Design of novel hexahydropyrazinoquinolines as potent and selective dopamine D3 receptor ligands with improved solubility

Article abstract of DOI:10.1016/j.bmcl.2005.09.053

We have recently reported hexahydropyrazinoquinolines as a new class of dopamine 3 (D₃) receptor ligands with high-affinity to the D ₃ receptor and excellent selectivity over the closely related D ₁-like and D₂-

Full text of DOI:10.1016/j.bmcl.2005.09.053

Bioorganic & Medicinal Chemistry Letters 16 (2006) 443–446
Design of novel hexahydropyrazinoquinolines as potent
and selective dopamine D3 receptor ligands with improved solubility
Jianyong Chen,a Ke Ding,a Beth Levantb and Shaomeng Wanga,*
aDepartments of Internal Medicine and Medicinal Chemistry, University of Michigan,
1500 East Medical Center Drive, Ann Arbor, MI 48109-0934, USA
bDepartment of Pharmacology, Toxicology, and Therapeutics, University of Kansas Medical Center,
Kansas City, KS 66160-7417, USA
Received 7 July 2005; revised 7 September 2005; accepted 8 September 2005
Available online 14 November 2005
Abstract—We have recently reported hexahydropyrazinoquinolines as a new class of dopamine 3 (D3) receptor ligands with high-
affinity to the D3 receptor and excellent selectivity over the closely related D1-like and D2-like receptors. However, our previously
reported most potent and selective D3 ligands have poor aqueous solubility, which greatly hinders our in vivo studies aimed at eval-
uation of their therapeutic potential in animal models. In this study, we wish to report the design, synthesis, and evaluation of a
series of new hexahydropyrazinoquinolines as D3 ligands with improved solubility. Among them, compound 4g has a Ki value
of 9.7 nM for the D3 receptor and displays a selectivity of >5000 and 466 times over the D1-like and D2-like receptors, respectively.
Importantly, the hydrochloride salt form of compound 4g has a good aqueous solubility (>50 mg/mL) and represents a promising
D3 ligand for further in vivo evaluations of its therapeutic potential for the treatment of drug abuse, restless legs syndrome, schizo-
phrenia, ParkinsonÕs disease, and depression.
Ó 2005 Elsevier Ltd. All rights reserved.
Recent studies1–5 have suggested that the dopamine 3
subtype receptor (D3) is a promising therapeutic target
for a variety of conditions including drug abuse, restless
legs syndrome, schizophrenia, ParkinsonÕs disease, and
depression. Potent and selective D3 ligands may have
therapeutic potential for the treatment of these condi-
tions. Accordingly, there is a strong research interest
in the design of potent and selective D3 ligands.6
tionships for the ÔheadÕ and ÔlinkerÕ regions. Those
modifications have yielded compounds 2 and 3 with
high affinity for the D3 receptor and excellent selectiv-
ity over the D1-like and D2-like receptors. However,
compounds 2 and 3 have a limited aqueous solubility,
which greatly hinders our in vivo studies for evalua-
tion of their therapeutic potential in animal models.
Of note, many of those potent and selective D3 ligands
recently reported by other laboratories are also highly
hydrophobic in nature and will likely encounter a sim-
ilar solubility problem.5,9–11 In a recent study, efforts
have been made toward improvement of the aqueous
solubility in the design of potent and selective D3
ligands.12 In this paper, we report the design, synthe-
sis, and evaluation of new analogues of compounds
2 and 3 with modifications at the ÔtailÕ region with a
goal to derive potent and selective D3 ligands with
improved aqueous solubility.
Our laboratory has recently reported the design, syn-
thesis, and evaluation of hexahydropyrazinoquinolines
as a new class of potent and selective D3 ligands.7,8
Based upon its chemical structure, the initial lead
compound 1 was divided into three structural regions,
the tricyclic hexahydropyrazinoquinoline core structure
as the ÔheadÕ, the naphthyl ring as the ÔtailÕ, and the
linker between the ÔheadÕ and the ÔtailÕ groups. In
our previous studies,7,8 we have designed and synthe-
sized analogues to explore the structure–activity rela-
Our previous molecular modeling analysis showed that
the hydrophobic naphthyl ÔtailÕ occupies a large binding
pocket primarily formed by hydrophobic residues in the
D3 receptor.8 For this reason, a hydrophobic tail may be
required for achieving high binding affinities to the D3
receptor. Hence, in our design, we have chosen to
Keywords: Hexahydropyrazinoquinolines; Dopamine D3 receptor;
Ligands.
*
Corresponding author. Tel.: +1 734 615 0362; fax: +1 734 647
9647; e-mail: shaomeng@umich.edu
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.09.053

444
J. Chen et al. / Bioorg. Med. Chem. Lett. 16 (2006) 443–446
N
O
N
N
A
B
H
Core structure
(Head)
Linker
Tail
1
MeO
MeO
N
O
N
O
N
N
NH
NH
3
2
MeO
N
O
N
R
NH
4
N
N
N
4a
4b
4c
4d
R=
N
N
N
N
N
4e
4f
4g
4h
replace the naphthyl ring in our lead compounds with
either a quinolinyl ring, or an isoquinolinyl ring, or qui-
noxalinyl ring. We reasoned that these ring systems are
still primarily hydrophobic in nature but have improved
aqueous solubility over the naphthyl ring. Thus, the
resulting compounds (4a–h) will have better aqueous
solubility than compounds 2 and 3. Since enantiomers
2 and 3 have very similar binding affinities to the D3
receptor and also very similar selectivities over the D1-
like and D2-like receptors,8 we have decided to employ
the racemic template for rapid exploration of the struc-
ture–activity relationship in the tail region. A series of
new analogues 4a–h were synthesized and evaluated at
the dopamine receptors.
using EDCI as coupling reagent in dichloromethane to
produce the target compounds 4a–4h.17
In view of the numerous dopamine receptors at which
binding affinity might be determined for these com-
pounds, we have designed our screening protocol to
make the most important comparison (i.e., D2-like/
D3) as well as compare between the major subclasses
of dopamine receptors (i.e., D1-like/D2-like, and D1-
like/D3). Accordingly, in this study, all the new com-
pounds were evaluated for their binding affinities at
the D1-like, D2-like and D3 receptors using previously
established methods.13–15 Because the affinities of com-
pounds at the dopamine receptor subtypes have been
shown to vary depending on the in vitro assay condi-
tions used and the source of receptors (i.e., human or
rat and the expression system used),5 the assay condi-
tions used in these assays were designed to favor ago-
nist binding and used receptors expressed in their
native tissue, brain. Our receptor binding assay enable
selective assessment of each dopamine receptor subtype
of interest by using brain regions that predominantly
express the receptor of interest (e.g., the caudate puta-
The synthesis for compounds 4a–h is straightforward
and outlined in Scheme 1. Briefly, compound 5 was syn-
thesized using our previously published method.7 Com-
pound 6 was obtained by condensation of compound
5 with trans-(4-tert-butoxycarbonylamino-cyclohexyl)-
acetic acid, followed by removal of the Boc-protecting
group and reduction with lithium alumina hydride.
Compound 6 was then coupled with the requisite acids
MeO
MeO
MeO
b
a
N
N
N
N
O
NH
N
N
R
H
NH
6
5
2
4a-4h
Scheme 1. Synthesis of hexahydropyrazinoquinolines. Reagents and conditions: (a) i. trans-(4-tert-butoxycarbonylamino-cyclohexyl)-acetic acid,
EDCI, HOBT, N,N-diisopropylethylamine, CH2Cl2, rt, 12 h; ii. 4 M HCl in dioxane, rt, 4 h; iii. LiAlH4, THF, reflux, 2 h; (b) RCO2H, EDCI, HOBT,
N,N-diisopropylethylamine, CH2Cl2, rt, 2 h.

J. Chen et al. / Bioorg. Med. Chem. Lett. 16 (2006) 443–446
445
men, which expresses very low levels of D3 sites, for the
D1-like and D2-like receptor assays) and/or in vitro
conditions that affect the binding of the radioligand
to a particular binding site (e.g., [3H]PD 128907 bind-
to compound 4a further confirms the importance of
the hydrophobic interaction between the naphthyl ring
in compound 4a and the hydrophobic residues in the
D3 receptor, as suggested in our molecular modeling
studies.8
ing to D2 sites is disfavored in the absence of Mg2+
,
thus enabling selective labeling of D3 sites in ventral
striatal membranes).14 These assay conditions have
been extensively validated and compared with those
of other in vitro assay systems.13 Of note, the non-se-
lective antagonist haloperidol, and the D3-selective
antagonists (+)-S14297 and U99194A exhibited binding
affinities and selectivities in our assays and are highly
similar to those previously reported using cloned hu-
man receptors expressed in transfected cells.5,16 The
binding affinities for compounds 4a–h, together with
reference compounds haloperidol, (+)-S14297, and
U99194A at the D3, D1-like, and D2-like receptors
determined in our assay conditions are summarized in
Table 1.
Based upon our previous modeling results,8 although
the naphthyl ring in compound 4a interacts with sev-
eral hydrophobic residues, part of the ring B in the
naphthyl ring is exposed to solvent. It suggests that
a nitrogen atom may be inserted into the ring B in
the naphthyl ring without causing significant reduction
in binding affinity to the D3 receptor. To test this
idea, we have synthesized compounds 4g and 4h, in
which the 2-naphthyl ring in 4a is replaced by either
6-quinolinyl or 8-quinolinyl ring. As can be seen from
Table 1, although compound 4h is >10 times less po-
tent than 4a in its binding affinity to D3 and has no
selectivity between the D3 receptor and D2-like recep-
tors, compound 4g is a potent D3 ligand with a Ki
value of 9.7 nM. Furthermore, compound 4g has an
excellent selectivity over the D1-like and D2-like recep-
tors, being >5000 times between the D3 and D1-like
receptors, and 466 times between the D3 and D2-like
receptors. Importantly, compound 4g was found to
have a much better aqueous solubility than compound
4a in our solubility testing. For example, the hydro-
chloride salt form of 4g has an aqueous solubility
greater than 50 mg/mL, whereas the hydrochloride salt
form of 4a has an aqueous solubility less than 1 mg/
mL. Therefore, compound 4g represents a potent,
selective, and soluble D3 ligand for our further in vivo
studies.
Replacement of the naphthyl ring by a 3-isoquinolinyl,
3-quinolinyl, or 2-quinolinyl ring results in compounds
4b, 4c, and 4d in which a carbon atom in ring A of
the naphthyl ring is replaced by a nitrogen atom at
three different positions. These replacements reduce
the binding affinity to the D3 receptor by 10, 10,
and 36 times, respectively, as compared to the initial
lead compound 4a. These replacements also signifi-
cantly reduce the selectivity of these compounds for
the D3 receptor versus the D2-like receptors. To inves-
tigate the effect of the substitution position for the
aromatic ring, we changed the substitution position
from b- to a-position for the quinolinyl in compound
4c, which resulted in compound 4e. This change re-
sults in an improvement of three times in both bind-
ing affinity to the D3 receptor and selectivity over
the D2-like receptors as compared to compound 4c.
But compound 4e is still three times less potent than
4a and also much less selective over the D2-like recep-
tors than 4a. Replacement of the naphthyl ring by a
2-quinoxalinyl ring resulted in compound 4f, which
is significantly less potent and selective than com-
pound 4a. The significant reduction in binding affinity
and selectivity for these new analogues as compared
In summary, a series of new hexahydropyrazinoquino-
lines were designed, synthesized, and evaluated as D3 li-
gands. Among them, compound 4g has a Ki value of
9.7 nM for the D3 receptor and displays a selectivity of
>5000 and 466 times over the D1-like and D2-like recep-
tors, respectively. Importantly, compound 4g has a good
aqueous solubility (>50 mg/mL) and represents a prom-
ising D3 ligand for further in vivo evaluations of its ther-
apeutic potential for the treatment of drug abuse and
several other conditions.
Table 1. Binding affinities at the D1-like, D2-like, and D3 receptors in binding assays using rat brain
Compound
Ki SEM (nM)
Selectivity
D1-like [3H]SCH 23390
D2-like [3H]spiperone
D3 [3H]PD 128907
D1-like/D3
D2-like/D3
1
4a
>250,000
>50,000
>50,000
>250,000
244 59
5.1 0.67
48 1.8
51 6.5
183 17
16 2.7
152 24
9.7 1.8
87 12
>1000
>9804
>1041
>980
301
>1000
717
3.9
27
3660 594
188 39
4b19
4c
4d
>50,000
1360 345
4760 797
1470 132
2830 171
4500 841
162 48
55,000 4640
19,700 2130
24,800 2800
>50,000
26
94
4e
1259
163
4f
4g
4h19
19
>5154
>575
466
1.9
0.30
26
>50,000
42 3.4
Haloperidol
S12947
U99194A
7.2 1.7
256 52
11,200 513
24 3.114
9.8 1.6
498 72
1.75
>10,000
>200
>100,000
>100,000
22
Validated in vitro assay conditions were optimized to enable selective assessment of each receptor of interest and to favor agonist binding.18
Data represent means SEM of 3–5 independent determinations.

446
J. Chen et al. / Bioorg. Med. Chem. Lett. 16 (2006) 443–446
1H), 6.07 (d, J = 8.0 Hz, 1H), 6.61 (d, J = 3.0 Hz, 1H),
References and notes
6.68 (dd, J = 9.0, 3.0 Hz, 1H), 6.77 (d, J = 9.0 Hz, 1H),
7.60–7.67 (m, 1H), 7.80–7.85 (m, 1H), 7.93 (d, J = 8.5 Hz,
1H), 8.17 (d, J = 8.5 Hz, 1H), 8.58 (d, J = 2.2 Hz, 1H),
9.25 (d, J = 2.2 Hz, 1H). 13C NMR (75 MHz, CDCl3): d
27.6, 27.8, 32.3, 33.5, 34.1, 35.9, 47.7, 49.8, 53.9, 55.8, 56.0,
56.9, 60.2, 112.6, 114.0, 115.3, 126.4, 127.3, 127.8, 127.9,
129.1, 129.8, 131.6, 135.8, 140.9, 148.5, 149.6, 152.6, 165.3.
18. [3H]SCH 23390 binding assays for D1-like dopamine
receptors were performed as previously described in
detail15 using membranes prepared from the caudate-
putamen of adult male Sprague–Dawley rats (Harlan,
Indianapolis, IN). All compounds were dissolved in
100% EtOH at a concentration up to 5 mM. The assay
buffer was 50 mM Tris–HCl, 5 mM KCl, 2 mM MgCl2,
and 2 mM CaCl2, pH 7.4, at 23 °C; the concentration of
[3H]SCH 23390 (73 Ci/mmol; Amersham) was 0.3 nM;
and non-specific binding was determined in the presence
of 1 lM (+)-butaclamol. SigmaPlot was used to deter-
mine Ki values using the KD value for [3H]SCH 23390 of
0.3 nM.15 [3H]spiperone binding assays for D2-like
dopamine receptors were performed as previously
described in detail and as described for [3H]SCH
23390, except that the concentration of [3H]spiperone
(24 Ci/mmol; Amersham) was 0.2 nM.13,15 Ki values
were determined using the KD value for [3H]spiperone
of 0.1 nM.15 [3H]PD 128907 binding assays D3-like
dopamine receptors were performed as previously
described in detail14,15 using ventral striatal (nucleus
accumbens and olfactory tubercles) membranes prepared
in assay buffer (50 mM Tris, 1 mM EDTA; pH 7.4 at
23 °C). The concentration of [3H]PD 128907 was
0.3 nM; 116 Ci/mmol; (Amersham, Arlington Heights,
IL) and non-specific binding was defined by 1 lM
spiperone. Ki values were determined using the KD
value for [3H]PD 128907 of 0.3 nM.15 Haloperidol and
U99194A were purchased from Sigma (St. Louis, MO).
(+)-S12497 was the generous gift of Dr. Mark Millan of
Institut de Recherches Servier.
1. Joyce, J. N. Pharmacol. Ther. 2001, 90, 231.
2. Pilla, M.; Perachon, S.; Sautel, F.; Garrido, F.; Mann, A.;
Wermuth, C. G.; Schwartz, J.-C.; Everitt, B. J.; Sokoloff,
P. Nature 1999, 400, 371.
3. Koob, G. F.; Caine, S. B. Nat. Med. 1999, 993.
4. Montplaisir, J.; Nicolas, A.; Denesle, R.; Gomez-Mancil-
la, B. Neurology 1999, 52, 938.
5. Levant, B. Pharmacol. Rev. 1997, 49, 231.
6. Newman, A. H.; Grundt, P.; Nader, M. A. J. Med. Chem.
2005, 48, 3663.
7. Ji, M.; Chen, J.; Ding, K.; Wu, X.; Varady, J.; Levant, B.;
Wang, S. Bioorg. Med. Chem. Lett. 2005, 15, 1701.
8. Ding, K.; Chen, J.; Ji, M.; Wu, X.; Varady, J.; Yang, C.-
Y.; Lu, Y.; Deschamps, J. R.; Levant, B.; Wang, S. J.
Med. Chem. 2005, 48, 3171.
9. Bettinetti, L.; Schlotter, K.; Hubner, H.; Gmeiner, P. J.
Med. Chem. 2002, 45, 4594.
10. Leopoldo, M.; Berardi, F.; Colabufo, N. A.; De Giorgio,
P.; Lacivita, E.; Perrone, R.; Tortorella, V. J. Med. Chem.
2002, 45, 5727.
11. Hackling, A.; Ghosh, R.; Perachon, S.; Mann, A.; Holtje,
H. D.; Wermuth, C. G.; Schwartz, J. C.; Sippl, W.;
Sokoloff, P.; Stark, H. J. Med. Chem. 2003, 46, 3883.
12. Grundt, P.; Carlson, E. E.; Cao, J.; Bennett, C. J.;
McElveen, E.; Taylor, M.; Luedtke, R. R.; Newman, A.
H. J. Med. Chem. 2005, 48, 839.
13. Levant, B.; Grigoriadis, D. E. J. Pharmacol. Exp. Ther.
1992, 262, 929.
14. Bancroft, G. N.; Morgan, K. S.; Flietstra, R. J.; Levant, B.
Neuropsychopharmacology 1998, 18, 305.
15. Levant, B. Characterization of Dopamine Receptors. In
Current Protocols in Pharmacology; Ferkany, J., Enna, S.
J., Eds.; John Wiley and Sons: New York, 1998; pp 1.6.1–
1.6.16.
16. Audinot, A.; Newman-Tancredi, A.; Gobert, A.; Rivet,
J.-M.; Brocco, M.; Lejuene, F.; Gluck, L.; Desposte, I.;
Bervoets, K.; Dekeyne, A.; Millan, M. J. J. Pharmacol.
Exp. Ther. 1998, 287, 187.
19. Competition data for these compounds in the [3H]spipe-
rone binding assay was consistent with interactions at two
binding sites as analyzed by SigmaPlot, suggesting
potential agonist activity at D2-like receptors. Ki values
presented here are for the high-affinity component.
17. The NMR data for compound 4c: 1H NMR (300 MHz,
CDCl3): d 1.10–1.40 (m, 5H), 1.45–1.57 (m, 2H), 1.72–2.00
(m, 5H), 2.16–2.30 (m, 3H), 2.41–2.47 (m, 2H), 2.72–3.05
(m, 6H), 3.67–3.74 (m, 1H), 3.76 (s, 3H), 4.00–4.10 (m,