Multiparameter exploration of piperazine derivatives as δ-opioid receptor agonists for CNS indications

Article abstract of DOI:10.1016/j.bmcl.2011.11.088

A novel series of piperazine derivatives exhibits sub-nanomolar binding and enhanced subtype selectivity as δ-opioid agonists. The synthesis and SAR are described as well as the application of computational models to improve in vitro ADME and safety prope

Full text of DOI:10.1016/j.bmcl.2011.11.088

Bioorganic & Medicinal Chemistry Letters 22 (2012) 1169–1173
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Multiparameter exploration of piperazine derivatives as d-opioid receptor
agonists for CNS indications
⇑
John P. McCauley Jr. , Cathy L. Dantzman, Megan M. King, Glen E. Ernst, Xia Wang, Kelly Brush,
William E. Palmer, William Frietze, Donald W. Andisik, Valerie Hoesch, Kenneth Doring,
⇑
James Hulsizer, Khanh H. Bui, Jay Liu, Thomas J. Hudzik, Steven S. Wesolowski
AstraZeneca Pharmaceuticals, 1800 Concord Pike, Wilmington, DE 19850, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 7 December 2011
A novel series of piperazine derivatives exhibits sub-nanomolar binding and enhanced subtype selectivity
as d-opioid agonists. The synthesis and SAR are described as well as the application of computational
models to improve in vitro ADME and safety properties suitable for CNS indications, specifically micro-
somal clearance, permeability, and hERG channel inhibition.
Keywords:
Delta opioid agonist
hERG
Ó 2011 Elsevier Ltd. All rights reserved.
Permeability
Microsomal clearance
ADME predictive models
The delta opioid receptor (dOR) is a target for several disease
areas with significant unmet medical need, including pain and
mood disorders.1 Delta, as well as the two other major opioid recep-
an aromatic aldehyde 3 gives a secondary alcohol 4 which is con-
verted to bromide 5. Subsequent displacement by piperazine gives
the racemic amine derivative 6. This amine can be resolved classi-
cally through the selective crystallization of chiral tartrate salt
tor subtypes, mu (l) and kappa (j), belongs to the G-protein cou-
pled receptor (GPCR) family class. While the endogenous selective
ligands for the d-opioid receptor are the peptide agonists
Met- and Leu-enkephalin, morphine is well known as a powerful
O
N
l
agonist that also binds to the d and
j subtypes.
Although efficacious, morphine’s side effects such as respiratory
OH
HO
O
depression and addiction leave ample room for improved thera-
peutics. Compounds that selectively target the d-opioid receptor
may have benefits without some of the drawbacks of non-selective
drugs, although other mechanisms such as binding to heterorecep-
tors and receptor internalization have been postulated to account
for additional side effects.2 Prior research in this area has identified
selective d-opioid agonists, including derivatives of the endoge-
nous peptides, morphine-derived structures, as well as diverse
small molecule scaffolds beginning with SNC-80 (Fig. 1).3
We have expanded previous work on a series of potent pipera-
zine derivatives (1, Fig. 2).4 The activity and selectivity of this ser-
ies remain attractive, and the synthesis, SAR, and computational
modeling toward compounds with suitable CNS profiles are
described.
O
N
N
H
N
Morphine
SNC-80
O
HO
N
N
O
HO
Drug candidates were prepared by the application of a synthetic
scheme that enables diversity at R1, R2, and R3 via a common inter-
mediate (Scheme 1). The addition of an aryl Grignard reagent 2 to
H
N
N
H
TAN-67
ADL-5859
⇑
Corresponding authors.
E-mail address: steven.wesolowski@astrazeneca.com (S.S. Wesolowski).
Figure 1. Morphine and subtype-selective d-opioid agonists.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.11.088

1170
J. P. McCauley Jr. et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1169–1173
R3
derivatives. Alternatively, preparative chiral stationary phase li-
quid or supercritical fluid chromatography can be performed at
this (or a later) stage to give the single enantiomer intermediates
7. Treatment with HCl yields carboxylic acid 8, and BOC protection
affords the versatile intermediate 9. Selective protection and
deprotection schemes are applied to give a series of compounds
10 that possess systematically varied substitutions at three differ-
ent regions on the drug molecule (R1, R2, and R3).
R1
N
N
R2
Either stereoisomer can exhibit potent binding at dOR. In this
particular series, when both were potent binders, the R enantiomer
typically showed slightly greater potency than its S enantiomer.
1
Figure 2. Generic structure of the piperazine series 1.
O
O
N
O
N
Y
N
+
H
X
SOBr2
Y
Y
X
X
X
MgI
X
X
O
X = CH or N
Y = H or NO2
2
Br
OH
4
5
3
O
O
O
N
N
Y
HO
Y
X
piperazine
X
Y
X
X
resolve
X
6N HCl
reflux
X
N
N
N
6
7
8
N
H
N
H
N
H
O
R3
HO
Z
Y
(BOC)2O
X
X
X
X
N
X = CH or N
Z = H or N-R'
N
N
X = CH or N
Y = H or NO2
N
10
R2
O
9
O
Versatile intermediate
Scheme 1. Synthesis of chiral intermediate 9 which enables sequential installation of diverse R1, R2, and R3 groups.
Table 1
dOR binding affinity,
monolayers, and hERG inhibition for piperazines 11–26
l
OR selectivity, microsomal clearance, permeability through MDCK–MDR1 (Madin–Darby canine kidney cells transfected with human MDR1 gene) cell
R3
R1
N
N
R2
Cmpd
R1
R2
R3
dORa IC50 (nM)
l
/da
hCLintb
150
(l
l/min/mg)
Pappc (nm/s)
<1
hERGd IC50
(lM)
O
O
11
0.29
5200
11
N
N
H
F

J. P. McCauley Jr. et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1169–1173
1171
M)
Table 1 (continued)
Cmpd
R1
R2
R3
dORa IC50 (nM)
l
/da
hCLintb
36
(l
l/min/mg)
Pappc (nm/s)
<1
hERGd IC50
(l
O
O
O
O
HO
12
0.27
11000
480
15
N
N
H
F
F
O
13
0.90
49
<1
3.6
N
N
H
O
O
O
N
14
15
16
0.068
>100
1.1
>120000
<100
260
160
26
NT
NT
<1
20
N
N
H
O
N
18
N
N
N
S
H
O
O
>9400
>30
N
H
O
O
O
17
18
23
33
<4
85
NT
<1
13
28
N
N
H
O
N
O
6.6
>1500
N
N
N
N
N
H
O
O
O
O
O
19
20
0.35
0.68
>5800
5100
100
110
2.2
3.0
14
11
N
N
N
H
O
O
S
N
H
N
S
21
0.78
11000
430
<1
8.7
HN
O
O
O
O
O
O
O
N
22
23
0.071
1.3
6200
2700
210
21
<1
5.4
7.3
N
N
N
N
H
H
N
O
O
S
N
1.3
S
N
24
1.4
>6000
650
4.3
9.9
HN
O
O
O
O
O
O
N
25
26
0.26
0.70
14000
17
<4
<1
<1
>33
>33
N
N
N
H
O
O
N
H
N
>12000
N
H
N
H
a
The dOR and
lOR binding assays employed iodinated Deltorphin II and FK33-824, respectively, as the radioligand and membranes were prepared from HEK 293 cells
over-expressing either dOR or
l
OR. Quantification of the receptor-associated radioligand was done with a filtration format and an SPA format.
b
Intrinsic clearance of compound in human liver microsomes.
The permeability through MDCK–MDR1 cell monolayers as measured from the apical side to the basolateral side in a transport assay format.
Voltage-dependent potassium channel encoded by the human ether-a-go-go related gene.
c
d

1172
J. P. McCauley Jr. et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1169–1173
display of functional agonism led to a focus on the set of R enanti-
omers shown in Table 1.
In addition, ADME predictive models were built and applied as
filters (Fig. 3). The ADME models utilized a binary QSAR methodol-
ogy with six calculated properties (molecular weight, polar surface
area, ClogP, number of hydrogen bond donors, number of hydro-
gen bond acceptors and number of rotatable bonds) using the
MOE software package.5 The models provided binary (pass/fail)
predictions for passive permeability and microsomal clearance.
For the hERG model, a support vector machine (SVM) approach
yielded better pass/fail predictions for test sets and was used for
predictions. Whereas models for microsomal clearance and hERG
were built from robust distributions of experimental data, the
model for passive permeability was of weaker predictive value
due to a narrower range and distribution of experimental data
within the series.
When considered together with physical properties, synthetic
feasibility, and novelty, a total of 37 new compounds were priori-
tized to be made from the set of ca. 300,000 virtual compounds.
Several newly synthesized and tested compounds simultaneously
met hERG and multiple ADME criteria. Thirty six of the compounds
obtained measured CLint and hERG, and the predicted classifica-
tions were correct for 34/36 (94%), and 30/36 (83%), respectively.
Although all the compounds chosen for synthesis were also pre-
dicted to have measurable passive permeability (>1 nm/s), only 5
of the 16 tested in this assay were indeed >1 nm/s. Furthermore,
dOR potency was sacrificed for many compounds, particularly for
the newly explored substituents outside the training sets. The five
Figure 3. Schematic of structure–searchable database. Prioritized target 31 from
the enumeration library is shown with ADME and hERG predictions as well as
computed physical properties.
More significantly, the binding potency of the R enantiomer tended
to be reflected in functional agonism, whereas the agonism was
lost for the S enantiomer in many instances. This more consistent
Table 2
dOR binding affinity,
lOR selectivity, microsomal clearance, permeability through MDCK–MDR1 (Madin–Darby canine kidney cells transfected with human MDR1 gene) cell
monolayers, and hERG inhibition for piperazines 27–31. Predicted values using classification models for clearance, permeability, and hERG are shown in italics. Predictions
consistent with later experimental results are highlighted in green and those inconsistent with experimental results are in red and labeled
R3
R1
N
N
R2
Cmpd R1
R2
R3
dORa
IC50
(nM)
l
/da
Predicted
hCLintb
min/mg)
hCLintb
l/min/
mg)
Predicted
Pappc (nm/
s)
Pappc
(nm/s) hERGd IC50
M)
Predicted
hERGd
IC50
(lM)
(
ll/
(l
(l
O
O
N
27
28
0.094
4.1
15000 >50e
>2200 >50
38
>1
>1
1.2
1.8
>10
>10
17
11
N
N
H
O
S
O
O
N
53
N
N
O
H
O
O
O
N
29
30
4.7
2.0
5.2
>2100 <50
>5100 <50
>1900 <50
40
33
20
>1e
>1
<1
>10
>10
>10
15
N
N
N
N
N
S
N
1.2
2.5
13
N
31
>1
>33
N
a
The dOR and
l
OR binding assays employed iodinated Deltorphin II and FK33-824, respectively, as the radioligand and membranes were prepared from HEK 293 cells
OR. Quantification of the receptor-associated radioligand was done with a filtration format and an SPA format.
over-expressing either dOR or
l
b
Intrinsic clearance of compound in human liver microsomes.
c
d
e
The permeability through MDCK–MDR1 cell monolayers as measured from the apical side to the basolateral side in a transport assay format.6
Voltage-dependent potassium channel encoded by the human ether-a-go-go related gene.
Incorrectly classified by the predictive model.

J. P. McCauley Jr. et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1169–1173
1173
compounds that met potency and most property criteria are shown
References and notes
in Table 2. Compound 27 (IC50 = 0.094 nM) is one of the most po-
tent compounds in the series, while compound 31 is notable for
its weak activity opposite the hERG channel (IC50 >33 lM). The
1. (a) Millan, M. Pain 1986, 27, 303; (b) Narita, M.; Kuzumaki, N.; Narita, M.;
Kaneko, C.; Hareyama, N.; Miyatake, M.; Shindo, K.; Miyoshi, K.; Nakajima, M.;
Nagumo, Y.; Sato, F.; Wachi, H.; Seyama, Y.; Suzuki, T. J. Neurochem. 2006, 97,
1369; (c) Berrocoso1, E.; Sánchez-Blázquez, P.; Garzón, J.; Mico1, J. A. Curr.
Pharm. Design 2009, 15, 1612.
2. (a) Yekkirala, A.; Kalyuzhny, A.; Portoghese, P. ACS Chem. Neurosci. 2010, 1, 146;
(b) Pradhan, A.; Becker, J.; Scherrer, G.; Tryoen-Toth, P.; Filliol, D.; Matifas, A.;
Massotte, D.; Gavériaux-Ruff, C.; Kieffer, B. PLoS ONE 2009, 4, e5425; (c) Dondio,
G. Il Farmaco 2000, 55, 178; (d) Broom, D.; Jutkiewicz, E.; Folk, J.; Traynor, J.;
Rice, K.; Woods, J. Psychopharmacology 2002, 164, 42.
3. (a) Calderon, S.; Rothman, R.; Porreca, F.; Flippen-Anderson, J.; McNutt, R.; Xu,
H.; Smith, L.; Bilsky, E.; Davis, P.; Rice, K. J. Med. Chem. 1994, 37, 2125; (b)
Nagase, H.; Kawai, K.; Hayakawa, J.; Wakita, H.; Mizusuna, A.; Matsuura, H.;
Tajima, C.; Takezawa, Y.; Endoh, T. Chem. Pharm. Bull. 1998, 46, 1695; (c) Le
Bourdonnec, B.; Windh, R.; Ajello, C.; Leister, L.; Gu, M.; Chu, G.; Tuthill, P.;
Barker, W.; Koblish, M.; Wiant, D.; Graczyk, T.; Belanger, S.; Cassel, J.; Feschenko,
M.; Brogdon, B.; Smith, S.; Christ, D.; Derelanko, M.; Kutz, S.; Little, P.; DeHaven,
R.; DeHaven-Hudkins, D.; Dolle, R. J. Med. Chem. 2008, 51, 5893.
4. (a) Plobeck, N.; Delorme, D.; Wei, Z.-Y.; Yang, H.; Zhou, F.; Schwarz, P.; Gawell,
L.; Gagnon, H.; Pelcman, B.; Schmidt, R.; Yue, S. Y.; Walpole, C.; Brown, W.; Zhou,
E.; Labarre, M.; Payza, K.; St-Onge, S.; Kamassah, A.; Morin, P.-E.; Projean, D.;
Ducharme, J.; Roberts, E. J. Med. Chem. 2000, 43, 3878; (b) Jones, P.; Griffin, A. M.;
Gawell, L.; Lavoie, R.; Delorme, D.; Roberts, E.; Brown, W.; Walpole, C.; Xiao, W.;
Boulet, J.; Labarre, M.; Coupal, M.; Butterworth, J.; St-Onge, S.; Hodzic, L.; Salois,
D. Bioorg. Med. Chem. Lett. 2009, 19, 5994.
R2 pyridyls (30–31) surfaced as having the best compromise of
overall properties. Although permeability was generally improved
with this second set of compounds, the magnitude of the enhance-
ment was modest. Furthermore, compounds that exhibited moder-
ate permeability were measured to be P-glycoprotein efflux
substrates6 and are unlikely to sustain sufficient brain levels for
efficacy. Thus, none of these compounds met our full panel of
in vitro criteria for progression as a CNS drug, with permeability
through MDCK–MDR1 cell monolayers most often as a primary
liability.
This series of piperazines exhibits excellent dOR binding and
subtype selectivity, as well as a range of microsomal clearance
and hERG inhibition. The piperazine moiety itself remains a chal-
lenge for achieving permeability through the MDCK–MDR1 cell
monolayer while retaining favorable potency and ADME character-
istics. Modifications to the core scaffold that fundamentally alter
the physical properties (e.g., pKa and polar surface area) offer addi-
tional scope for CNS indications and are described in the following
companion paper.7
5. Based on model described in (a) Cavalli, A.; Poluzzi, E.; De Ponti, F.; Recanatini,
M. J. Med. Chem. 2002, 45, 3844; (b) MOE (Molecular Operating Environment) by
CCG (Chemical Computing Group), Montreal, Canada.; (c) Labute, P. Pac. Symp.
Biocomput. 1999, 4, 444.
6. Polli, J. W.; Wring, S. A.; Humphreys, J. E.; Huang, L.; Morgan, J. B.; Webster, L. O.;
Serabit-Singh, C. S. J. Pharmacol. Exp. Ther. 299, 620.
Acknowledgment
7. Dantzman, C. L.; King, M. M.; Ernst, G. E.; Wang, X.; McCauley Jr., J. P.; Andisik, D.
W.; Brush, K.; Bui, K. H.; Frietze, W.; Hoesch, V.; Liu, J.; Palmer, W. E.; Spear, N.;
Hudzik, T. J.; Wesolowski, S. S. Bioorg. Med. Chem. Lett., 2011 following paper.
We wish to thank the scores of scientists at AstraZeneca-Wil-
mington for providing such a stimulating environment for pharma-
ceutical research over the years. You will be missed.