Novartis Pharma AG, Nervous System Research, Basel,
Switzerland (F.G., S.L., T.L., W.I., D.L., B.S., R.K., S.U., D.S.,
C.P., M.S., P.J.F.); Istituto Mediterraneo di Neuroscienze
"Neuromed," Pozzilli, Italy (V.B., G.B., F.N.); SIBIA Neurosciences
Incorporated, La Jolla, California (M.A.V., S.D.H., E.C.J.); and
University of Catania, Catania, Italy (F.N.)
Group III metabotropic glutamate receptors (mGluRs) are thought to
modulate neurotoxicity of excitatory amino acids, via mechanisms of
presynaptic inhibition, such as regulation of neurotransmitter release.
Here, we describe
(R,S)-4-phosphonophenylglycine (PPG) as a
novel, potent, and selective agonist for group III mGluRs. In
recombinant cell lines expressing the human receptors hmGluR4a, hmGluR6, hmGluR7b, or hmGluR8a, EC50 values for
(R,S)-PPG of 5.2 ± 0.7 µM,
4.7 ± 0.9 µM, 185 ± 42 µM, and 0.2 ± 0.1 µM,
respectively, were measured. The compound showed EC50 and
IC50 values of
200 µM at group I and II hmGluRs and was
inactive at cloned human N-methyl-D-aspartate,
-amino-3-hydroxy-5-methyl-isoxazole-4-propionate, and kainate
receptors (>300 µM). On the other hand, it showed micromolar
affinity for a Ca2+/Cl
-dependent
L-glutamate binding site in rat brain, similar to other phosphono-substituted amino acids like
L-2-amino-4-phosphonobutyrate. In cultured cortical
neurons, (R,S)-PPG provided protection
against a toxic pulse of
N-methyl-D-aspartate (EC50 = 12 µM), which was reversed by the group III mGluR antagonist
(R,S)-
-methylserine-O-phosphate but not by the group II antagonist
(2S)-
-ethylglutamate. Moreover, (R,S)-PPG protected against
N-methyl-D-aspartate- and quinolinic acid-induced striatal lesions in rats and was anticonvulsive in the
maximal electroshock model in mice. In contrast to the group III mGluR
agonists L-2-amino-4-phosphonobutyrate and
L-serine-O-phosphate, (R,S)-PPG showed no proconvulsive effects
(2200 nmol i.c.v.). These data provide novel in vivo evidence for group
III mGluRs as attractive targets for neuroprotective and anticonvulsive
therapy. Also, (R,S)-PPG represents an
attractive tool to analyze the roles of group III mGluRs in nervous
system physiology and pathology.
 |
Introduction |
The
neurotoxicity of excitatory amino acids such as L-glutamate
and some of its analogs, e.g., kainate and
N-methyl-D-aspartic acid (NMDA), is
well established in the central nervous system (Lipton and Rosenberg,
1994
). Analogs of L-glutamate, which have been
investigated in neural systems, share with their parent compound the
-amino acid moiety and a distal, negatively ionizable group (Watkins
et al., 1990
). These structural features are considered essential for
L-glutamate to interact with each member of its large family of ionotropic and metabotropic neurotransmitter receptors (Hollmann and Heinemann, 1994
).
In an effort to discover new agents interfering with the glutamatergic
system, a large panel of phosphono-substituted
-amino acid
derivatives has been generated.
D-2-amino-5-phosphonopentanoic acid (D-AP5),
D-(E)-2-amino-4-methyl-5-phosphono-3-pentenoic
acid (CGP 40116; Fig. 1),
D-(E)-4-(3-phosphonoprop-2-enyl)piperazine-2-carboxylic acid (D-CPPene), and
2-amino-3-(2'-chloro-5-phosphonomethyl-biphenyl-3-yl)-propionic acid)
(SDZ 220-581), for instance, are potent, selective, and competitive
antagonists for NMDA receptors, which constitute one pharmacological
group within the class of ionotropic glutamate receptors (iGluRs).
Those and many related compounds served as tools to elucidate the role
of NMDA receptors in brain disorders, such as neurodegenerative
processes following ischemia and epileptic seizures (Sauer et al.,
1992
; Urwyler et al., 1996a
). On the other hand, many
phosphono-substituted
-amino acids, like
L-2-amino-4-phosphonobutyrate (L-AP4),
L-serine-O-phosphate
(L-SOP), and 4-phosphono-phenylglycine [(R,S)-PPG; Bigge et al., 1989
], were found to
be inactive as NMDA receptor ligands.

View larger version (15K):
[in this window]
[in a new window]
|
Fig. 1.
Chemical structures of phosphono-amino acid
derivatives. (R,S)-PPG,
D-(E)-2-amino-4-methyl-5-phosphono-3-pentenoic
acid (CGP 40116), and two reference agonists for group III mGluRs,
L-AP4 and L-SOP, are depicted. The carbon atom
backbone of each structure is numbered in italics. All four compounds
contain the amino acid function and a distal phosphonic acid. CGP
40116, a classical competitive NMDA receptor antagonist, and
(R,S)-PPG have a considerably larger
distance between the two acidic groups than the glutamic acid analogs
L-AP4 and L-SOP.
|
|
L-AP4 and L-SOP, however, are potent and
selective agonists at a group of metabotropic glutamate receptors
(mGluRs). Eight mGluR subtypes are currently known, which are numbered
according to the order of their molecular discovery, and are subdivided into three distinct groups (Tanabe et al., 1992
; Conn and Pin, 1997
).
Group I mGluRs (mGluR1 and mGluR5) are positively coupled to the
phosphoinositide/Ca2+ cascade. Group II (mGluR2
and mGluR3) and group III (mGluR4, mGluR6, mGluR7, and mGluR8)
receptors are both negatively coupled to adenylate cyclase in
heterologous expression assays. The three groups can be discriminated
pharmacologically with the use of selective agonists.
3,5-Dihydroxyphenylglycine selectively activates group I mGluRs,
whereas 2R,4R-aminopyrrolidine-2,4-dicarboxylate and LY-354740 are examples for group II selective agonists (Conn and
Pin, 1997
; Monn et al., 1997
). L-AP4,
L-SOP, and close analogs are the only selective
agonists known for group III mGluRs, with low micromolar potency
(EC50 values, 0.1-7 µM) at mGluR4, mGluR6, and
mGluR8; mGluR7 can only be activated at concentrations higher than 100 µM (Okamoto et al., 1994
; Johansen et al., 1995
; Conn and Pin, 1997
;
Flor et al., 1997
). Group III (and group II) mGluRs are thought to
mediate presynaptic depression of glutamatergic synaptic potentials in
several brain areas, most likely via inhibition of voltage-gated
calcium entry and regulation of glutamate release (Trombley and
Westbrook, 1992
; Conn and Pin, 1997
). Moreover, selective activation of
group III mGluRs results in neuroprotection in vitro; agonists like
L-AP4 and L-SOP promote
survival of rat cerebellar granule cells and protect cultured cortical
and cerebellar neurons against toxic insults, such as prolonged
-amyloid peptide exposure, transient iGluR activation, or mechanical
damage (Graham and Burgoyne, 1994
; Copani et al., 1995
; Bruno et al.,
1996
; Faden et al., 1997
). In contrast to the findings with NMDA
receptor antagonists, in vivo neuroprotection with group III mGluR
agonists has not yet been reported to our knowledge.
Both, proconvulsive and anticonvulsive effects of group III mGluR
agonists have been observed, depending not only on the animal model
used but also on timing and dosage of the drug treatment (e.g., Graham
and Burgoyne, 1994
; Abdul-Ghani et al., 1997
; Ghauri et al., 1996
; Tang
et al., 1997
).
Here, we report in vitro and in vivo neuroprotective actions of
(R,S)-PPG, a compound with structural similarity
to competitive NMDA receptor antagonists and group III mGluR agonists
(Fig. 1). (R,S)-PPG was also tested for
anticonvulsive properties in the maximal electroshock-induced
convulsion model in mice, in comparison with
L-AP4 and L-SOP. In an
effort to characterize activity and selectivity of
(R,S)-PPG at the molecular level, the compound was tested for interaction with all eight mGluR subtypes, with a
representative selection of recombinant human iGluRs and with a
Ca2+/Cl
-dependent
L-glutamate binding site of rat brain. Group III
mGluRs as attractive drug targets for the treatment of neurological
disorders, such as epilepsy and Huntington's disease, will be discussed.
 |
Experimental Procedures |
Chemical Synthesis of
(R,S)-PPG.
(R,S)-PPG was synthetized in four steps starting
from 4-hydroxybenzaldehyde using a different synthetic pathway than the
one described (Bigge et al., 1989
). The starting material was first esterified with trifluoromethanesulfonicanhydride. The
trifluoromethanesulfonate ester was then converted to the corresponding
phosphonate ester using a palladium-catalyzed coupling. Conversion of
the aldehyde to the amino nitrile and subsequent hydrolysis performed
in concentrated HCl gave (R,S)-PPG with an
overall yield of about 30%. Further details of the synthesis will be
published elsewhere.
Cloning of hmGluR8a cDNA.
The sequence encoding the human
metabotropic glutamate receptor subtype 8a (hmGluR8a) was constructed
from clones obtained by library screening in combination with
polymerase chain reaction (PCR).
Library Screening.
Five × 105
plaques each of two human cDNA libraries from whole adult brain (in
gt10; Clontech, Palo Alto, CA) and adult hippocampus (in
ZAPII;
Stratagene, Heidelberg, Germany) were screened with 5' and 3' probes
from the rat mGluR4 sequence (Tanabe et al., 1992
) as described
previously (Laurie et al., 1997
). After a second round of screening,
individual cDNA inserts were rescued into Bluescript SK(
) phagemids
(Stratagene) by in vitro (
gt10) or in vivo (
ZAPII) excision. cDNA
inserts were characterized by restriction enzyme mapping and DNA
sequencing (ABI systems, Langen, Germany). Two nonoverlapping clones
were identified as homologous to portions of the mouse mGluR8 sequence
(Duvoisin et al., 1995
): HMGBr7 (homologous to bases 576-799 of mouse
mGluR8) and HMGHi7 (homologous to bases 2059-2830 of mouse mGluR8).
PCR.
The 5' end of the hmGluR8 coding sequence was amplified
from human retinal cDNA. Thermocycling conditions were: 94°C for 1 min, 45°C for 1 min, 72°C for 1 min, 10 cycles, then 94°C for 30 s, 55°C for 30 s, 72°C for 1 min, 38 cycles, using
"Expand-High Fidelity" polymerase (Boehringer Mannheim, Mannheim,
Germany). PCR oligos were
5'-GTCGCTGACTGCAATACCACCTGCGGAGAAAATG-3' [sense oligo from
mouse mGluR8 sequence (Duvoisin et al., 1995
); translation initiation
codon underlined] and 5'-CAACTATCTGAGCCAATCCAG-3'. The resultant
900-bp amplicon, "PCR5p8," was subcloned into the PCR cloning
vector pCRII (Invitrogen, San Diego, CA).
The missing sequence between clones HMGBr7 and HMGHi7 was obtained by
PCR from human retinal cDNA, and the resultant 1570-bp amplicon was
also subcloned into the vector pCRII. Thermocycling conditions were
94°C for 1 min, 60°C for 1 min, 72°C for 2 min, 38 cycles, using
Expand-High Fidelity polymerase (Boehringer Mannheim). PCR oligos were
5'-GACTCCTACCAAGCCCAAGCCATG-3' and 5'-CGCTGCTCTCCATAGTCAATGATG-3'. This
intervening sequence, "PCRint8", was digested into three fragments
with BamHI and the largest (1214-bp) fragment,
"3'PCRint8" ligated to HMGHi7 by a common BamHI site,
forming 3'PCRint8-HMGHi7. A parallel digest of PCRint8 with
MunI released the 420-bp fragment "5'PCRint8".
The full-length hmGluR8a sequence was constructed by ligating
3'PCRint8-HMGHi7 to 5'PCRint8 via a common MunI site,
ligating the product to PCR5p8 via an NcoI site and finally
ligating the whole sequence into the expression vector pCIneo
(Clontech) using XbaI and SalI sites. The coding
region of the assembled hmGluR8a clone (hmGluR8a.pCIneo) was sequenced
on both strands. In comparison with the very recently published
hmGluR8a sequence (Wu et al., 1998
), we find only one amino acid
difference: Asp-768 to Ile, which is encoded by the
library-derived clone HMGHi7.
Stable Expression of hmGluR8a in HEK293 Cells.
The construct
hmGluR8a.pCIneo was linearized by digestion with Asp-700. One microgram
of the linearized DNA was used to transfect 107
HEK293 cells. Selection for stable integration was made by addition of
0.8 mg/ml G418 (Life Technologies, Basel, Switzerland) to the medium
(minimal essential medium with 2 mM L-glutamine
supplemented with 10% dialyzed fetal calf serum; Life Technologies),
and 30 G418-resistant clonal cell lines were isolated as described
previously (Laurie et al., 1995
). Further selection was performed by
measuring the glutamate (0.1 mM)-induced depression of forskolin (10 µM)-elevated cAMP accumulation in cells grown in collagen-coated
wells. Responses ranged from no depression to about 85% depression.
Using this approach, two cell lines, HEK-hmGluR8a-2 and
HEK-hmGluR8a-20, were identified as giving consistently good responses
for up to at least 20 passages of subculturing.
Stable Mammalian Cell Lines for Cloned mGluR1 to mGluR7 and
Ionotropic Glutamate Receptors.
Generation, culture, and
pharmacological characterization of stable cell lines for hmGluR1b, -2, -4a, -5a, -6, -7b, rat mGluR3, hNMDAR1A/2A, hNMDAR1A/2B, hGluR3i, and
hGluR6 have been described recently (Knöpfel et al., 1995
; Laurie
et al., 1995
, 1997
; Daggett et al., 1996
; Varney et al., 1996
, 1998
;
Flor et al., 1997
and references therein; Lin et al., 1997
).
In Vitro Pharmacological Assays for Cloned Glutamate
Receptors.
Measurements of cyclic AMP accumulation (Flor et al.,
1997
), inositol monophosphate formation (Knöpfel et al., 1995
),
and cytoplasmic calcium elevation (Flor et al., 1996
) were performed as
described previously.
L-[3H]Glutamate-Binding Assay for
mGluR3.
HEK293 cells stably transfected with the cDNA encoding rat
mGluR3 were cultured and harvested as described previously (Laurie et
al., 1995
). Membranes from these cells were washed by five cycles of
centrifugation (10 min at 50,000g, 4°C) and resuspension in assay buffer before being frozen and stored at
80°C until their
use in the binding experiments. After thawing, membranes were washed
five times by centrifugation and resuspension in ice-cold assay buffer
as above. The
L-[3H]glutamate-binding
assay was performed in 0.6 ml of 50 mM Tris-HCl buffer (pH 7.5 at
0°C) containing an aliquot of the membrane suspension (about 50 µg
of protein), 5 nM
L-[3H]glutamate (NEN,
Regensdorf, Switzerland), 2.5 mM CaCl2, and the
test compounds at the appropriate concentrations. Nonspecific binding
was determined by including 0.5 mM unlabeled
L-glutamate. The samples were incubated at 0°C
for 4 h before bound and free radioligand were separated by
centrifugation at 4°C for 20 min at approximately 10,000g.
The supernatant was decanted, and the pellets were quickly and
superficially rinsed with ice-cold assay buffer and then added to
scintillation fluid containing tissue solubilizer (Solvable; NEN).
After solubilization at 50°C overnight, the radioactivity was
measured by liquid scintillation counting. Competition curves were
analyzed, and IC50 values were determined by
nonlinear curve fitting using the program GraphPad Prism (GraphPad Software, Inc., San Diego, CA).
Ca2+/Cl
-Dependent
L-[3H]Glutamate Binding to Rat Brain
Membranes.
This assay was essentially performed as described
previously (Urwyler et al., 1996b
). In brief, the assay mixture (in a
final volume of 1.1 ml) contained 50 mM Tris-HCl (pH 7.4), 2.5 mM
CaCl2, extensively washed rat hippocampal
membranes (freshly prepared on the day of the experiment) corresponding
to approximately 3 mg of original tissue (wet weight), 5 nM
L-[3H]glutamate, and the test
compounds at the desired concentrations. Nonspecific binding was
defined with 0.2 mM DL-AP7. The samples were incubated for
25 min at 37°C before bound and free radioligand were separated by
centrifugation at 12,000g for 4 min. The pellets were
quickly and superficially rinsed with 100 µl of ice-cold incubation
buffer and then added to scintillation fluid containing tissue
solubilizer (Solvable; NEN). After solubilization at 50°C overnight,
the radioactivity was measured by liquid scintillation counting.
Preparation of Cultured Cortical Cells and Examination of NMDA
Toxicity.
Mixed cultures of cortical cells were prepared from
fetal mice (14-16 days of gestation), as described (Bruno et al.,
1996
), and used 13 to 14 days after plating. Cultures were exposed to 100 µM NMDA for 10 min at room temperature in a HEPES-buffered salt
solution. After extensive washing, cultures were incubated for 18 to
24 h at 37°C in minimal essential medium-Eagle's buffer (Life Technologies) supplemented with 25 mM
NaHCO3 and 21 mM glucose. Neuronal toxicity was
examined by phase-contrast microscopy and quantitated after staining
with trypan blue. Stained neurones were counted from three random
fields per well. Lactate dehydrogenase release into the medium was also
measured as described previously (Bruno et al., 1996
).
Examination of Neuronal Toxicity after Intrastriatal Infusion
with NMDA and Quinolinic Acid.
Male Sprague-Dawley rats (250-300
g) were anesthetized with pentobarbital (50 mg/kg, i.p.) and infused
with NMDA (100 nmol/0.5 µl/2 min) or NMDA + group III mGluR agonists
(balanced to neutral pH) in the left caudate nucleus, at +2.0 mm AP,
2.6 mm L, and 5 mm V, according to the Pellegrino and Cushman atlas.
The injection was repeated at a second site (1 mm posterior to the
first site) to increase the extent of striatal toxicity. Seven days
later, animals were sacrificed, and neuronal damage was assessed by
either histological analysis or measurements of striatal glutamate
decarboxylase (GAD) activity. For histological analysis, the brains
were rapidly frozen in isopentane at
40°C and then stored at 0°C.
Twenty-micrometer cryostat sections were Nissl-stained and examined in
light microscopy. For measurements of GAD activity, the corpus striatum
was dissected bilaterally and homogenized in 5 mM imidazol buffer
containing 0.2% Triton X-100 and 10 mM dithiothreitol. An aliquot of
the homogenate was incubated in 400 µl of 10 mM phosphate buffer (pH 7.0) containing 10 mM 2-mercaptoethanol, 0.02 mM pyridoxalphosphate, and 1 µCi of L-[3H]glutamate
(Amersham Intl., Buckinghamshire, UK; specific activity 46 Ci/mmol) for
1 h at 37°C; the reaction was stopped with 15 µl of ice-cold
11.8 N HClO4. After centrifugation in a microfuge at maximal speed, 10 µl of the supernatant were diluted with 0.01 N
HCl and derivatized with O-phtalaldehyde and mercaptoethanol for 1 min at room temperature before injection into HPLC. The HPLC
apparatus consisted of a programmable solvent module 126 (Beckman
Instruments, Inc., Fullerton, CA), an analytical
C18 reversed-phase column kept at 30°C
(Ultrasphere ODS 5 µm spherical, 80 Å pore, 2 mm × 15 cm;
Beckman Instruments Inc.), and an RF-551 spectrofluorimetric
detector (Shimadzu Corp., Tokyo, Japan). Excitation and emission were
set at 360 and 450 nm, respectively. The mobile phase consisted
of 50 mM sodium phosphate, 10% methanol, pH 7.2 (A), and 50 mM
sodium phosphate, 70% methanol, pH 7.2 (B). After 8 min of isocratic
conditions with 98% (A) and 2% (B), (B) was increased up to 40%
within 30 min and then to 98% within 1 min and then maintained at 98%
for 11 min before returning to the initial conditions. The
radioactivity coeluting with
-aminobutyric acid (GABA) was collected
and counted by scintillation spectrometry. Protein concentrations in
the original samples were determined by using a commercially available
kit (Bio-Rad, Richmond, CA).
Intrastriatal injection of quinolinic acid and evaluation of
excitotoxic neurodegeneration by magnetic resonance imaging was performed as previously described (Sauer et al., 1992
).
Maximal Electroshock Test (MES).
Experiments were conducted
on 19- to 25-g male mice [Tif:MAGf (SpF)] at 21-22°C. Generalized
tonic-clonic convulsions of the hind extremities were induced by
passing alternating electrical currents of 50 Hz and 18 mA through
corneal electrodes (for reference, see Kupferberg and Schmutz, 1997
).
L-AP4, L-SOP, and
(R,S)-PPG were dissolved in 0.9% saline, the pH
corrected to 7.0 and administered i.c.v., i.p., or i.v. with
pretreatment times of 15 or 30 min. Five to 10 animals per dose were
used; ED50 values were calculated on the basis of
at least five doses, and each experiment was done at least twice. The
number of animals protected from tonic hind limb extension seizure and
the duration of hind limb tonus were determined in each dose group.
Statistics.
Significant differences were estimated using the
two-tailed Dunnett's t test by comparing test-drug groups
with the control group. Values of 2P < 0.05 were
considered as statistically significant.
Materials.
Molecular biology reagents and enzymes were
purchased from Amersham, Bio-Rad, Boehringer Mannheim, Invitrogen, New
England Biolabs, and Stratagene. Tissue culture reagents were from Life Technologies and Sigma. Forskolin, 3-isobutyl-1-methylxanthine, and
L-SOP were obtained from Sigma. All other reference
agonists and antagonists were purchased from Tocris (Anawa Trading SA, Zürich, Switzerland). CGP 40116 was synthesized within Novartis Pharma AG by Dr. Roland Heckendorn. All other chemicals were of reagent
grade and were obtained from Fluka (Buchs, Switzerland), Merck
(Darmstadt, Germany), Serva (Heidelberg, Germany), or Sigma.
 |
Results |
Profiling of (R,S)-PPG against All
Eight Cloned mGluR Subtypes.
As shown in Fig.
2, (R,S)-PPG most
potently inhibited forskolin-stimulated cAMP accumulation in
recombinant cells stably expressing hmGluR8a, producing 70 to 80%
inhibition at a maximally effective concentration of 100 µM and an
EC50 value of 0.2 ± 0.1 µM (mean ± S.E.M.). Concentration-response curves for agonist activity at the
other known group III mGluR subtypes were measured, and EC50 values of 5.2 ± 0.7 µM, 4.7 ± 0.9 µM, and 185 ± 42 µM (means ± S.E.M.) were found for
hmGluR4a, hmGluR6, and hmGluR7b, respectively (Fig. 2; Table
1). In contrast,
(R,S)-PPG showed no significant inhibition of
forskolin-stimulated cAMP accumulation in untransfected CHO-K1 and
HEK293 cells when tested at 300 µM and 100 µM, respectively (not
shown). Different extents of maximal inhibition of cAMP formation by
(R,S)-PPG were observed in the four cell lines
expressing recombinant group III mGluRs, ranging from 45 to 80%
inhibition. The Hill coefficients for the interaction of
(R,S)-PPG with the four group III mGluRs were
between 1.5 and 2.5 (Table 1).

View larger version (20K):
[in this window]
[in a new window]
|
Fig. 2.
Agonist activity of
(R,S)-PPG at group III hmGluRs.
Concentration-response curves for inhibition of forskolin (10 µM)-stimulated cAMP accumulation by
(R,S)-PPG at CHO cells expressing
hmGluR4a, hmGluR6, and hmGluR7b as well as HEK cells expressing
hmGluR8a. Each data point represents mean value ± S.E.M. from at
least three independent experiments (n 6).
Sigmoidal curves were fit using the GraphPad Prism program (GraphPad
Software, Inc., San Diego, CA).
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 1
Activity of (R,S)-PPG on metabotropic and
ionotropic glutamate receptors expressed in recombinant mammalian cells
Concentration-response curve fitting and statistical analysis were done
using the GraphPad Prism program (GraphPad Software, Inc.).
EC50 values are given ± S.E.M. (n 3);
nH, Hill coefficient.
|
|
To address specificity of (R,S)-PPG, the compound
was tested for agonist and antagonist activity at all known group I,
II, and III mGluR subtypes. In recombinant cells expressing hmGluR1b or
hmGluR5a, quisqualate produced 5- to 15-fold stimulation of phosphoinositide hydrolysis at concentrations of 20 µM and 0.3 µM, which are approximate EC80 concentrations
for these two clonal lines. (R,S)-PPG, tested at
500 µM, neither stimulated phosphoinositide hydrolysis on its own nor
reversed the quisqualate effects in either hmGluR1b- or
hmGluR5a-expressing cells (Fig. 3A).
Moreover, at concentrations of 10 µM and 100 µM,
(R,S)-PPG was also found inactive at hmGluR1b and
hmGluR5a (data not shown).

View larger version (35K):
[in this window]
[in a new window]
|
Fig. 3.
Activity of (R,S)-PPG
at group I, II, and III hmGluRs. A, inositol monophosphate formation of
CHO cells expressing hmGluR1b and Ltk cells expressing
hmGluR5a. Exposure to buffer was taken as control and set to 1.0; for
submaximal and maximal stimulation of hmGluR1b, 20 µM quisqualate and
1000 µM quisqualate were used, respectively; 0.3 µM quisqualate
gave submaximal stimulation of hmGluR5a, and 10 µM quisqualate showed
the maximal effect. To test for agonist activity at group I mGluRs,
(R,S)-PPG was applied at 500 µM to
CHO-hmGluR1b and Ltk -hmGluR5a cells. Antagonist activity
of 500 µM (R,S)-PPG at hmGluR1b and
hmGluR5a was determined by coapplication of the submaximal
concentrations of quisqualate (around EC80). Mean
values ± S.E.M. from at least two independent experiments are
shown (n 4). B, inhibition of
forskolin-stimulated cAMP accumulation in CHO cells expressing either
hmGluR2 or hmGluR4a. Forskolin (10 µM) stimulated cAMP formation
about 40-fold (taken as control). All values are given as fraction of
control. The effect of forskolin is inhibited by 30 µM
(1S,3R)-ACPD in hmGluR2-expressing cells
and by 1 µM L-AP4 in hmGluR4a-expressing cells. These
submaximal agonist concentrations represent approximately
EC80. To test for agonist activity,
(R,S)-PPG was applied to
forskolin-stimulated CHO-hmGluR2 and CHO-hmGluR4a cells. Antagonist
activity of (R,S)-PPG at hmGluR2 and
hmGluR4a was determined by coapplication of submaximal concentrations
of (1S,3R)-ACPD and L-AP4,
respectively. Columns represent mean values ± S.E.M. of at least
two independent experiments (n 5). Asterisks
indicate statistically significant antagonism or agonist activity of
(R,S)-PPG (2P < .01;
Dunnett's t test). C, inhibition of forskolin (10 µM)-stimulated cAMP accumulation in CHO cells expressing hmGluR7b and
HEK cells expressing hmGluR8a. Forskolin-stimulated cAMP formation was
taken as control. To obtain submaximal depression of the
forskolin-stimulated cAMP accumulation, 500 µM L-AP4 and
1 µM L-AP4 were used for hmGluR7b and hmGluR8a,
respectively (this represents about EC80).
(R,S)-PPG was tested for agonist activity
as well as antagonist activity of submaximal L-AP4.
Asterisks indicate statistically significant agonist activity of
(R,S)-PPG (2P < .01;
Dunnett's t test, n 4).
|
|
In Chinese hamster ovary (CHO) cells stably expressing hmGluR2,
1-aminocyclopentane-1S,3R-dicarboxylic acid
[(1S,3R)-ACPD], at 30 µM, depressed
forskolin-stimulated cAMP accumulation by approximately 65%, which
represents about 80% of the maximal response (Fig. 3B). Five hundred
micromolar (R,S)-PPG showed no significant depression of cAMP when applied alone but showed 52% reversal of cAMP
depression when coapplied with 30 µM
(1S,3R)-ACPD (Fig. 3B), indicating antagonist
activity at hmGluR2. At 300 µM, the antagonist activity of
(R,S)-PPG was reduced to 45%; at 100 µM, no
significant antagonist activity of (R,S)-PPG at
hmGluR2 was observed (2P > .05, Dunnett's
t test, n = 8). At recombinant rat mGluR3
expressed in human embryonic kidney (HEK) cells,
(R,S)-PPG as well as L-AP4
had only very weak affinity in an
L-[3H]glutamate binding
assay. (R,S)-PPG, L-AP4,
and L-glutamate showed around 50% inhibition of
L-[3H]glutamate binding
at 200 µM, 300 µM, and 200 nM, respectively (data not shown).
As shown in Fig. 3, B and C, and Table 1, inhibition of
forskolin-stimulated cAMP formation in recombinant cells expressing hmGluR4a, hmGluR6, hmGluR7b, or hmGluR8a by
(R,S)-PPG was at least as efficacious as with
L-AP4, applied at EC80.
Coapplication of L-AP4 and
(R,S)-PPG (300 or 1000 µM) did not reverse the
L-AP4 response.
Characterization of (R,S)-PPG at
Ionotropic Glutamate Receptors.
(R,S)-PPG
was originally designed and characterized for NMDA-receptor binding
(Bigge et al., 1989
). However, the compound was found to be inactive up
to 100 µM in the
[3H]-3-(2-carboxypiperazin-4-yl)-propyl-1-phosphoric
acid receptor binding assay on rat cortical brain membranes,
indicating that (R,S)-PPG does not bind to the
L-glutamate site of native rat NMDA receptors.
Here, we extend the characterization by employing stable cell lines
transfected with the hNMDAR1A/2B subunit combination (Varney et al.,
1996
) to address functional agonist and antagonist activity of
(R,S)-PPG at cloned human NMDA receptors. In this cell line, L-glutamate (applied at 1 µM)
induced a rise in intracellular calcium concentration.
(R,S)-PPG neither induced a calcium rise when
applied at 300 µM nor antagonized the
L-glutamate-induced calcium signal (Fig.
4). In contrast, the competitive
NMDA-receptor antagonist CGP 40116 concentration-dependently prevented
the L-glutamate-evoked rise in intracellular
calcium (Fig. 4). Similar lack of activity of 300 µM
(R,S)-PPG was observed in calcium assays for
cloned hNMDAR1A/2A (data not shown, Table 1).

View larger version (27K):
[in this window]
[in a new window]
|
Fig. 4.
Lack of activity at NMDA-type ionotropic glutamate
receptors by (R,S)-PPG. Elevation of
cytoplasmic calcium ion concentrations,
[Ca2+]i, in Ltk cells
expressing cloned human NMDA receptors (subunit combination 1A/2B). A,
increase in the fluorescent response ratio
F340/F380 corresponds to an elevation in
[Ca2+]i. Three single calcium measurements
are shown: application of 1 µM L-glutamate
(L-Glu) alone, coapplication of L-Glu (1 µM)
and (R,S)-PPG (300 µM), and 300 µM
(R,S)-PPG alone. The duration of the
applications is indicated with a column. B, each column represents
calcium signals (mean values ± S.E.M.) of at least six
measurements of two independent experiments. The response of 1 µM
L-Glu was taken as control and set to 100%. Asterisks
indicate statistically significant antagonist activity
(2P < .01; Dunnett's t
test).
|
|
Moreover, stable cell lines expressing human AMPA(GluR3) or
kainate(GluR6) receptors were utilized in the same assay (Daggett et
al., 1996
; Varney et al., 1998
). Application of L-glutamate resulted in robust increases of cytoplasmic calcium, whereas
(R,S)-PPG (300 µM) did neither evoke calcium
signals on its own nor antagonize L-glutamate in
either of the two cell lines (data not shown; Table 1).
Ca2+/Cl
-Dependent
L-[3H]Glutamate Binding to Rat Brain
Membranes.
L-AP4, L-SOP, and
(R,S)-PPG displaced
L-[3H]glutamate binding
at rat hippocampal membranes, measured in the presence of
CaCl2, with IC50 values of
0.5 µM, 3.1 µM, and 1.9 µM, respectively (Table 2). Thus, L-AP4 was
the most potent among the three compounds, having a 4-fold higher
affinity than (R,S)-PPG, which was about equipotent with L-SOP.
View this table:
[in this window]
[in a new window]
|
TABLE 2
Activity of group III mGluR agonists at a
Ca2+/CI -dependent L-glutamate binding
site of rat brain
The affinities of L-AP4, L-SOP, and
(R,S)-PPG as displacers of
L-[3H]glutamate binding to rat hippocampal
membranes were measured in the presence of calcium chloride as
described (Urwyler et al., 1997b ). The results shown are means ± S.E.M. from three independent experiments, each performed in
triplicate; nH, Hill coefficient.
|
|
Neuroprotective Activity of (R,S)-PPG
in Neuronal Culture.
In primary cultures of cortical neurons
cocultured with glia, (R,S)-PPG was highly
neuroprotective when applied during the NMDA-induced excitotoxic pulse
(10 min). The action of (R,S)-PPG was
concentration-dependent, with an apparent EC50
value of 12 µM (Fig. 5A). Maximally
effective concentrations of (R,S)-PPG rescued
slightly more than 50% of the neuronal population from excitotoxic
degeneration (Fig. 5).
(R,S)-
-Methylserine-O-phosphate (MSOP), a preferential antagonist of group III mGluRs (Thomas et al.,
1996
), prevented the neuroprotective activity of
(R,S)-PPG. In contrast, the selective group II
mGluR antagonist (2S)-
-ethylglutamic acid (EGlu) (Jane et
al., 1996
; Thomas et al., 1996
) did not affect the action of
(R,S)-PPG (Fig. 5B). Neither
(R,S)-PPG nor MSOP or EGlu had any effect on
neuronal viability when applied to the cultures in the absence of NMDA
(not shown).

View larger version (22K):
[in this window]
[in a new window]
|
Fig. 5.
Protection of cultured cortical neurons against
NMDA-induced excitotoxic degeneration by
(R,S)-PPG. A, concentration-protection
relationship of (R,S)-PPG against NMDA
(100 µM)-induced degeneration in mixed cortical cultures assessed by
extracellular lactate dehydrogenase activity (measured as mOD/min). The
value measured with 100 µM NMDA was taken as 100%. To determine
EC50 and Hill coefficient (nH), a sigmoidal curve was fit
using the GraphPad Prism program (GraphPad Software, Inc.). B, neuronal
degeneration in mixed cortical cultures was induced by 100 µM NMDA
(resulting in approximately 85% of maximal NMDA toxicity) and assessed
by extracellular lactate dehydrogenase activity (measured as mOD/min,
set to 100%). Statistically significant neuroprotection by
(R,S)-PPG is indicated by asterisks
(2P < .01; Dunnett's t test,
n 4, at least two independent experiments).
(R,S)-PPG neuroprotection can be
antagonized by MSOP but not by EGlu.
|
|
Protection by (R,S)-PPG against NMDA-
and Quinolinic Acid-Induced Striatal Lesions.
Intrastriatal
infusion of various ionotropic glutamate receptor agonists results in
neurochemical and neuropathological changes resembling Huntington's
disease (DiFiglia, 1990
). First, we have used infusion of 100 nmol of
NMDA which induces an extended area of necrosis characterized by
neuronal loss, reactive gliosis, edema, and neuronal pyknosis (Fig.
6A). Neuronal damage was visible across
the extension of the caudate nucleus, up to 3 mm posterior to the
injection site. Coinfusion of 250 nmol of
(R,S)-PPG with NMDA resulted in efficacious
neuroprotection against excitotoxic neuronal damage; particularly in
the more lateral parts of the caudate nucleus, drastically reduced
neuronal loss and pyknosis as well as less edema formation were seen
(Fig. 6B). To quantitate the protective effect of
(R,S)-PPG, we have measured striatal GAD activity
as a biochemical marker of viable GABAergic neurones (Fig.
7), which has been widely used as
post-mortem assessment of lesion size (e.g., Urwyler et al., 1996a
).
NMDA infusion led to a 45 to 50% decrease in GAD activity as compared
with the respective contralateral side. No reduction in GAD activity
was observed in animals coinfused with NMDA plus
(R,S)-PPG. The protective activity of
(R,S)-PPG against NMDA toxicity was mimicked by
L-AP4 (50 or 250 nmol, Fig. 7).

View larger version (126K):
[in this window]
[in a new window]
|
Fig. 6.
(R,S)-PPG protects
against NMDA-induced striatal lesions in vivo. Nissl staining of
20-µm cryostat sections from rat caudate nucleus lateral to the site
of NMDA (100 nmol) injection (A); note the large area of extensive
necrotic damage at the left side of the micrograph (white arrowheads)
and widespread neuronal pyknosis; black arrowheads point at two
pyknotic neurones. When (R,S)-PPG (250 nmol) is coinjected with NMDA (B), necrotic damage and neuronal
pyknosis is drastically reduced; open arrowheads indicate two healthy
neurons. Also, note the higher neuronal density in B compared with
A.
|
|

View larger version (26K):
[in this window]
[in a new window]
|
Fig. 7.
Quantitation of in vivo neuroprotection by
(R,S)-PPG. Degeneration of GABAergic
neurones in rat corpus striatum was induced by NMDA (100 nmol)
injection and assessed by measuring the GAD activity. GAD activity of
rat corpus striatum contralateral to the site of NMDA injection was
taken as control and set to 100%. Each column represents mean
values ± S.E.M. Statistically significant neuroprotection by
(R,S)-PPG and L-AP4 is
indicated by asterisks (2P < .01, Dunnett's
t test; the number of tested animals in each group is
indicated within the columns; for each group, at least two independent
experiments were performed).
|
|
Furthermore, we tested (R,S)-PPG for protection
against intrastriatal infusion of quinolinic acid (200 nmol), which
produced larger lesions than NMDA. Lesion size was quantitated by
magnetic resonance imaging as described (Sauer et al., 1992
). Compared with the control group receiving 200 nmol of quinolinic acid (11 animals treated), coinfusion of (R,S)-PPG (250 nmol) showed significant protection (2P < .005, Dunnett's t test) with a reduction of the lesion size by
58.4% ± 13% (mean ± S.E.M., n = 13 animals).
Protection by (R,S)-PPG against
MES-Induced Convulsions in Mice.
MES is commonly used as a basic
in vivo test for anticonvulsive compounds (Kupferberg and Schmutz, 1997
and references therein). In this test, with i.c.v. injections of
L-AP4 and L-SOP, a pretreatment time of 15 min,
and doses between 60 and 220 nmol, no anticonvulsive effects were seen.
When doses were increased to approximately 2000 nmol, L-AP4
and L-SOP induced clonic/clonic-tonic seizures at 5 to 10 min after drug administration in 40 to 60% of the treated animals
(Table 3). In contrast,
(R,S)-PPG up to 2200 nmol did not show any
proconvulsant effect. Moreover, (R,S)-PPG when
applied at 173 nmol (i.c.v.) produced 100% protection against MES with an ED50 value of 78 nmol (Table 3). At doses
above 2000 nmol (i.c.v.), all three compounds were lethal in 20 to 60%
of the animals. (R,S)-PPG given i.p. or i.v. was
inactive against MES-induced convulsions up to 100 mg/kg and 10 mg/kg,
respectively.
View this table:
[in this window]
[in a new window]
|
TABLE 3
Activity of L-AP4, L-SOP, and (R,S)-PPG in the
maximal electroshock test (mouse)
n.d., ED50 values were not determined. The results shown are
derived from at least two independent experiments, each performed with
at least five animals per group.
|
|
 |
Discussion |
In Vitro Pharmacology.
Recent cDNA cloning and recombinant
expression of the heterogeneous family of G protein-coupled
(metabotropic) glutamate receptors has created a large gap between the
molecular knowledge of mGluR subtypes and the understanding of their
role in brain function and dysfunction. Thus, the discovery of
subtype-selective compounds and their testing in experimental models
for nervous system physiology and pathology has become increasingly important.
In the present study, using 12 different cell lines stably expressing
cloned ionotropic and metabotropic glutamate receptor subtypes, we have
characterized the pharmacological profile of (R,S)-PPG. This compound behaved as a potent
agonist at hmGluR8a (EC50 = 200 nM), hmGluR6
(EC50 = 4.7 µM), and hmGluR4a
(EC50 = 5.2 µM), with no appreciable activity
(up to 200 µM) at hmGluR1, hmGluR2, rat mGluR3, hmGluR5, and
hNMDAR1A/2B, hNMDAR1A/2A, AMPAR(hGluR3), and kainateR(hGluR6).
(R,S)-PPG therefore exhibited approximately
1000-fold selectivity for hmGluR8 and
40-fold selectivity for
hmGluR4a and hmGluR6 versus group I/II mGluRs and all tested iGluRs.
(R,S)-PPG activated hmGluR7b only at high
micromolar concentrations (EC50 = 185 µM), which is only slightly more potent than the compound's weak antagonist activity at hmGluR2 and its activity against glutamate binding at rat
mGluR3 (Table 1). In addition to its potent activity at group III
mGluRs, (R,S)-PPG, as well as
L-AP4 and L-SOP, displayed micromolar affinity for a
Ca2+/Cl
-dependent
L-[3H]glutamate binding
site in rat brain (Fagg et al., 1982
; Urwyler et al., 1996b
). The
functional role of this binding site was proposed to be that of a
glutamate transporter, but alternative explanations are also possible
(see Urwyler et al., 1996b
and references therein for detailed
discussion). The affinity of (R,S)-PPG at this
site was found to be 4-fold lower than the affinity of
L-AP4. Whether this difference is reflected at
the level of endogenous glutamate accumulation, and thus could
contribute to explain the discrepancy we have found between
L-AP4 and (R,S)-PPG in the
MES model in mice (see below), is currently unclear, especially because
the explanation cannot be extended to L-SOP.
In summary, (R,S)-PPG represents a novel group
III mGluR agonist, with in vitro pharmacological properties
indistinguishable from the current agonists
L-SOP, L-AP4, and
cyclopropyl-AP4 (Johansen et al., 1995
; Okamoto et al., 1994
; Conn and
Pin, 1997
; Flor et al., 1997
;). Interestingly, however,
(R,S)-PPG differs structurally from all of these
compounds but shares close similarity with classical competitive NMDA
receptor antagonists like CGP 40116, as the distance between the amino
acid moiety and the phosphonate group is comparable. Because this
distance is substantially smaller in L-AP4 and
L-SOP, the potency and selectivity of
(R,S)-PPG for group III mGluRs is quite
surprising and suggests a particular mode of binding of
(R,S)-PPG, involving the phenyl spacer between
the
-amino acid and phosphonic acid moieties.
-Methylphosphonophenyl glycine, a mixed antagonist of group II
and III mGluRs (Bedingfield et al., 1996
), is structurally the most
closely related compound and differs from
(R,S)-PPG only by a methyl group in the
-position. Thus, replacement of that group with the hydrogen of
(R,S)-PPG has changed the antagonistic properties
into selective agonist activity. Similar properties have been shown for
linear L-glutamate analogs and cyclopropylglycine derivatives (Jane et al., 1996
; Thomas et al., 1996
), where the introduction of a methyl group at the
-position changed agonists into antagonists.
Neuroprotection and Anticonvulsive Actions Mediated by
(R,S)-PPG.
Neuroprotective effects
of L-AP4 and L-SOP observed in several in vitro
paradigms (see Introduction) prompted us to examine (R,S)-PPG, as a structurally different but
pharmacologically indistinguishable compound, in the model of
NMDA-induced degeneration of mouse cortical neurons cocultured with
glia (Bruno et al., 1996
). Here, we found (R,S)-PPG highly neuroprotective, and its action
was antagonized by MSOP, which is a group III mGluR antagonist (Thomas
et al., 1996
), but not by the group II mGluR antagonist EGlu. These
results therefore strengthen the suggestion that activation of group
III mGluRs is neuroprotective in vitro (e.g., Copani et al., 1995
; Bruno et al., 1996
; Faden et al., 1997
).
To investigate neuroprotective effects of
(R,S)-PPG also in vivo, we analyzed striatal
degeneration following local infusion of NMDA and quinolinic acid
into the rat caudate nucleus. The use of such excitotoxic injury
models, to produce neuronal depletion, reactive gliosis, and
alterations of neurotransmitter levels, has been highly valuable for
examining pathological patterns reminiscent of Huntington's disease
(HD). Even if the primary cause of HD is unrelated, excitotoxic
injury mediated by iGluR activation may play a role in progressive
neuronal depletion (DiFiglia, 1990
).
We found (R,S)-PPG protective against NMDA- and
quinolinic acid-induced striatal lesions; to our knowledge, this
provides the first in vivo evidence that activation of group III mGluRs is neuroprotective in animal models.
Inhibition of glutamate release by presynaptic mGluR4, -7 and/or
-8 (Shigemoto et al., 1997
) may represent a common mechanism of
neuroprotection in vitro and in vivo. Accordingly, an enhanced release
of endogenous glutamate has been shown to facilitate the progression of
NMDA toxicity in cortical cultures (Monyer et al., 1992
). Additionally,
in vivo striatal toxicity induced by kainate or NMDA receptor agonists
such as quinolinic acid and NMDA critically involves the presence of
cortical glutamatergic fibers afferent to the caudate nucleus of
striatum (Colwell et al., 1996
and references therein). Although little
recurrent excitation exists in the striatum, the endogenously released
L-glutamate may have a permissive role on NMDA and
quinolinic acid toxicity, perhaps by activating postsynaptic group I
mGluRs or other facilitatory receptors.
Upon depletion of the caudate neuronal population during the
progression of HD, the resulting "excess" of corticostriatal glutamatergic input may cause further neuronal loss (DiFiglia, 1990
),
and inhibition of this input via group III mGluRs may provide protection. In addition to the regulation of presynaptic glutamate release, an inhibition of NMDA receptors by postsynaptic group III
mGluRs via a protein phosphorylation cascade (Martin et al., 1997
) may
also be involved in their neuroprotective effects. Thus, activation of
group III mGluRs could open several novel strategies to interfere with
the progressive course of neurodegenerative disorders.
The potency at cloned mGluRs relative to the protective activity of
(R,S)-PPG in primary culture
(EC50 = 12 µM) suggests that the
neuroprotective action of (R,S)-PPG
preferentially involves mGluR4, mGluR6, and/or mGluR8, which are all
expressed by cultured cortical neurons (Faden et al., 1997
). For in
vivo neuroprotection, however, mGluR6 is probably irrelevant because of
its restricted expression in the retinal bipolar cells layer. A more
detailed examination of the relative contribution of individual group
III mGluR subtypes to in vitro and in vivo neuroprotection awaits the
discovery of more selective agonists and the utilization of group III
mGluR subtype-deficient mice.
Modulation of epileptic seizures by group III mGluRs has been
frequently reported (e.g., Ghauri et al., 1996
; Abdul-Ghani et al.,
1997
; Tang et al., 1997
). Thus, we were tempted to test (R,S)-PPG, in comparison with
L-AP4 and L-SOP, in the MES
model in mice. MES is a basic screening test for anticonvulsive drugs, it is indicative of drug activity primarily against generalized tonic-clonic and, secondarily, also partial seizures, and it led to the
discovery of several clinically approved antiepileptics, e.g.,
carbamazepine, oxcarbazepine, and phenytoin (Kupferberg and Schmutz,
1997
and references therein). Surprisingly, our results for
(R,S)-PPG obtained with this model revealed
important differences as compared with the prototypic group III mGluR
agonists L-AP4 and L-SOP.
In agreement with previous results (Ghauri et al., 1996
; Tang et al.,
1997
), both L-AP4 and L-SOP
were proconvulsive at high doses (around 2000 nmol) and did not protect
against MES-induced seizures at any of the given doses (60-2400 nmol).
In contrast, (R,S)-PPG exhibited substantial
anticonvulsive activity with an ED50 value of 78 nmol, full protection at 173 nmol, and did not exhibit any
proconvulsive effect up to 2200 nmol. These discrepancies could
possibly be explained by physicochemical properties and/or divergent in
vivo metabolism of the structurally quite different group III mGluR
agonists L-AP4, L-SOP, and
(R,S)-PPG. Although we cannot exclude that
anticonvulsive properties of (R,S)-PPG are
mediated by a mechanism distinct from group III mGluRs, other reports
also support group III mGluRs as mediators of anticonvulsive and
antiepileptogenic effects. Abdul-Ghani et al. (1997)
reported protective effects of L-AP4 on development of
electrical kindling and also in fully kindled rats. Furthermore, Tang
et al. (1997)
reported for L-SOP an immediate,
transient (<10 min) proconvulsive effect followed by a prolonged (>1
day) anticonvulsive effect against sound-induced seizures with an
anticonvulsant ED50 value of 36 nmol. Moreover,
recent experimental evidence from mGluR7-deficient mice that exhibit
spontaneous epileptic seizures upon certain olfactory stimuli indicates
that at least one group III mGluR is critically involved in maintaining
the delicate balance between neuronal inhibition and excitation (H. van
der Putten, personal communication).
In conclusion, (R,S)-PPG represents a novel
pharmacological tool to analyze the role of group III mGluRs in nervous
system physiology and pathology, and our in vivo data support a
neuroprotective and anticonvulsive role for group III mGluRs and
encourage the search for systemically active group III mGluR agonists
as promising drugs for the treatment of neurological disorders, such as
Huntington's disease and epilepsies.
We thank H. Allgeier for critically reading the manuscript and
helpful discussion, R. Heckendorn for providing CGP 40116, H. van der
Putten for sharing unpublished data, and P. Schoeffter for initial cAMP
measurements with the HEK-hmGluR8a cell lines.
Accepted for publication January 25, 1999.
Received for publication October 19, 1998.