2,4-Diaminopyridine δ-opioid receptor agonists and their associated hERG pharmacology

Article abstract of DOI:10.1016/j.bmcl.2009.01.106

A number of libraries were produced to explore the potential of 2,4-diaminopyridine lead 1. The resulting diaminopyridines proved to be potent and selective δ-opioid receptor agonists. Several rounds of lead optimisation using library chemistry identified compound 17 which went on to show efficacy in an electromyography model of neuropathic pain. The structure-activity relationship of the series against the hERG ion channel proved to be a key selectivity hurdle for the series.

Full text of DOI:10.1016/j.bmcl.2009.01.106

Bioorganic & Medicinal Chemistry Letters 19 (2009) 1702–1706
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
2,4-Diaminopyridine d-opioid receptor agonists and their associated
hERG pharmacology
*
Dafydd R. Owen , Margarita Rodriguez-Lens, Martin D. Corless, Steven M. Gaulier, Valerie A. Horne,
Ross A. Kinloch, Graham N. Maw, David W. Pearce, Huw Rees, Tracy J. Ringer, Thomas Ryckmans,
Blanda L. C. Stammen
Pfizer Global Research and Development, Sandwich Laboratories, Ramsgate Road, Sandwich, CT13 9NJ, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A number of libraries were produced to explore the potential of 2,4-diaminopyridine lead 1. The resulting
diaminopyridines proved to be potent and selective d-opioid receptor agonists. Several rounds of lead
optimisation using library chemistry identified compound 17 which went on to show efficacy in an elec-
tromyography model of neuropathic pain. The structure–activity relationship of the series against the
hERG ion channel proved to be a key selectivity hurdle for the series.
Received 19 December 2008
Revised 27 January 2009
Accepted 28 January 2009
Available online 5 February 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
2,4-Diaminopyridine
Delta opioid receptor
Agonist
Library chemistry
hERG
For a number of years considerable effort has been made to dis-
cover and develop selective d-opioid receptor (DOR) agonists for
the treatment of chronic pain. This is largely because DOR agonists
new chemical series based on 2,4-diaminopyridines is described
herein.
Compound 1 (Fig. 1) was identified as a singleton hit using a hu-
man recombinant cell line d-opioid receptor functional assay12
(DOR EC50) during targeted file screening. All compounds within
this Letter proved to be full d-agonists using this assay format.
have the potential to exhibit lower abuse liability than
l-opioid
receptor agonists as well as reduced gastrointestinal, respiratory
and cognitive effects.1,2 DOR agonists demonstrate effective antin-
ociceptive activity in preclinical models of inflammatory,3,4 neuro-
pathic3 and cancer pain5 but have little effect on acute nociception
models.3 Furthermore translocation of delta opioid receptors from
intracellular compartments to neuronal plasma membranes may
account for their effectiveness in chronic pain states.6 Convulsive
effects related to DOR agonism have been studied in rats with cer-
tain chemical classes of non-peptidic agonists. The conclusions
drawn suggest that the potential remains to further develop DOR
agonists.7
In addition to Pain, DOR agonists have recently received interest
in other indications such as depression,8 cardioprotection9 and
overactive bladder.10 A highly potent, peripherally selective, zwit-
terionic series from these laboratories, initially designed to target
irritable bowel syndrome, has also been disclosed.11 In our contin-
ued search for both peripherally selective and central nervous sys-
tem penetrant d-opioid agonists, we sought further chemical series
from targeted file screening. The identification and exploration of a
These compounds were also cross-screened in equivalent
-opioid functional assays. Unless otherwise stated, all compounds
were at least 50-fold selective over these opioid sub-types.
l and
j
Despite the low number of hits from those compounds
screened, we were attracted by the chemical possibilities of a hit
to lead programme based on 1 with its encouraging potency
(DOR EC50 263 nM, Emax 100%12), in vitro metabolic stability in hu-
man liver microsomes13 (HLM, Clint < 7
ll/min/mg) and particu-
larly its selectivity over the
EC50 > 10 M).
l
-opioid receptor -opioid
(l
l
N
H2
N
H2
N
H2
N
N
N
N
N
N
O
O
CN
OEt
1
2
3
* Corresponding author. Tel.: +44 0304648445.
E-mail address: dafydd.owen@pfizer.com (D.R. Owen).
Figure 1. Screening hit and prototypes.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.01.106

D. R. Owen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1702–1706
1703
O
Singleton follow up compounds were designed to increase con-
fidence in structure activity relationships (SAR) and the synthetic
viability of the series. These prototypes proved to be active (2
DOR EC50 64 nM and 3 DOR EC50 61 nM, Fig. 1).
Singleton synthesis was characterised by an amine displace-
ment (generally a piperidine) at the 2-position of 2-chloro-4-nitro-
pyridine-N-oxide. Reduction to the 2,4-diaminopyridine using iron
in acetic acid yielded the desired final compounds (Scheme 1). The
potential for library synthesis was also apparent. Two templates
were prepared presenting either a carboxylic acid or amine, ready
for the parallel synthesis of amides. (Scheme 1).
The selection of monomers for each library was achieved
through computationally designed properties for the final target
compounds. Limits were set for molecular weight, polar surface
area, lipophilicity14 and on the presence of reactive or vulnerable
functional groups. This resulted in libraries of ꢀ300 target com-
pounds being designed each time.
Two amide libraries were initially prepared using the differing
cores to explore the scope of the chemical series based on the
activity of the ester 2 and nitrile 3. Using the acid template 4
(X = Ph, R = CO2H) as the core (Library 1), amine monomers that
contained aryl groups produced the majority of the active com-
pounds (Fig. 2). For each library described, four selected com-
pounds are chosen to illustrate the nature of the SAR for DOR
agonist pharmacology, in vitro human liver microsomal stability
(HLM) and hERG ion channel activity.15
H
2N
H
N
N
N
OEt
N
R
*
N
N
*
5
7
6
OMe
Ph
H
N
N
O
*
*
8
R/
DOR EC50 HLM Clint
hERG Ki
(nM)
17
Compound
(nM)
5
24*
13
ul/min/mg
22
62
46
40
5
6
7
8
169
35
83
201
* 17-fold selective overµ-opioid receptor
Figure 2. Library 1 SAR.
O
H
OMe
N
O
*
2N
Cl
N
R
9
*
*
10
OMe
OMe
N
O
O
O
Ph
*
HN
12
11
R/
DOR EC50 HLM Clint
(nM)
14
hERG Ki
(nM)
Compound
ul/min/mg
43
69
14
13
82
9
59
60
343
10
11
12
Library 2 was based on the amino version of template 4 (X = Ph,
R = NH2). With the amide linkage now reversed, aryl acid derived
substituents featured in most, but not all of the active compounds
(Fig. 3).
548
171
>10000
Figure 3. Library 2 SAR.
Given the compact nature of the lead 1, greater increases in DOR
potency were expected for the addition of such lipophilic aryl
groups. In addition, any gains in potency were offset by the
in vitro metabolic instability (as measured by HLM) and the con-
siderable hERG liability that were now characteristic of these lar-
ger compounds.16
N
H2
H2N
N
N
N
N
N
Alongside the 4,4-disubstituted piperidine approach of Scheme
1, other templates were considered that could offer alternate syn-
thetic possibilities and SAR. Piperidine based templates, known
from other opioid receptor sub-types, were attached to the 2,4-
diaminopyridine group to explore the effect on delta activity and
opioid sub-type selectivity. Early efforts with this activity/selectiv-
O
O
N
O
N
Fentanyl
13
14
Figure 4.
l to d opioid selectivity switch mediated by 2,4-diaminopyridine.
ity hypothesis proved promising. In an example taken from
l ago-
nists, the phenethyl group of a fentanyl17-related compound was
replaced with the 2,4-diaminopyridine to give 13 (Fig. 4).
method of Scheme 1. In both cases moderately active d-opioid ago-
nists were identified (13 DOR EC50 207 nM, and 14 DOR EC50
308 nM, Fig. 4). These two compounds were greater than 12-fold
selective over any other opioid sub-type pharmacology. Off-setting
the basic centre from the piperidine using the 2,4-diaminopyridine
The spirocyclic template known from opioid receptor like-1
(ORL-1) pharmacology18 was also combined with the 2,4-diamino-
pyridine to give 14. Each compound was prepared by the singleton
N
H2
N
N
O2N
iii.
iv.
H
Library 1
NR1R2
2N
Ph
N+
H
N
N
O
O
i.
ii.
N
R
N
R
H2N
X
4
N
R
X
X
Singleton
or template
N
Ph
HN
R3
Library 2
O
Scheme 1. Reagents: (i) 2-chloro-4-nitropyridine-N-oxide, NaHCO3, t-amyl alcohol, 50 °C; (ii) Fe, AcOH; (iii) X = Ph, R = CO2Et; (a) 6 N HCl, reflux; (b) HNR1R2, HBTU, Hunig’s
base, NMP; (iv) X = Ph, R = NHBOC; (a) TFA, DCM; (b) R3CO2H, HBTU, Hunig’s Base, NMP.

1704
D. R. Owen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1702–1706
H
2N
appeared to be a useful strategy in conferring d-opioid selectivity
MeO
MeO
on piperidine templates originally identified for other opioid phar-
macologies (l, ORL-1).
N
N
N
*
*
21
The observations around 13 and 14 prompted further library
chemistry efforts. The spirocyclic 14 triggered a library varying
the cyclic amine at the 2-position of the 2,4-diaminopyridine. Of
the piperidines selected from the company file, those containing
a spirocyclic ring at the 4-position were prioritised in light of com-
pound 14.19 These compounds were prepared by a parallel equiv-
19
R
N
*
*
O
22
23
X
R/
DOR EC50 HLM Clint
hERG Ki
(nM)
Compound
19 X=4-Me
21 X=H
(nM)
9
ul/min/mg
alent of the singleton method in Scheme
1 where amine
12
80
17
<7
5030
displacement of 2-chloro-4-nitropyridine-N-oxide was followed
by reduction of the whole library. The reduction conditions were
changed to a transfer hydrogenation method for convenience in
the library protocol.
30
>10000
9690
22 X=H
52
23 X=H
108
>10000
Figure 6. 2-Aminopyridine amide library SAR.
From the data generated, piperidines were confirmed as the
best ring size for the core, as opposed to azetidines and pyrrol-
idines. Spirocyclic systems at the piperidine 4-position performed
favourably. The spirocyclic indane 17 proved to be very active
and had some encouraging in vitro properties (Fig. 5).
The discovery of compound 13 (Fig. 4) offered greater synthetic
scope for SAR exploration. In moving away from 4,4-disubstituted
piperidines and spirocyclic systems, compound design was now
based on a simpler 4-aminopiperidine core. Synthetically, two sub-
stituents could be varied easily, enabling molecular properties
(cLogP, molecular weight) to be better kept in balance during com-
pound design.
O2N
N+
O
N
O
+
i.
20
R
X
X=OH
X=Cl
HN
N
N
O
H2
O2N
N+
ii.
N
H
O
iii.
N
N
In order to replace the aniline in 13, a 2-pyridyl group was
investigated as a phenyl isostere off the 4-amino position of the
piperidine core. This isostere proved to be active (19, Fig. 6). This
did however trigger a change in the synthetic strategy for library
production. The 2-aminopyridine intermediate 20 (Scheme 2)
was not sufficiently reactive under standard amide coupling condi-
tions with our desired set of carboxylic acid monomers. In re-
sponse to this, acid chlorides were synthesised in situ using
Ghosez’s conditions.20 This method proved to be particularly effec-
tive when operating in a 96-well synthesis block format on a 10 mg
scale. The library protocol now featured three synthetic steps—acid
chloride generation, amide bond formation and nitro/N-oxide
reduction. Despite having three chemically orthogonal, linear steps
in a parallel synthesis format, success rates were regularly above
80% for this optimised protocol (Scheme 2).
N
N
N
O
N
R
R
Scheme 2. Reagents: (i) 1-chloro-N,N-2-trimethylpropenylamine, DCM; (ii) Et3N,
DCM; (iii) 10%Pd/Al2O3, NH4HCO2, MeOH.
However, this was at a 10-fold cost in DOR activity when compared
to 19 (Fig. 6).
Having investigated SAR in several templates using mainly li-
brary chemistry, the 4-aminopiperidine amides were selected for
further optimisation through cyclisation strategies. The versatility
of the starting 4-aminopiperidine template enabled variation of
ring systems and substituents at the 4-position of the piperidine
core.
Azabenzimidazolone template 24 was prepared (Scheme 3)
such that both arylation21 and alkylation chemistries could be
funded from the same intermediate. Once these divergent func-
tionalisations had been performed (Scheme 4), the installation of
the 2,4-diaminopyridine was carried out by the now standard
method.
Compound 19 was a rare example of a sub-10 nM DOR agonist
with ꢀ500-fold selectivity over hERG binding. It also showed some
of the best in vitro metabolic stability in HLM thus far. Smaller
acid-derived substituents such as cyclopropyl compound 23
showed low hERG affinity and excellent in vitro metabolic stability.
*
*
N
N
This potent sub-series identified some of the most active com-
pounds yet, optimizing potency through a cyclisation strategy.
However, minimizing rotatable bonds did little for metabolic sta-
bility in this case with the compounds showing tangible turnover
in the HLM assay. Most significant was their pronounced activity
in the hERG binding assay (Fig. 7). This clearly ruled out further
progression of the azabenzimidazolone sub-series.
O
O
N
N
H2
O
N
*
15
16
18
*
N
R
N
MeO
17
N
O
BOC
N
BOC
N
BOC
N
BOC
N
i.
ii.
iii.
R/
DOR EC50 HLM Cl int
hERG Ki
(nM)
>10000
18
4710
1840
Compound
24
(nM)
12
ul/min/mg
13
10
<7
-
15
16
17
18
NH2
N
N
H
N
HN
N
O
N
59
27
N
O2N
H
2N
H
846*
* 2-fold selective overµ-opioid receptor
Scheme 3. Reagents: (i) 2-chloro-3-nitropyridine, Et3N, n-BuOH, 80 °C; (ii)
Figure 5. Cyclic amine library SAR.
Pd(OH)2, NH4HCO2, EtOH, reflux; (iii) Triphosgene, Et3N, THF.

D. R. Owen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1702–1706
1705
NH2
*
*
*
N
N
H2N
N
N
BOC
N
N
N
N
N
R
O
O
N
iii.
N
N
N
BOC
N
O
N
N
O
O
N
N
i.
19
26
N
N
17
OMe
Ar
Ar
N
In vivo
QTC
(nM)
291
-
O
N
R/
hERG Ki hERG patch
NH2
BOC
N
N
Compound
(nM)
(nM)
H
N
ii.
17
19
26
4710
5030
903
1420
2020
54
N
-
N
O
iii.
N
N
N
Figure 9. Selected cardiovascular safety parameters.
O
R1
N
N
R1
series (>500 compounds), against their calculated lipophilicity,
showed a stringent cut off for the compound to be devoid of hERG
channel activity. As a general trend, a cLogP of >1 for the com-
pound meant an 80% chance of obtaining a binding Ki below
Scheme 4. Reagents: (i) ArB(OR)2, Cu(OAc)2, Et3N, DCE, 60 °C; (ii) R1CH2Br, K2CO3,
DMF, 50 °C; (iii) (a) TFA, DCM; (b) 2-chloro-4-nitropyridine-N-oxide, NaHCO3, t-
amyl alcohol, 50 °C; (c) 10%Pd/Al2O3, NH4HCO2, MeOH.
10
l
M in the hERG assay.
This selectivity issue was studied further by looking at selected
*
*
H2N
compounds in the more physiologically relevant hERG patch clamp
assay. Compounds showed the rare property of being up to ten fold
more potent in the patch clamp assay than the hERG binding assay,
thus reducing any selectivity window even further (Fig. 9). In vivo
studies in dog23 showed an even greater sensitivity to QTC prolon-
gation with small but significant effects observed at free plasma
concentrations as low as 291 nM for compound 17 (Fig. 9).
In summary, a singleton hit from targeted file screening was
rapidly explored using both new and established multi-step library
protocols. This Letter summarises >1500 analogues from the 2,4-
diaminopyridine series, mainly synthesised in libraries. The use
of the 2,4-diaminopyrdine group conferred excellent d-opioid
selectivity over other opioid sub-types on piperidine-derived cores.
However, the same heterocycle also seemed to bring a significant
hERG liability to the entire series, even in compounds of seemingly
moderate lipohilicity. Given that exceptionally potent compounds
were not identified (<1 nM), the series was deemed to have insuf-
ficient selectivity over this key cardiovascular safety parameter for
further progression.
N
N
N
25
*
26
28
OMe
*
O
N
27
N
R
R/
DOR EC50 HLM Clint hERG Ki
Compound
(nM)
3
16
28
41
ul/min/mg
(nM)
865
903
518
746
23
19
20
13
25
26
27
28
Figure 7. Azabenzimidazolone SAR.
With a selection of potent d-opioid agonists identified from the
series, pain efficacy was assessed in vivo. Compound 17 was se-
lected as it displayed the best overall in vitro properties with re-
spect to primary pharmacology, selectivity and metabolic
stability. Despite having in vitro HLM stability, 17 had a clearance
of >70 ml/min/kg in rat making it unsuitable for oral dosing. Using
an infusion protocol to achieve suitable plasma levels, 17 was
tested in an electromyography (EMG) model of pain in rat
(Fig. 8).22
Compound 17 showed maximal efficacy at a free plasma con-
centration of 23–76 nM in the EMG model. This reflected 1–3Â
the DOR EC50 from the primary functional screen. An analgesic,
but non-sedative 3 mg/kg iv dose of ketamine, was used as a stan-
dard for the in vivo model to guage the relative efficacy of a delta
opioid agonist. This dose of ketamine produced a 60% inhibition
relative to control animals—the same effect as the top dose of
DOR agonist 17.
References and notes
1. Quock, R. M.; Burkey, T. H.; Varga, E.; Hosohata, Y.; Hosohata, K.; Cowell, S. M.;
Slate, C. A.; Ehlert, F. J.; Roeske, W. R.; Yamamura, H. I. Pharamcol. Rev. 1999, 51,
503.
2. Kieffer, B. L.; Gaveriaux-Ruff, C. Prog. Neurobiol. 2002, 66, 285.
3. Petrillo, P.; Angelici, O.; Bingham, S.; Ficalora, G.; Garnier, M.; Zaratin, P. F.;
Petrone, G.; Pozzi, O.; Sbacchi, M.; Stean, T. O.; Upton, N.; Dindio, G. M.;
Sceideler, M. A. J. Pharmacol. Exp. Ther. 2003, 307, 1079.
4. Cahill, C. M.; Morinville, A.; Hoffert, C.; O’Donnell, D.; Beaudet, A. Pain 2003,
101, 199.
5. Brainin-Mantos, J.; Smith, N. D.; Rew, Y.; Goodman, M.; Taulane, J.; Yaksh, T. L.
Pain 2006, 122, 199.
6. Cahill, C. M.; Holdridge, S. V.; Morinville, A. Trends Pharmacol. Sci. 2007, 28, 23.
7. Jutkiewicz, E. M.; Baladi, M. G.; Folk, J. E.; Rice, K. C.; Woods, J. H. J. Pharmacol.
Exp. Ther. 2006, 317, 1337.
8. Saitoh, A.; Kimura, Y.; Suzuki, T.; Kawai, K.; Nagase, H.; Kamei, J. . J. Pharm. Sci.
2004, 95, 374.
The binding affinity for the hERG ion channel was closely
tracked throughout this 2,4-diaminopyridine series. Analysis of
all compounds screened in the dofetilide binding assay from this
9. Watson, M. J.; Holt, J. D. S.; O’Neill, S. J.; Wei, K.; Pendergast, W.; Gross, G. J.;
Gengo, P. J.; Chang, K.-J. J. Pharmacol. Exp. Ther. 2006, 316, 423.
10. Holt, J. D. S.; Watson, M. J.; Chang, J. P.; O’Neill, S. J.; Wei, K.; Pendergast, W.;
Gengo, P. J.; Chang, K.-J. J. Pharmacol. Exp. Ther. 2005, 315, 601.
11. Middleton, D. S.; Maw, G. N.; Challenger, C.; Jessiman, A.; Johnson, P. S.; Million,
W. A.; Nicholls, C. L.; Price, J. A.; Trevethick, M. Bioorg. Med. Chem. Lett. 2006, 16,
1434.
12. All pharmacology values in the tables are geometric means of at least three
experiments. Differences of <2-fold should not be considered significant. Delta
opioid agonist and antagonist activities were measured using a functional cell
Infusion
protocol
% inh. of.
control
Mean nM free
in plasma
76µg/kg +
0
1.6
23
76
72µg/kg/hr
240µg/kg +
290µg/kg/hr
48
61
660µg/kg +
920µg/kg/hr
based ALPHAscreen assay. ALPHAscreen is
a bead-based, non-radioactive
Amplified Luminescent Proximity Homogeneous Assay. Binding of molecules
captured on beads leads to an energy transfer from one bead to the other
Figure 8. EMG wind up efficacy of 17.
producing a luminescent signal. The ALPHAscreen method measured the

1706
D. R. Owen et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1702–1706
inhibition of cAMP produced via adenylate cyclase, following direct activation
13. Korfmacher, W. A.; Palmer, C. A.; Nardo, C.; Dunn-Meynell, K.; Grotz, D.; Cox,
K.; Lin, C.-C.; Elicone, C.; Liu, C.; Duchoslav, E. Rapid Commun. Mass. Spectrom.
1999, 13, 901.
14. Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc. 2007, 6, 881.
15. Singleton, D. H.; Boyd, H.; Steidle-Nichols, J. V.; Deacon, M.; de Groot, M. J.;
Price, D.; Nettleton, D. O.; Wallace, N. K.; Troutman, M. D.; Williams, C.; Boyd, J.
G. J. Med. Chem. 2007, 50, 2931.
16. Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. J. Med. Chem. 2006, 49, 5029.
17. Muijsers, R. B. R.; Wagstaff, A. J. Drugs 2001, 61, 2289.
18. Goto, Y.; Arai-Otsuki, S.; Tachibana, Y.; Ichikawa, D.; Ozaki, S.; Takahashi, H.;
Iwasawa, D.; Okamoto, O.; Okuda, S.; Ohta, H.; Sagara, T. J. Med. Chem. 2006, 49,
847.
of the enzyme by forskolin. The assay was based on competition between
endogenous cAMP and exogenously added biotin-cAMP. Donor beads were
streptavidin-coated and acceptor beads were coated with an anti-cAMP
antibody. The assay used CHO cells stably transfected with a CHO-Flp In
delta opioid receptor construct. The opioid receptor is Gi coupled to adenylate
cyclase, therefore agonists inhibited forskolin-stimulated cAMP production.
Delta opioid cells were cultured in T225 flasks in DMEM/nutrient mix F12
supplemented with 2 mM L-glutamine, 500 lg/ml hygromycin and 10% heat
inactivated foetal bovine serum. For assay the cells were harvested by
removing the media, washing twice with PBS and adding cell dissociation
solution. A viable cells count was performed and cells were resuspended at
1.6 Â 106/ml (8000 cells/well in the assay) in stimulation buffer (5 mM HEPES
pH 7.4, HBSS, 0.1% BSA, 10 mM MgCl2Á6H2O and 0.5 mM IBMX). Compound,
cells and acceptor bead mix were added to the plate before addition of
19. Allerton, C. M. N.; Owen, D. R.; Ryckmans, T.; Stammen, B. L. C. PCT Int. Appl.
2007, WO2007057775.
20. Devos, A.; Remion, J.; Frisque-Hesbain, A.-M.; Colens, A.; Ghosez, L. J. Chem.
Soc., Chem. Commun. 1979, 24, 1180.
forskolin at
a final assay concentration of 20 lM. Following a 30 min
incubation period the streptavidin donor beads were added and the plate
incubated overnight before reading on a Fusion reader. Percentage agonist
activity was calculated and EC50 values generated. All compounds screened
appeared as full agonists (Emax 100%). Known non-peptidic delta opioid agonist
BW-373U86 was a 3.2 nM full agonist in the assay.
21. Chan, D. M. T.; Monaco, K. L.; Li, R.; Bonne, D.; Clark, D. G.; Lam, P. Y. S.
Tetrahedron Lett. 2003, 44, 3863.
22. Mazario, J.; Herrero, J. F. Neurosci. Lett. 1999, 274, 179.
23. Chen, X.; Cass, J. D.; Bradley, J. A.; Dahm, C. M.; Sun, Z.; Kadyszweski, E.;
Engwall, M. J.; Zhou, J. Br. J. Pharmacol. 2005, 146, 792.