Departments of Psychopharmacology (A.N.-T., D.C., F.D., M.J.M.) and
Molecular and Cellular Pharmacology (V.A., J.-P.N., J.-A.B.), Institut
de Recherches Servier, Paris, France
The accompanying multivariate analysis of the binding profiles of
antiparkinson agents revealed contrasting patterns of affinities at
diverse classes of monoaminergic receptor. Herein, we
characterized efficacies at human (h)D2SHORT(S),
hD2LONG(L), hD3, and hD4.4 receptors and at h
2A-, h
2B-,
h
2C-, and h
1A-adrenoceptors (ARs). As
determined by guanosine
5'-O-(3-[35S]thio)triphosphate
([35S]GTP
S) binding, no ligand displayed "full"
efficacy relative to dopamine (100%) at all "D2-like"
sites. However, at hD2S receptors quinpirole, pramipexole,
ropinirole, quinerolane, pergolide, and cabergoline were as efficacious
as dopamine (Emax
100%); TL99, talipexole,
and apomorphine were highly efficacious (79-92%); piribedil,
lisuride, bromocriptine, and terguride showed intermediate efficacy
(40-55%); and roxindole displayed low efficacy (11%). For all drugs,
efficacies were lower at hD2L receptors, with terguride and
roxindole acting as antagonists. At hD3 receptors,
efficacies ranged from 33% (roxindole) to 94% (TL99), whereas, for
hD4 receptors, highest efficacies (~70%) were seen for
quinerolane, quinpirole, and TL99, whereas piribedil and terguride
behaved as antagonists and bromocriptine was inactive. Although
efficacies at hD2S versus hD2L sites were
highly correlated (r = 0.79), they correlated only
modestly to hD3/hD4 sites
(r = 0.44-0.59). In [35S]GTP
S
studies of h
2A-ARs, TL99 (108%), pramipexole (52%),
talipexole (51%), pergolide (31%), apomorphine (16%), and
quinerolane (11%) were agonists and ropinirole and roxindole were
inactive, whereas piribedil and other agents were antagonists. Similar
findings were obtained at h
2B- and
h
2C-ARs. Using [3H]phosphatidylinositol
depletion, roxindole, bromocriptine, lisuride, and
terguride displayed potent antagonist properties at
h
1A-ARs. In conclusion, antiparkinson agents display
diverse agonist and antagonist properties at multiple subtypes of
D2-like receptor and
1/
2-AR, actions, which likely contribute
to their contrasting functional profiles.
 |
Introduction |
Although
treatment of Parkinson's disease has long centered on administration
of the dopamine precursor L-dihydroxyphenylacetyl acid
(L-DOPA), there is increasing interest in the therapeutic use of dopaminergic agonists, both in association with
L-DOPA and as monotherapy (Hughes, 1997
). Inasmuch as
dopaminergic agents currently used as antiparkinson agents interact
principally with "D2-like" receptors, an
important question concerns their comparative actions at
D2 receptors (of which functionally distinct
short D2S and long D2L
isoforms exist), D3 receptors, and
D4 receptors. D2S versus
D2L receptor isoforms differ in both their
localization and their functional roles. The D2S
isoform is principally responsible for presynaptic control of dopamine
release, whereas postsynaptic D2S and
D2L receptors in the basal ganglia, via
contrasting patterns of interaction with D1
sites, differentially influence motor function; notably, blockade of
D2L sites underlies the extrapyramidal motor effects of dopaminergic antagonists (Wang et al., 2000
). As shown in
the accompanying article (Millan et al., 2002
), therapeutically used
antiparkinson agents recognize D2S and
D2L isoforms with similar affinity, and many
antiparkinson agents also interact with dopamine
D3 receptors. Although the density of striatal
D3 receptors is reduced upon degeneration of
nigrostriatal dopaminergic pathways, exposure to L-DOPA may
induce their up-regulation, reflecting complex regulatory mechanisms
involving dopamine D1 receptors and brain-derived
neurotrophic factor (Quik et al., 2000
; Guillin et al., 2001
; Joyce,
2001
). Nevertheless, the precise nature of functional
interrelationships among D3,
D2, and D1 receptors, and
the implication of D3 receptors in the
therapeutic compared with dyskinetic effects of antiparkinson agents,
remain to be clarified (Joyce, 2001
). The majority of antiparkinson
agents also show significant affinity for D4
receptors (Millan et al., 2002
), but their engagement does not improve
motor function; furthermore, antagonist properties at
D4 receptors may minimize psychiatric side
effects and improve cognitive function (Newman-Tancredi et al., 1997
;
Arnsten et al., 2000
).
Several antiparkinson drugs display pronounced affinities for
h
2A-, h
2B-, and
h
2C-ARs (Millan et al., 2002
). This is of note
in light of the importance of adrenergic mechanisms in the etiology and
treatment of Parkinson's disease (Brefel-Courbon et al., 1998
; Bezard
et al., 2001
). In addition to their postsynaptic localization,
2A-ARs are expressed as inhibitory
autoreceptors on adrenergic neurons (Nicholas et al., 1997
;
Millan et al., 2000a
,b
). Furthermore,
2A-ARs
exert an inhibitory influence upon ascending serotonergic pathways,
frontocortical and subcortical dopaminergic pathways (Kable et al.,
2000
; Millan et al., 2000a
,b
) as well as corticolimbic cholinergic and
glutamatergic pathways (Horn et al., 1982
; Tellez et al., 1997
; Boehm,
1999
). Correspondingly,
2A-ARs fulfill an
important role in the control of motor function, mood, and cognition
(Arnsten, 1997
; Kable et al., 2000
; Millan et al., 2000b
). Furthermore,
2B-ARs are enriched in the thalamus, a
structure interlinked with the basal ganglia and involved in the
disruption of motor function in Parkinson's disease, whereas
2C-ARs are concentrated in the striatum itself
(Nicholas et al., 1997
; Bezard et al., 2001
). Gene knockout studies
have indicated a modulatory influence of central
2C-ARs, complementary to
2A-ARs, upon motor and cognitive function
(Kable et al., 1999
; Bjorklund et al., 2000
). Although the precise
significance of individual
2-AR subtypes
remains unclear, there is evidence that
2-AR
antagonist properties may be useful in the management of Parkinson's
disease. First, in rats sustaining unilateral 6-hydroxydopamine lesions of the substantia nigra pars compacta (SNPC),
2-AR agonists and antagonists, respectively,
inhibit and enhance amphetamine-induced rotation (Mavridis et al.,
1991
). Second, in primates displaying Parkinson's disease-like
symptoms after treatment with
1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine,
2-AR antagonists increase locomotor activity
and reduce dyskinesias induced by L-DOPA (Brefel-Courbon et
al., 1998
; Bezard et al., 2001
). Third, after enhancement of adrenergic
transmission by blockade of
2A-AR
autoreceptors, noradrenaline (NA) may (independently of
2-ARs) exert neuroprotective actions at
dopaminergic neurons in the SNPC (Troadec et al., 2001
). Fourth,
small-scale clinical studies indicate that the
2-AR antagonist idazoxan improves motor performance in Parkinson's disease patients receiving
L-DOPA (Brefel-Courbon et al., 1998
).
Several antiparkinson agents also interact with
h
1A-, h
1B-, and
h
1D-ARs (Millan et al., 2002
). Although the
relevance of
1-ARs to management of
Parkinson's disease is less apparent than for their
2-AR counterparts, they modulate ascending
serotoninergic and dopaminergic transmission and play an important role
in motor control (Millan et al., 2000a
; Spreng et al., 2001
;
Stone et al., 2001
). Indeed,
1-AR antagonists
interfere with the induction of rotation by antiparkinson agents in
rats sustaining unilateral lesions of the SNPC (Mavridis et al., 1991
).
Furthermore, frontocortical
1-ARs are
implicated in the control of cognitive function (Arnsten, 1997
). The
perturbation of cardiovascular function associated with pronounced
activation or blockade of
1-ARs should also be accentuated (Guimarães and Moura, 2001
).
The above-mentioned observations exemplify the importance of
characterizing functional actions of antiparkinson agents at subtypes
of "hD2-like" receptor and
h
1/2-AR. However, studies have been restricted
to a few ligands at poorly characterized native sites compared with
defined classes of (cloned) human receptor (see Discussion).
Knowledge of the comparative agonist/antagonist profiles of
antiparkinson agents remains, thus, fragmentary. The present study
expanded, therefore, the multivariate analyses of binding profiles
presented in the preceding article (Millan et al., 2002
) in evaluating
efficacies of diverse antiparkinson agents at cloned
hD2S, hD2L,
hD3, and hD4 dopamine
receptors, and at h
2A-,
h
2B- h
2C-, and
h
1A-ARs, stably expressed in a common cellular
system, Chinese hamster ovary (CHO) cells.
 |
Materials and Methods |
Determination of Drug Efficacies at hD2-Like
Receptors and at h
2-AR Subtypes by
[35S]GTP
S Binding.
Efficacies at CHO-expressed
recombinant hD2S, hD2L,
hD3, and hD4
(hD4.4 isoform) receptors, and at CHO-expressed
h
2A-, h
2B-, and
h
2C-ARs were determined by measuring the
influence of drugs alone and, where appropriate, in interaction with DA
or NA upon [35S]GTP
S binding. The protocols
used have been described in detail previously (Newman-Tancredi et al.,
1997
, 1999a
,b
; Millan et al., 2001
). Briefly, the concentration of
[35S]GTP
S was 0.1 nM
(hD2S, hD2L, and
hD4), 0.2 nM (h
2A-AR,
h
2B-AR, and h
2C-AR)
or 1.0 nM (hD3). The pH was 7.4 in each case and the temperature 22°C. Incubation time was 40 min for
hD2S, hD2L, and
hD3 sites, 20 min for hD4
sites, and 60 min for h
2-AR subtypes. The
buffer contained 20 mM HEPES, 100 or 150 mM NaCl, 3 µM GDP, and 3 or
10 mM MgCl2. Membranes were incubated with the
antiparkinson agent alone and/or with DA (3 µM-hD2S, 10 µM-hD2L,
and 1 µM-hD4) or NA (10 µM for each subtype)
for 15 min before the addition of [35S]GTP
S.
Agonist efficacies were expressed as a percentage of the effect
observed with maximally effective concentrations of DA (10 µM) or NA
(10 µM). Experiments were terminated by rapid filtration through GF/B
filters (Whatman, Maidstone, UK) using a 96-well cell harvester
(Packard Instrument Company, Inc., Downers Grove, IL), and
radioactivity was determined by liquid scintillation counting.
Determination of Drug Efficacies at h
1A-ARs by
[3H]Phosphatidylinositol ([3H]PI)
Depletion.
The efficacies of drugs alone, and in interaction with
NA, were determined in CHO-expressed h
1A-ARs
as described previously (Millan et al., 2001
). Briefly, cells were
labeled with 2 µCi/ml of [3H]myoinositol
(10-20 Ci/mmol) for 24 h. Cells were washed and then incubated at
37°C for 30 min with the drug alone in Krebs-LiCl buffer: 15.6 mM
NaH2PO4 pH 7, 120 mM NaCl,
4.8 mM KCl, 1.2 mM MgSO4, 1.2 mM
CaCl2, 0.6% (w/v) glucose, 0.04% (w/v) bovine
serum albumin, and 10 mM LiCl. In the absence of NA, ~40,000 dpm was typically detected, compared with ~25,000 in the presence of a maximally effective concentration of NA (30 µM). Drug efficacies were
expressed as a percentage of the effect observed with a maximally effective concentration of NA (30 µM). For antagonist studies, cells
were preincubated for 15 min with drug before the addition of NA (10 µM) and incubation continued for 30 min. Membranes were recovered by
rapid filtration through GF/B filters (Whatman) using a 96-well cell
harvester (Packard Instrument Company, Inc.), and the
[3H]PI content was determined by scintillation
counting (Millan et al., 2001
).
Data Analyses.
[35S]GTP
S and
[3H]PI isotherms were analyzed by nonlinear
regression using the program PRISM (GraphPad Software, San Diego, CA).
KB values for inhibition of DA- or
NA-stimulated [35S]GTP
S binding at
hD2-like or h
2-ARs, and
of NA-induced [3H]PI depletion at
h
1A-ARs, were calculated as described
previously (Lazareno and Birdsall, 1993
; Newman-Tancredi et al.,
1999a
,b
) according to the equation KB = IC50/{[(2 + (agonist/EC50)nH)nH
1]
1}, where IC50 is the inhibitory
concentration50 of the antagonist, agonist is DA
or NA concentration, EC50 is effective
concentration50 of DA or NA alone, and
nH is Hill coefficient of the DA or NA stimulation isotherm. EC50 values for NA at
h
2A-, h
2B-,
h
2C-, and h
1A-ARs
were 354, 316, 302, and 329 nM, respectively.
EC50 values for DA at hD2S,
hD2L, hD3, and
hD4 receptors were 350, 340, 11, and 100 nM,
respectively. Protein concentrations were determined by use of a
bichinconic acid kit (Sigma, St. Quentin Fallavier, France). Pearson
product-moment correlation coefficients were calculated for
pEC50 values determined herein compared with pKi values determined in the
accompanying article (Millan et al., 2002
).
Drugs.
Pramipexole dihydrochloride, piribedil hydrochloride,
and ropinirole were synthesized by Servier Institut de Recherches
(Paris, France). Lisuride maleate and terguride were donated by
Schering (Berlin, Germany). Bromocriptine, (
)-quinpirole HCl,
pergolide methanesulfonate, and TL99
(6,7-dihydroxy-N,N-dimethyl-2-aminotetralin) were
purchased from Sigma/RBI (Natick, MA). Apomorphine hydrochloride was
purchased from Sigma. Roxindole methanesulfonate was donated by Merck
(Darmstadt, Germany) and talipexole by Boehringer Ingelheim GmbH
(Ingelheim, Germany). Cabergoline was obtained from Farmitalia Carlo
Erba (Rueil-Malmaison, France). Quinelorane dihydrochloride was a gift
from Eli Lilly & Co. (Indianapolis, IN).
 |
Results |
Drug Actions at hD2S Receptors.
At
hD2S receptors
(Bmax = 1.4 pmol/mg), at a maximally
effective concentration (10 µM), DA enhanced
[35S]GTP
S binding by ~2.5-fold; it
displayed a pEC50 value of 6.5 (Fig.
1; Table
1). Quinpirole, pramipexole,
quinelorane, pergolide, and cabergoline behaved as highly efficacious
agonists at hD2S receptors in stimulating G
protein activation ([35S]GTP
S binding) to a
degree at least as marked as that of DA (Emax defined as 100%). TL99,
talipexole, and apomorphine also showed high efficacies, whereas other
ligands displayed less pronounced efficacies, ranging from 40% for
terguride to 55% for lisuride. Roxindole showed very low efficacy.
Drug potencies for stimulation of [35S]GTP
S
binding (pEC50 values) at
hD2S receptors correlated well (r = 0.82, P < 0.05) with their
pKi values determined in competition binding experiments (data not shown; Millan et al., 2002
).

View larger version (24K):
[in this window]
[in a new window]
|
Fig. 1.
Influence of antiparkinson agents upon G protein
coupling at hD2S (A), hD2L (B), and
hD3 (C) receptors expressed in CHO cells.
[35S]GTP S binding was carried out as described in
Table 1. Binding is expressed as a percentage of that observed with a
maximally effective concentration (10 µM) of dopamine (defined as
100%). Values shown are from representative experiments performed in
triplicate and repeated on at least three occasions.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 1
Efficacies (Emax values) and potencies
(pEC50 or pKb values) of
antiparkinson agents at recombinant hD2S, hD2L,
hD3, and hD4 receptors
Efficacy (Emax) and potency (pEC50 or
pKb) values at hD2S, hD2L,
hD3, and hD4 receptors were determined by
[35S]GTP S binding. pEC50 values for stimulation
are indicated in normal case and pKb values
for inhibition are indicated in bold. Emax values
are percentages of the stimulation observed with a maximally
efficacious (Emax = 100%) concentration of dopamine
(see Materials and Methods) and are expressed as means ± S.E.M. values of at least three independent determinations performed
in triplicate. pEC50 values are means of at least three
independent determinations: S.E.M.s values (not shown) were less than
0.2 log units. Dopamine exhibited pEC50 values of 6.46, 6.49, 7.95, and 7.00 at hD2S, hD2L, hD3, and
hD4 receptors, respectively. For piribedil, the
pKb at hD4 receptors is indicated in
the table. It exhibited an agonist pEC50 value of 6.4 at these
sites. For roxindole, the pKb
at hD2S receptors is given in the table. It exhibited an
agonist pEC50 value of 8.11 at these sites.
|
|
Drug Actions at hD2L Receptors.
At
hD2L receptors
(Bmax = 2.2 pmol/mg), at a maximally
effective concentration (10 µM), DA enhanced
[35S]GTP
S binding by ~1.9-fold; it
displayed a pEC50 value of 6.5 (Fig. 1; Table 1).
At hD2L receptors, efficacies for all ligands were markedly lower than at hD2S sites. Indeed,
all ligands, except quinelorane, behaved as partial agonists. Roxindole
and terguride induced no stimulation of
[35S]GTP
S binding and displayed antagonist
properties. The correlation coefficient for efficacies at
hD2L versus hD2S receptors
was 0.79 (P < 0.05). Drug potencies for stimulation of
[35S]GTP
S binding
(pEC50 values) at hD2L
receptors correlated well (r = 0.93, P < 0.05) with their pKi values
determined in competition binding experiments (data not shown; Millan
et al., 2002
).
Drug Actions at hD3 Receptors.
At
hD3 receptors
(Bmax = 15.6 pmol/mg), at a maximally
effective concentration (10 µM), DA enhanced
[35S]GTP
S binding by ~1.6-fold; it
displayed a pEC50 value of 7.8 (Fig. 1; Table 1).
All drugs behaved as agonists at hD3 receptors, with efficacies varying from 30% (roxindole) and 34% (piribedil) to
88% (talipexole) and 94% (TL99). The lower efficacies of quinelorane and quinpirole at hD3 versus
hD2S and hD2L sites are of
note, whereas roxindole and bromocriptine showed higher efficacies at hD3 than hD2S and
hD2L sites. Indeed, there was no consistent pattern of drug efficacies at hD3 relative to
hD2S and hD2L receptors. Accordingly, correlation coefficients for efficacies at
hD3 compared with hD2L and
hD2S sites, although significant
(P < 0.05), were only 0.59 and 0.49, respectively.
Drug potencies for stimulation of [35S]GTP
S
binding (pEC50 values) at
hD3 receptors correlated significantly (r = 0.62, P < 0.05) with
pKi values determined in competition binding experiments (data not shown; Millan et al., 2002
).
Drug Actions at hD4 Receptors.
At
hD4 receptors
(Bmax = 1.4 pmol/mg), at a maximally
effective concentration (10 µM), DA enhanced
[35S]GTP
S binding by ~2.2-fold; it
displayed a pEC50 value of 7.0 (Table 1).
Although bromocriptine did not interact with hD4
receptors, agonist efficacies of the other drugs varied widely. Thus,
although TL99, quinelorane, and quinpirole showed relatively high
efficacies (~70%), pergolide, talipexole, cabergoline, apomorphine,
roxindole, pramipexole, and lisuride showed less marked efficacies of
32 to 56%. Piribedil displayed very low efficacy (7%) and antagonized the stimulation by DA of [35S]GTP
S binding.
Terguride, which was inactive alone, similarly blocked the action of
DA. On the other hand, roxindole was more efficacious at
hD4 than at hD2S and
hD2L receptors. Thus, there was no consistent
pattern of drug efficacies at hD4 versus
hD2S, hD2L, and
hD3 receptors and correlation coefficients,
although significant (P < 0.05), were modest:
hD2S, r = 0.57;
hD2L, r = 0.55; and
hD3, r = 0.44. Drug potencies for
stimulation of [35S]GTP
S binding
(pEC50 values) at hD4
receptors correlated well (r = 0.78, P < 0.05) with their pKi values
determined in competition binding experiments (data not shown; Millan
et al., 2002
).
Drug Actions at h
2A-, h
2B-, and
h
2C-ARs.
At h
2A-,
h
2B-, and h
2C-ARs, a
maximally effective concentration of NA (10 µM) increased
[35S]GTP
S binding by 7.2-, 6.6-, and
2.7-fold, respectively; pEC50 values were 6.2, 6.5, and 6.5, respectively (Figs. 2,
3 and 4; Table 2). Antiparkinson agents
differed markedly concerning their functional activities at
h
2A-, h
2B-, and
h
2C-ARs. TL99 behaved as a high-efficacy
agonist at each subtype of h
2-AR, whereas talipexole behaved as a partial agonist at each subtype. On the other
hand, pergolide showed pronounced efficacy at
h
2B-ARs, intermediate efficacy at
h
2A-ARs, and low efficacy at
h
2C-ARs. Pramipexole revealed partial agonist
properties at h
2A-ARs; actions were not
evaluated at h
2B- and
h
2C-ARs owing to its low affinities at
these sites (Millan et al., 2002
). Apomorphine showed low efficacy only
at h
2A-ARs, whereas high concentrations of
quinelorane and quinpirole revealed weak partial agonist actions at
h
2A- and h
2B-ARs,
respectively. Agonist pEC50 values for
stimulation of [35S]GTP
S binding
corresponded well to their respective
pKi values as defined in competition
binding assays (Millan et al., 2002
). In view of its high affinity and
low efficacy at h
2A-AR subtypes, apomorphine
was further evaluated in interaction with NA and shown to behave as an
antagonist. Furthermore, several other drugs also reversed
NA-stimulated [35S]GTP
S binding. For drugs
behaving as antagonists, pKB values correlated well (P < 0.05) with their respective
pKi values derived from competition
binding studies (data not shown; Millan et al., 2002
):
h
2A-ARs, r = 0.94;
h
2B-ARs, r = 0.86; and
h
2C-ARs, r = 0.87.

View larger version (18K):
[in this window]
[in a new window]
|
Fig. 2.
Influence of antiparkinson agents upon G protein
coupling at h 2A-adrenoceptors expressed in CHO cells.
[35S]GTP S binding was carried out as described in
Table 1. Binding is expressed as a percentage of that observed with a
maximally effective concentration (10 µM) of noradrenaline (defined
as 100%). Values shown are from representative experiments performed
in triplicate and repeated on at least three occasions.
|
|

View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3.
Influence of antiparkinson agents upon G protein
coupling at h 2B-adrenoceptors expressed in CHO cells.
[35S]GTP S binding was carried out as described in
Table 1. [35S]GTP S binding is expressed as a
percentage of that observed with a maximally effective concentration
(10 µM) of noradrenaline (defined as 100%). Values shown are from
representative experiments performed in triplicate and repeated on at
least three occasions.
|
|

View larger version (18K):
[in this window]
[in a new window]
|
Fig. 4.
Influence of antiparkinson agents upon G protein
coupling at h 2C-adrenoceptors expressed in CHO cells.
[35S]GTP S binding was carried out as described in
Table 1. [35S]GTP S binding is expressed as a
percentage of that observed with a maximally effective concentration
(10 µM) of noradrenaline defined as 100%). Values shown are from
representative experiments performed in triplicate and repeated on at
least three occasions.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 2
Efficacies (Emax values) and potencies
(pEC50 or pKb values) of antiparkinson
agents at recombinant h 2A-, h 2B-, and
h 2C-adrenoceptors
Efficacy (Emax) and potency (pEC50 or
pKb) values were determined by
[35S]GTP S binding. pEC50 values for stimulation
are indicated in normal case, and pKb values
for inhibition are indicated in bold. Emax values
are percentages of the stimulation observed with a maximally
efficacious (Emax = 100%) concentration of
noradrenaline (10 µM) and are expressed as means ± S.E.M.
values of at least three independent determinations performed in
triplicate. pEC50 or pKb values are
means of at least three independent determinations: S.E.M. values (not
shown) were less than 0.2 log units. The Bmax at
h 2A-, h 2B-, and h 2C-ARs was 1.8, 1.0, and 1.3 pmol/mg, respectively. Noradrenaline exhibited pEC50
values of 6.45, 6.50, and 6.52 at h 2A-, h 2B-, and
h 2C-ARs, respectively. Apomorphine displayed a
pKb of 6.58 at h 2A-ARs,
and pergolide displayed a pKb of 6.96 at h 2C-ARs.
|
|
Drug Actions at h
1A-ARs.
Ligands that
exhibited significant binding affinity at
h
1A-ARs (pKi
values
6.0; Millan et al., 2002
) were evaluated in a functional test
of phospholipase C activation, depletion of membrane-bound [3H]PI (Fig. 5;
Table 3). In this procedure, NA
itself revealed a pEC50 value of 6.51. No
compound stimulated phospholipase C activity when tested alone,
indicating an absence of agonist properties. In contrast, in order of
decreasing potency, roxindole, bromocriptine, lisuride, terguride,
cabergoline, and piribedil all reversed the stimulation of
[3H]PI hydrolysis induced by noradrenaline (10 µM), demonstrating antagonist properties.
pKB values at
h
1A-ARs correlated well (r = 0.96, P < 0.05) with
pKi values obtained from competition binding assays (data not shown; Millan et al., 2002
).

View larger version (16K):
[in this window]
[in a new window]
|
Fig. 5.
Influence of antiparkinson agents upon stimulation of
phospholipase C activity by noradrenaline at
h 1A-receptors expressed in CHO cells.
[3H]PI depletion studies were carried out as described
under Materials and Methods. The antagonist actions of
drugs were examined against noradrenaline (10 µM). Values shown are
from representative experiments performed in triplicate and repeated on
at least three occasions.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 3
Efficacies and potencies (pKb
values) of antiparkinson agents at recombinant h 1A-ARs
Efficacies (Emax) are for drugs alone compared with
a maximally effective (100%) concentration of NA (30 µM). Potencies
(pKb) at h 1A-ARs
(Bmax = 2.5 pmol/mg) were determined by blockade of
NA-stimulated [3H]PI depletion. Noradrenaline exhibited a
pEC50 value of 6.48. Pramipexole, ropinirole, quinelorane,
apomorphine, pergolide, talipexole, TL99, and quinpirole, which have
low affinities at h 1A-ARs (pKi values of
<6.0, see accompanying paper) were not tested.
pKb values are means of at least three
independent determinations: S.E.M. values were less than 0.2 log units.
Experiments were carried out in triplicate.
|
|
 |
Discussion |
This comprehensive comparison of the actions of 14 antiparkinson
agents at eight classes of cloned "hD2-like"
and h
1/h
2-AR revealed
marked differences in efficacies, observations of significance to their
contrasting functional profiles in experimental models and in human.
D2S and D2L Receptors.
In the only
previous comparison of antiparkinson agonists at
hD2S versus hD2L receptors
(ropinirole, talipexole, pergolide, lisuride, and bromocriptine), no
marked differences in efficacy were apparent (Gardner et al., 1997
;
Gardner and Strange, 1998
). In the present, more extensive work,
however, drug efficacies were invariably higher at
hD2S sites. The reasons underlying this difference require elucidation, but it may be of pertinence that hD2S and hD2L receptors
differentially interact with distinct subtypes of G protein, as
implicated in their contrasting patterns of coupling to calcium
channels (Wolfe and Morris, 1999
). Furthermore, the expression level
(2.2 pmol/mg) of hD2L sites herein was higher than that of hD2S sites (1.4 pmol/mg), whereas
the inverse was true for Gardner et al. (1997)
(1.3 versus 2.7 pmol/mg,
respectively). In the light of these comments, it should briefly be
pointed out that receptor density can play an important role in
determining drug efficacy (Newman-Tancredi et al., 2001
). Although
receptor density (Bmax) can easily be
determined at pure populations of transfected receptors (see
Results), equivalent information for defined networks of
neurons is not available because Bmax
estimations in native tissue almost inevitably incorporate neurons not
expressing the receptor in question. Thus, although the receptor
densities of hD2S and hD2L
sites here were in the same range as previous studies (Coldwall et al.,
1999
; Perachon et al., 1999
), they cannot be compared with
certainty to cerebral populations. Furthermore, it is important to note
differences in density between pre- versus postsynaptic populations of
D2 (and D3) receptors
(Seyfried and Boettcher, 1990
), as well as the up-regulation of
postsynaptic sites after damage to dopaminergic innervation (Kostrzewa,
1995
; Newman-Tancredi et al., 2001
), an experimental manipulation that mimics the pathology of Parkinson's disease (see below).
Although the relative degree of D2S versus
D2L receptor stimulation required for optimal
control of Parkinson's disease remains to be clarified, quinelorane
was the only drug to exhibit efficacy equivalent to dopamine at both
hD2S and hD2L receptors, in
line with its high efficacy at native, rat D2
receptors (Sánchez and Arnt, 1992
; Newman-Tancredi et al., 2001
).
The relatively high efficacies of quinpirole, ropinirole, pramipexole,
and talipexole at hD2S (Terasmaa et al., 2000
)
and hD2L sites for stimulation of
[35S]GTP
S binding coincide with measures of
extracellular acidification and mitogenesis (Mierau et al., 1995
;
Coldwall et al., 1999
; Perachon et al., 1999
; Alberts et al., 2000
;
Gilliland and Alper, 2000
). The present
[35S]GTP
S approach likewise revealed high
efficacies at hD2S and hD2L
sites of cabergoline and TL99 (Hughes, 1997
). Like bromocriptine and
lisuride, piribedil displayed intermediate efficacy. This is
interesting because piribedil is highly active in rodent and primate
models of antiparkinson activity; furthermore, piribedil improves motor
and cognitive function in patients both alone and in association with
L-DOPA (Rondot and Ziegler, 1992
; Maurin et al., 2001
;
Nagaraja and Jayashree, 2001
). Interestingly, terguride failed to
activate hD2L receptors, in line with its low
efficacy in rodent models of hypothermia, locomotion, and drug
discrimination (Arnt and Hyttel, 1990
; Sánchez and Arnt, 1992
).
Although terguride showed weak partial agonist activity in
hD2L receptor-expressing SH-SY5Y cells, its
efficacy was much lower than that of quinpirole (Gilliland and Alper,
2000
). Furthermore, although terguride showed modest antiparkinson
activity and attenuated L-DOPA-induced dyskinesia in
primates, it was effective in only a small percentage (10-20%) of
Parkinson's disease patients in (subsequently discontinued) clinical
trials (Filipova et al., 1988
; Akai et al., 1993
). Furthermore, roxindole, which likewise exhibited low efficacy at
hD2L and hD2S receptors
(Newman-Tancredi et al., 1999a
), failed to reduce
L-DOPA-induced dyskinesias in Parkinson's disease patients
and has not, as yet, been shown to possess antiparkinson activity in human.
Correspondingly, a certain, minimal "threshold" of efficacy may be
necessary for antiparkinson properties. However, "full" agonism at
the level of G protein-coupling ([35S]GTP
S
binding) is not essential for robust clinical activity in Parkinson's
disease because 1) efficacy is "amplified" by intracellular cascades downstream of G proteins (Cussac et al., 2002
); 2)
postsynaptic striatal D2 receptors (probably the
D2L isoform) are hypersensitive due to loss of
dopaminergic input from the SNPC (Kostrzewa, 1995
; Geurts et al., 1999
;
Newman-Tancredi et al., 2001
); and 3) submaximal efficacy is sufficient
to activate highly sensitive D2S autoreceptors implicated in neuroprotective properties of dopaminergic agonists (Seyfried and Boettcher, 1990
). Moreover, antiparkinson agents of intermediate efficacy may preferentially engage nigrostriatal D2 receptors implicated in the treatment of
Parkinson's disease compared with other populations mediating side
effects. Thus, "submaximal" efficacy at the G protein level for
drugs such as piribedil or bromocriptine may be advantageous in
optimizing the therapeutic index between clinical efficacy and side effects.
hD3 Receptors.
Although hD3
receptors couple less efficiently to G proteins in CHO cells than their
hD2S/hD2L counterparts, DA
stimulated [35S]GTP
S binding in the
high-expressing cell line used herein (Newman-Tancredi et al., 1999b
).
The substantial affinities of apomorphine, quinpirole, pramipexole,
talipexole, bromocriptine, and pergolide corroborate studies of their
actions in models of microphysiometry and mitogenesis (Mierau et al.,
1995
; Coldwell et al., 1999
; Perachon et al., 1999
). The high efficacy
of TL99 at hD3 receptors is of note in view of
its marked efficacy at
hD2S/hD2L and
hD4 sites, whereas the modest efficacies of
terguride and roxindole at hD3 sites mimic their
low efficacies at hD2L and (terguride)
hD4 sites. As concerns piribedil, its
intermediate efficacy at hD3 receptors resembles
its actions at hD2S/hD2L
receptors and is consistent with agonist properties in vivo at
D3 autoreceptors (Millan et al., 1995
). As
discussed elsewhere (Joyce, 2001
), the role of D3
sites in the expression of beneficial and deleterious actions of
antiparkinson agents remains unclear, a question of particular importance because, as shown herein, all antiparkinson agents activated
D3 receptors.
hD4 Receptors.
In line with studies of CHO cells
expressing the hD4.2 isoform (Gilliland and
Alper, 2000
) and of cloned, rat D4 sites (Gazi et
al., 2000
), quinpirole showed substantial efficacy at
hD4 (hD4.4) receptors. This
characteristic was shared by quinerolane and TL99. The agonist
properties of pergolide, apomorphine, talipexole, and pramipexole at
hD4 sites complement work using other measures of
drug efficacy and/or other hD4 isoforms (Mieurau
et al., 1995
; Coldwell et al., 1999
; Gazi et al., 2000
; Gilliland and
Alper, 2000
). Like pergolide, two other ergolines, cabergoline and
lisuride, similarly showed agonist properties at
hD4 sites. In contrast, piribedil displayed low
efficacy at hD4 receptors, whereas bromocriptine was inactive. Because bromocriptine and piribedil are clinically efficacious antiparkinson agents, these data support the argument that
activation of D4 receptors is not necessary for
therapeutic efficacy (Newman-Tancredi et al., 1997
). Moreover,
the essentially D4 antagonist properties of
piribedil may limit psychiatric side effects and contribute to its
improvement of cognitive function (Arnsten et al., 2000
; Nagaraja and
Jayashree, 2001
).
h
2-ARs.
Striking differences in drug efficacies
were seen at h
2-AR subtypes. In analogy to
piribedil (Millan et al., 2001
), lisuride, bromocriptine, and
apomorphine displayed antagonist properties, observations amplifying
functional studies of isolated organs and hippocampal NA release in
rats (McPherson, 1984
; Jackisch et al., 1985
). In line with their high
affinities for h
2-ARs (Millan et al., 2002
),
roxindole and two further ergot-related ligands, terguride and
cabergoline, also manifested potent
2-AR antagonist properties. In contrast, in line with in vivo studies (at
undefined
2-AR subtypes) in rodents (Horn et
al., 1982
; Meltzer et al., 1989
; Sánchez and Arnt, 1992
), TL99
displayed agonist, and talipexole partial agonist, properties at
h
2A-, h
2B-, and h
2C-ARs. Extending observations of partial
agonist properties at central
2-ARs in rodents
(Ferrari et al., 1993
), pramipexole displayed modest efficacy at
h
2A-ARs. Any potential significance of this
(low-potency) action in vivo, however, remains to be clarified. On the
other hand, in vivo studies in rodents have revealed agonist actions of
pergolide at central
2-ARs (Langtry and
Clissold, 1990
) and particularly pronounced agonist properties at
h
2B-ARs were observed here.
These contrasting actions of antiparkinson agents are of considerable
significance in light of evidence that blockade of
2-ARs improves motor performance, cognitive
function, and perhaps mood in Parkinson's disease (Brefel-Courbon et
al., 1998
). Indeed, experimental and clinical studies with
piribedil support the notion that "built-in"
2-AR antagonist actions may be beneficial in Parkinson's disease (Maurin et al., 2001
; Millan et al., 2001
; Nagaraja and Jayashree, 2001
). In contrast,
2-AR agonists interfere with the facilitory
influence of antiparkinson agents upon motor function (Mavridis et al.,
1991
; Bezard et al., 2001
). Indeed, talipexole-induced stereotypy in
rats (which reflects agonist properties at striatal
D2 receptors) is only apparent upon prevention of
its
2-AR agonist properties by
coadministration of idazoxan (Meltzer et al., 1989
). Similarly,
TL99-induced hypomotility has been attributed to its
2-AR agonist properties (Horn et al., 1982
),
whereas it only elicits rotation in unilateral SNPC-lesioned rats upon
cotreatment with
2-AR antagonists (Martin et
al., 1983
).
Nevertheless, future studies should address the significance of
2-AR subtypes in the clinical actions of
antiparkinson drugs. Although
2A-ARs are
certainly of key importance (Kable et al., 2000
; Millan et al., 2000a
),
2B- and
2C-ARs sites
should not be neglected. The former are concentrated in the thalamus, a
structure intimately involved in the motor deficits of Parkinson's
disease, whereas
2C-ARs are enriched in the
striatum itself (Nicholas et al., 1997
; Bezard et al., 2001
).
Moreover, gene knockout studies indicate that
2C-ARs contribute to the modulation of
monoaminergic transmission, cognitive function, and motor performance
(Bjorklund et al., 2000
; Kable et al., 2000
). Although no antiparkinson
agent showed agonist versus antagonist actions at distinct
2-AR subtypes, the lack of antagonist actions
of piribedil at
2B- versus
2A/2C-ARs sites, and the preferential agonist
actions of pergolide at
2B- versus
2A- and
2C-ARs, may
prove instructive in elucidating their relevance to Parkinson's
disease and its treatment.
Actions at h
1A-ARs.
Roxindole and the ergot
derivatives bromocriptine, lisuride, and terguride interact with cloned
h
1-ARs (Millan et al., 2002
), although, with
the exception of a study of bromocripine at peripheral
1-ARs (McPherson, 1984
), no information on
their functional activities is available. Thus, their potent antagonism
of NA-induced [3H]PI depletion at cloned
h
1A-ARs is of note, whereas cabergoline and
piribedil showed weak antagonist properties in line with their low
affinities at these sites (Millan et al., 2002
). Potent blockade of
h
1-ARs may be unfavorable inasmuch as
1-AR antagonists interfere with antiparkinson
properties in experimental models (Mavridis et al., 1991
). Although
blockade of
1-ARs in the cortex and on pars
reticulata GABAergic neurons may be involved, generalized sedative/motor-suppressive effects due to blockade of
1A-ARs in motor nuclei of the brainstem and
the spinal cord may also be of significance (Stone et al., 2001
).
Blockade of segmental and peripheral
1A-ARs
may also be deleterious in that it exacerbates the perturbation of
cardiovascular function elicited via stimulation of spinal dopaminergic
receptors (Guimarães and Moura, 2001
). On the other
hand, inasmuch as frontocortical
1-ARs inhibit
working memory, their blockade might improve cognitive function
(Arnsten, 1997
), although there is currently no clinical support for
this possibility. Further investigations should determine the potential importance for antiparkinson agents of actions at
h
1B- and h
1D-AR subtypes (Millan et al., 2002
), which likewise modulate motor, cognitive, and cardiovascular function (Guimarães and Moura, 2001
; Spreng et al., 2001
; Stone et al., 2001
).
 |
Conclusions |
The present data reveal striking differences among antiparkinson
agents concerning efficacies at multiple classes of
hD2-like receptor and
ha1/ha2-AR. These
observations amplify receptor-binding analyses of the accompanying
article (Millan et al., 2002
) in demonstrating that antiparkinson drugs
are heterogeneous rather than a common group of "dopaminergic
agonists". In this light, as discussed above, partial agonist and
agonist properties at D2S/D2L receptors are
favorable in the management of motor symptoms of Parkinson's disease,
whereas blockade of
2A-ARs may improve cognitive-attentional function and mood. The present data provide a
framework for additional studies of the significance of these and other
subtypes of dopaminergic receptor and
1/
2-AR in the etiology of Parkinson's disease, and in the beneficial and deleterious properties of antiparkinson agents.
We thank M. Soubeyran for secretarial assistance, and Laurence
Verrièle, Manuelle Touzard, Christine Chaput, Valérie
Pasteau, Laetitia Marini, Nelly Fabry, and Anne Bonnard for technical assistance.
Accepted for publication July 22, 2002.
Received for publication June 12, 2002.