10, DTG; 11, moxonidine. The dotted line is the line
of identity.
|
|
[3H]DTG binding.
Specific [3H]DTG
binding to the rat and human stomach membranes was saturable,
reversible (not shown) and of moderate affinity and high capacity. The
Kd values for [3H]DTG (932 nM and
242 nM, respectively) were consistent with the values previously
observed in rat brain and liver, in guinea pig brain, in porcine
stomach and on N1E-115 cells (McCann and Su, 1990; DeHaven-Hudkins and
Fleissner, 1992; Harada et al., 1994; Molderings et
al., 1996).
The affinities of the compounds for the [3H]DTG binding
sites determined in the competition experiments were in the micromolar to millimolar range, which is compatible with previously published radioligand binding studies of peripheral binding sites in porcine stomach (Harada et al., 1994) and on N1E-115 cells
(Molderings et al., 1996). Comparison of the
pKi values of the drugs with their affinities
(pKi values) for the 2 sites on
N1E-115 cells (Molderings et al., 1996) revealed a
significant correlation (fig. 8B). These data suggest that the
[3H]DTG binding sites in the rat stomach represent
2 binding sites, which is consistent with the finding of
Harada et al. (1994) in pig stomach.
Because the affinities of the imidazolines BDF 6143 and clonidine for
these sites were in the same range as those of typical -site
ligands such as haloperidol and DTG (table 2), and because it has
recently been shown that nonimidazoline -site ligands possess a
remarkable affinity for the nonadrenoceptor [3H]clonidine
binding sites in the rat stomach, a relationship between the
imidazoline binding sites and the sites in the rat stomach was
conceivable. However, there was no correlation between the affinity of
the drugs for the [3H]DTG binding sites and their
affinity for the nonadrenoceptor [3H]idazoxan and
[3H]clonidine binding sites in rat stomach
(r = 0.15, P > 0.74 and r = 0.10, P > 0.10; seven and nine compounds included in the regression analysis, respectively). Hence the three binding sites appear to
represent different entities.
Functional experiments.
In the stomach, the functional effects
of imidazolines and guanidines that have previously been observed were
tentatively ascribed to an activation of imidazoline receptors. The
present study focused on two clearly nonadrenoceptor-mediated effects: the contraction of gastric smooth muscle and the imidazoline-induced stimulation of gastric acid release. In particular, the question arose
whether these effects were related to one of the heretofore-described binding sites at the level of the effector cells.
It was previously reported that CDS, which is assumed to be an
endogenous ligand at imidazoline bindings sites/receptors (Meeley et al., 1986; Atlas, 1991; Piletz et al., 1995),
elicited a concentration-dependent contraction of rat gastric fundus
strips (Felsen et al., 1987). Because the CDS-induced
contraction was not counteracted by muscarine, bradykinin, serotonin,
angiotensin II, vasopressin or alpha-2 receptor antagonists,
it was suggested that the CDS-evoked contraction was mediated by
imidazoline receptors (Felsen et al., 1987). However, in the
present study BDF 6143, clonidine and cirazoline did not mimic the
effect of CDS reported previously, although the compounds possess high
affinity for I1- and/or I2-imidazoline binding
sites/receptors. In contrast, the thromboxan analog U46619 induced a
contraction of the rat gastric strips, which indicates that the smooth
muscle cells were viable and able to contract. Hence the binding of the ligands to nonadrenoceptor [3H]idazoxan and
[3H]clonidine binding sites in rat stomach did not cause
a contraction of gastric smooth muscle. Because in the stomach BDF 6143 and clonidine additionally exhibited an affinity for the
[3H]DTG binding sites similar to those of the typical
-site ligands (see table 2), it is also unlikely that these
2 sites are involved in contraction of gastric smooth
muscles.
It has also been speculated that the stimulation of gastric acid
release induced by high concentrations of clonidine and related compounds in rat and guinea pig (for references, see the Introduction) might be due to activation of imidazoline recognition sites (Houi et al., 1987; Bhandare et al., 1991).
Nonadrenoceptor [3H]idazoxan (Tesson et al.,
1992) and [3H]clonidine (G.J. Molderings, unpublished
results) binding sites have also been found in the rabbit stomach.
Therefore, [14C]aminopyrine accumulation in isolated
rabbit gastric glands was used to investigate whether imidazolines
increase acid secretion by direct activation of the gastric
glands, because this is the standard in vitro model for the
study of acid secretory mechanisms of the mammalian parietal cells
(Berglindh, 1977; for review see Chew, 1994). As expected, stimulation
of H2 histamine receptors by histamine induced an increase
of acid release within the glands and thereby led to an accumulation of
the weak base [14C]aminopyrine (fig. 7). In contrast, BDF
6143, idazoxan and the putative endogenous ligand at imidazoline sites,
agmatine (Li et al., 1994), failed to induce an increase in
[14C]aminopyrine accumulation and hence obviously did not
induce acid release. Clonidine at the extremely high concentration of 0.1 mM inconsistently increased [14C]aminopyrine
accumulation but was without effect at the lower drug concentrations.
The latter finding is in contrast to the increase in acid release
induced by clonidine and tolazoline from a guinea pig parietal cell
preparation (Houi et al., 1987). This release was ascribed
to activation of imidazoline receptors, although the effect was
potently inhibited by the H2 histamine receptor antagonists
cimetidine, ranitidine and famotidine (Houi et al., 1987),
which, in turn, did not compete with [3H]clonidine for
the nonadrenoceptor [3H]clonidine binding sites in guinea
pig stomach mucosa (Houi et al., 1987). Because only 65% of
the parietal cell preparation prepared by Houi et al. (1987)
consisted of parietal cells, a plausible (though speculative)
explanation for the discrepancy between our present results and those
obtained by Houi et al. might be that the imidazolines
stimulate acid release indirectly, perhaps by inducing the release of
endogenous histamine. This would be consistent with the observation
that the imidazoline-induced release of gastric acid was blocked by
H2 histamine receptor antagonists (Houi et al.,
1987). Taken together, our results make it rather unlikely that the
stimulatory effect of the imidazolines on gastric acid release observed
in vivo and in whole-stomach preparations is due to
activation of nonadrenoceptor [3H]idazoxan or
[3H]clonidine binding sites located on parietal
cells.
For the following reasons, we also investigated whether the potent
-site ligands (+)-3-PPP and DTG influenced gastric acid secretion in
rabbit glands. (1) -Site ligands possess a remarkable affinity for
the [3H]clonidine binding sites in rat stomach
(Molderings et al., 1995). (2) Imidazolines exhibit a
considerable affinity for peripheral 2 sites (Molderings
et al., 1996; present study). (3) 2-Sites are
believed to play some role in the control of gastric function (Harada
et al., 1994). However, both (+)-3-PPP and DTG failed to
increase [14C]aminopyrine accumulation and, accordingly,
did not induce acid secretion by direct stimulation of the parietal
cell.
Conclusion.
Our data provide evidence that, in addition to the
previously identified non-I1/non-I2
[3H]clonidine binding sites (Molderings et
al., 1995), the rat stomach is endowed with
[3H]idazoxan binding sites that belong to the
I2 subclass. In addition, 2 sites are
present in the rat stomach. Our experiments also provide basic evidence
for the existence of I2 and binding sites in the human
stomach. The three different binding sites appear not to act directly
on gastric smooth muscle. They also do not mediate direct
stimulation of gastric acid release from parietal cells.
The technical assistance of Mrs. D. Funccius and Mrs. M. Hartwig
is gratefully acknowledged. We wish to thank the pharmaceutical companies for generous gifts of drugs. The study was supported by the
Deutsche Forschungsgemeinschaft.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Binding
Sites in the Rat and Human Stomach
Top
Abstract
Introduction
;
BDF 6143, 4-chloro-2-(2-imidazolin-2-ylamino)-isoindoline;
(+)-3-PPP, (R)-3-(3-hydroxyphenyl)-N-propylpiperidine;
DTG, 1,2-di-(2-tolyl)guanidine;
IBS, imidazoline binding sites;
CDS, clonidine-displacing substance.
