Design,
Synthesis, and Evaluation of Potent and
Selective Ligands for the Dopamine 3 (D3)
Receptor with a Novel in Vivo Behavioral
Profile
Jianyong Chen, Gregory T. Collins, Jian
Zhang, Chao-Yie Yang, Beth Levant, James Woods,
and Shaomeng Wang
Departments of Internal
Medicine, Pharmacology, and Medicinal Chemistry,
University of Michigan
Department of Pharmacology,
Toxicology, and Therapeutics, University of
Kansas
Abstract: A series of compounds structurally
related to pramipexole were designed,
synthesized, and evaluated as ligands for the
dopamine 3 (D3) receptor. Compound 12 has a Ki
value of 0.41 nM to D3 and a selectivity of
>30000- and 800-fold over the D1-like and D2
receptors, respectively. Our in vivo functional
assays showed that this compound is a partial
agonist at the D3 receptor with no detectable
activity at the D2 receptor.
Dopaminergic neurotransmission is mediated
by five dopamine receptors (D1-D5), which can be
grouped into the D1-like (D1 and D5) and D2-like
(D2, D3, and D4 a) receptor subtypes. Recent
studies have suggested that the D3 receptor is a
promising therapeutic target for a variety of
conditions, including drug abuse, restless legs
syndrome, schizophrenia, Parkinson's disease,
and depression. Considerable effort has been
devoted in recent years to the discovery and
development of potent and selective D3 ligands.
Despite intense research efforts, design of
truly selective D3 ligands with good solubility
and bioavailability remains a challenge.
Compound 1 (pramipexole) is a potent
D3-preferring agonist but has limited
selectivity over the D2 receptor in vitro and in
vivo. Compound 2 was initially reported as a D3
partial agonist and has a 67-fold selectivity
over the D2 receptor.2 A number of potent and
selective D3 ligands, such as 3, have been
designed based upon the core structure of 2. Our
laboratory has reported the design of 4 as a
potent and selective D3 ligand using the
hexahydropyrazinoquinoline as the core
structure. Despite its relatively high affinity
and excellent selectivity for D3 over other
dopamine receptor subtypes, 4 has a poor aqueous
solubility, which limits its in vivo
evaluations.
The poor aqueous solubility is also a major
limitation for many recently described potent
and selective D3 ligands and an obstacle for
evaluation of these novel agents in behavioral
models in animals and their therapeutic
potential. To overcome this major limitation, we
investigated other core structures for the
design of potent and selective D3 ligands. Among
them, the core structure in 1 has a number of
very attractive features. First, 1 itself is a
very potent D3 ligand and has a Ki value of 0.78
nM to D3 in our binding assay.
Second and importantly, 1 has an excellent
aqueous solubility. Third, pramipexole
dihydrochloride has been approved for the
treatment of Parkinson's disease and restless
legs syndrome and has an excellent
pharmacological and toxicological profile in
humans and in animals. Hence, 1 represents a
particularly attractive template for the design
of potent and selective D3 ligands with
desirable physiochemical and pharmacological
properties.
Of note, although 1 has been widely used as
a D3 preferring ligand, it potently binds to the
high affinity state of the D2 receptor with a Ki
value of 3.1 nM in our binding assays (Table 1),
thus displaying only a 4-fold selectivity for
the D3 receptor over the D2 receptor. Recently,
the crystal structures for the human 2
adrenergic ( 2AD) G-protein coupled receptor
(GPCR) were solved.
We have modeled the human D3 receptor
structure based upon the high-resolution crystal
structures of 2AD receptor because these two
proteins belong to the same GPCR subfamily28 and
share close sequence homology. Because the
crystal structure of 2AD receptor was solved
with an inverse agonist bound to it, our modeled
D3 structure likely represents the
conformational state bound to either antagonists
or inverse agonists. Hence, care must be taken
when using the structure to model the
interactions of the D3 receptor with its ligands
with different intrinsic functions.
Nevertheless, we reasoned the modeled
human D3 structure based upon the very first
human GPCR structure could be useful to guide
the design of novel D3 ligands. To this end, we
modeled the binding of 1 to the D3 receptor
structure through computational docking,
followed by extensive refinement (Supporting
Information).
The predicted model showed that the primary
amino group in the thiazol ring of 1 forms a
hydrogen bonding network with the hydroxyl
groups of Ser192 and Ser193. The thiazol ring in
1 is parallel to the imidazole ring in His349,
making favorable ¹-¹ stacking interaction. The
protonated nitrogen in 1 forms a salt bridge
with the negatively charged Asp110. The n-propyl
group in 1 inserts into a hydrophobic channel
formed by Cys114, Phe345, Phe346, Trp342, and
Try373. The predicted model of 1 in complex with
the D3 receptor suggested that there is ample
room available to accommodate a much larger
hydrophobic group where the n-propyl group in 1
binds. Interestingly, in the adjacent area,
there is another welldefined but smaller
hydrophobic cavity formed by Cys114, Phe197, and
Trp342 residues.
We have thus designed and synthesized
compound 5 to explore the interactions with
these two pockets. Compound 5 was tested for its
binding affinities to the dopamine receptors
using the same methods as described previously.
It was found that 5 has a Ki value of 0.043 nM
to the D3 receptor, being 18 times more potent
than 1. Compound 5, however, also potently binds
to the high affinity state of the D2 receptor
with a Ki value of 2.7 nM, thus displaying a
62-fold selectivity for the D3 receptor over the
D2 receptor. Similar to 1, 5 has a weak affinity
to the D1-like receptors and has a Ki value of
11000 nM. Hence, although 5 has a very high
affinity to the D3 receptor, its selectivity
over the D2 receptor is modest. In our previous
design of 4, we have shown that introduction of
a trans-cyclohexyl group into the linker region
yielded new ligands with much improved
selectivity for the D3 receptor over the D2
receptor as compared to a linear 4-carbon
linker.
We have thus designed compound 6 to
investigate if introduction of this rigid
cyclohexyl group into 5 may also improve the
selectivity. Compound 6 binds to the D3 and D2
receptors with Ki values of 0.40 and 307 nM,
respectively. Hence, 6 is a potent D3 ligand and
displays an excellent selectivity of 763-fold
for the D3 receptor over the D2 receptor. We
next designed and synthesized compounds 7-10 to
investigate the importance of the n-propyl group
in 6 for binding and selectivity. Compound 7
with an n-butyl group has a slightly weaker
affinity for the D3 receptor than 6 and
exhibited a twosite competition curve at the D2
receptor, with roughly 10-fold less selectivity
for the D3 receptor over the D2 receptor with
the high affinity binding component. Compound 8
with an isopentyl group is five times less
potent than 6 to the D3 receptor but has a
similar binding affinity to the D2
receptor.
Compound 9 with a bulky cyclohexylethyl
group is 55 times less potent than 6 to the D3
receptor but is only three times less potent
than 6 to the D2 receptor. Compound 10 with a
hydrogen atom at this site has a Ki value of 7.6
nM to the D3 receptor, being 19 times less
potent than 6, but their binding affinities to
the D2 receptor are essentially the same.
Therefore, our binding data clearly showed that
the substitution on this nitrogen atom has a
major effect on the binding to the D3 receptor
but modest influence on the binding to the D2
receptor. Our data also showed that the n-propyl
group in 6 enhances the binding affinity to the
D3-receptor by 19-fold as compared to a hydrogen
atom in 10.
We next investigated the influence of the
naphthyl group in 6 for binding and selectivity.
Compound 11, in which the naphthyl group is
replaced by a 2-benzofuran, binds to the D3
receptor with the same affinity (Ki ) 0.51 nM)
as 6, but its selectivity over the D2 receptor
is decreased to 133-fold due to its increased
binding affinity to the D2 receptor. Compound
12, in which a cinnamyl group is used to replace
the naphthyl, retains a high binding affinity
for the D3-receptor (Ki ) 0.41 nM) and displays
800- and >30000-fold selectivity over the D2
and D1-like receptors. These data suggested that
the modifications of the naphthyl group can have
a significant effect on the selectivity, and
this region should be further investigated for
the design of potent and selective D3 ligands.
The synthesis of compounds 5-12 is provided in
the Supporting Information. Compounds 5, 6, and
12 were found to have good aqueous solubility.
For example, the dihydrochloride salt form of 6
has an aqueous solubility greater than 100
mg/mL. Their excellent aqueous solubility
provided us with an opportunity to evaluate
their in vivo functional profiles in animals.
Another challenge in the development of
selective D3 ligands was that the current in
vitro functional assays for the D3 receptor are
not predictive of the in vivo function of D3
ligands. Furthermore, there was also the lack of
a robust in vivo functional assay for the D3
receptor. To addresses these challenges, we have
recently validated in vivo functional assays for
the D3 and D2 receptors.
Our studies showed that yawning in
rats provides a sensitive measure of in vivo
agonist activity at the dopamine D3 receptor,
while the induction of hypothermia has been
shown to be mediated by agonist activity at the
D2 receptor. Employing these well-validated
assays, we evaluated 5, 6, and 12 for their in
vivo functional activity at the D3 and D2
receptors. Compound 1, a known D3 and D2
agonist, was used as a control in our
evaluations. Consistent with the data obtained
in previous studies, increases in yawning
were observed over low doses (0.01 to 0.1 mg/kg)
of 1 with inhibition of yawning and the
induction of hypothermia occurring at higher
doses.
These data indicate that 1 functions as a
preferential D3 agonist in vivo and a D2 agonist
at higher doses. Compound 5 induced
yawning and produced an inverted U-shaped
dose-response curve. The maximum levels of
yawning induced by 5 are very similar to
that induced by 1. Furthermore, hypothermia was
induced by 5 at higher doses, concurrent with
deceases in yawning. These data showed
that 5 functions as a full agonist at the D3 and
D2 receptors in vivo, consistent with the
two-site competition curve observed in the
[3H]spiperone binding assay for 1 and 5
(Supporting Information).
Furthermore, the in vivo data suggested that
5 is bioavailable. Unlike 1 and 5, the
dose-response curves for 6 and 12 induced
yawning were relatively flat and failed
to reach significance during the initial 30 min
observation period. While significant levels of
yawning induced by 6 and 12 were observed
after 60 min, the dose-response curves for both
compounds remained relatively flat. Moreover, 6
and 12 failed to induce changes in body
temperature over the initial hour of
observation, an effect that is indicative of D2
agonist activity. Together, the low levels of
yawning, combined with the absence of any
hypothermic effect, suggested two possibilities:
(1) 6 and 12 function as weak partial agonists
at the D3 receptor, with no detectable agonist
activity at the D2 receptor, or (2) they are
simply not bioavailable. To investigate these
two possibilities, we next evaluated the ability
of 12 to alter compound 1-induced yawning
and hypothermia and the data are shown in Figure
4. Similar to the effects of 12 alone, but
unlike the effects of D3-selecitve antagonists,
low levels of yawning were observed
during the initial 30 min after administration
of either 10.0 or 32.0 mg/kg of 12.
Interestingly, this effect appeared to
persist upon administration of low doses of 1 as
significant increases in yawning were
observed when rats were pretreated with 12 (10.0
or 32.0 mg/kg). However, 12 resulted in a
dose-dependent decrease in the amount of
yawning observed following the maximally
effective dose of 1 at 0.1 mg/kg. No significant
effects of 12 were observed at higher doses of 1
(0.32 and 1 mg/kg). These data suggested that 12
is capable of antagonizing the D3-mediated
effects of 1. However, the profile of activity
for 12 is different from that observed for
selective D3 antagonists, which generally
produce selective rightward and/or downward
shifts of the ascending limb of the
yawning dose-response curve for
D3-preferring agonists without increasing the
amount of yawning observed at low
doses.
In fact, the effects of 12 alone, and in
combination with 1, suggest that it is more
similar to the partial agonist, aripiprazole,34
than an antagonist. Moreover, 12 failed to alter
the induction of hypothermia by 1, an effect
that is indicative of D2 agonist activity, which
can be reliably blocked by both selective and
nonselective D2 antagonists.
Together, our data provide evidence that 12
is a partial agonist at the D3 receptor with no
detectable agonist or antagonist activity at the
D2 receptor, thus possessing a novel in vivo
functional profile. In summary, a series of
enantiomerically pure pramipexole derivatives
have been designed, synthesized, and evaluated
for their binding and selectivity to the D3,
D1-like, and D2 receptor. This led to the
identification of several potent and highly
selective D3 ligands with excellent aqueous
solubility. Our in vivo functional evaluations
showed that while 5 functions as a full D3
agonist, 12 behaves as a selective D3 partial
agonist with no activity at the D2 receptor.
Further in vivo studies are underway to evaluate
the therapeutic potential of 12 for the
treatment of drug abuse and other indications.
The results will be reported in due course.
-Collins
C, Truccone A et al. Pro-erectile Effects of
Dopamine D2-like Agonists are Mediated by the D3
Receptor in Rats and Mice. J J Pharmacol Exp
Ther 2009
-Canales JJ,
Iversen SD Psychomotor activating effects
mediated by dopamine D2 and D3 receptors in the
nucleus accumbens Pharmacol Biochem Behav 2000;
67;161-168
-Canales
JJ, Iversen SD Dynamic dopamine receptor
interactions in the core and shell of nucleus
accumbens differentially coordinate the
expression of unconditioned motor behaviors
Synapse 2000; 36; 297-306
-Collins GT, JM
Witkin et al Dopamine agonist-induced
yawning in rats: a dopamine d3 receptor mediated
behavior. J Pharmacol Exp Ther
2005;314(1):310-319
-Collins GT, Newman
AH,Woods JH et al.Yawning and hypothermia in
rats: effects of dopamine D3 and D2 agonists and
antagonists. Psychopharmacology
(Berl).
2007;193(2):159-170
-Collins GT. et
al. Food restriction alters
pramipexole-induced yawning, hypothermia, and
locomotor activity in rats: Evidence for
sensitization of dopamine D2 receptor-mediated
effects. JEPT 2008;325:691-697
-Collins
GT et al. Narrowing in on compulsions:
dopamine receptor functions Exp Clin
Psychopharmacol 2008,16(4):498,502
-Chen J, Gregory T.
Collins, et al. Design, Synthesis, and
Evaluation of Potent and Selective Ligands for
the Dopamine 3 (D3) Receptor with a Novel in
Vivo Behavioral Profile. J Med. Chem.,
2008;51(19):5905&endash;5908
-Collins GT,
Truong YN, et al. Behavioral sensitization
to cocaine in rats: evidence for temporal
differences in dopamine D(3) and D (2) receptor
sensitivity. Psychopharmacology (Berl).
2011;215(4):609-20.
Li SM,
Collins GT et al. Yawning and locomotor
behavior induced by dopamine receptor agonists
in mice and rat. Behav Pharmacol. 2010; 21(3):
171&endash;181.
Based on studies in the
rat, Sokoloff et al. have made the valuable
suggestion that the D3 receptor is a
particularly important target for antipsychotics
in the mesolimbic DA system. These study in the
human demonstrates that the distribution of D3
receptors and D3 mRNA-bearing neurons is
consistent with relative segregation of the D3
subtype to the limbic striatum as well as it
primary and secondary targets and many sources
of its afferents.
