3934
A. A. Mazurov et al. / Bioorg. Med. Chem. Lett. 23 (2013) 3927–3934
molecule from their center of mass. A lower radius of gyration sug-
gests a more spheric-like shape of the compound. By contrast, a
higher radius of gyration suggests a more elongated the molecular
(Fig. 4). The object recognition model is based on a rodent’s spon-
taneous tendency to explore novel aspects and ignore familiar as-
pects of their environment. This pattern of exploratory activity can
shape. That the potency and efficacy of compounds at the
a4b2 and
be used as an index of memory function. Full a4b2 nAChR agonist
a
3b4 subtypes are dictated by the shape of the compound as sug-
34 enhanced working memory in the NOR paradigm in rats at
every dose (0.1–3 mg/kg, po) tested in a first experiment. Further
exploration of lower doses revealed that compound 34 statistically
and significantly increased novel object exploration time at doses
as low as 0.003 mg/kg. The maximum percent of recognition index
was demonstrated with doses of 0.3 and 3 mg/kg. By comparison,
animals treated with vehicle showed no evidence of enhanced
memory as subjects spent approximately the same amount of time
investigating the novel and familiar objects.
gested by their correlation with shape-based descriptors, Shadow
XZ, CHI-V-3_C and radius of gyration, is consistent with ROCS-de-
rived shape analysis. In fact, we have found that Tanimoto shape
can discriminate potent compounds (EC50 65
tent compounds (EC50 >5 M) with a Roc (receiver operating char-
acteristic curve) score of 0.82 (see Supplementary data Fig. S2 and
S3). The 5 M cutoff was selected based on compound 4, also
known as AZD1446. This diazabicyclo[3.3.0]octane derivative is a
highly selective 4b2 nAChR agonist which is currently being eval-
lM) from non-po-
l
l
a
In conclusion, several series of a4b2 nAChR agonists are dis-
uated in clinical trials for the treatment of cognitive disorders asso-
ciated with psychiatric or neurological disorder. It has an EC50
closed, compared, and contrasted. An exocyclic carbonyl moiety,
as a hydrogen bond acceptor, is positioned at an optimal distance
from a secondary amino group via either a diazabicyclic or an aza-
bicyclic scaffold. Attachments of short chain aliphatic or small ali-
cyclic groups to the carbonyl fragment suffice to promote high
binding affinity and activation of nAChRs. Modeling studies sug-
gested that ligand shape plays a key role in contributing to agon-
ism. Tested compound 34 has demonstrated improved working
memory performance in a novel object recognition task.
value of 4.9
lM at the human a4b2 nAChR, which led us to choose
5
lM as cuttoff between actives and inactive compounds. Using a
cutoff of 5 nM which could discriminate more potent compounds
was simply hindered by the fact that none of the compounds re-
ported herein exhibited such highly potent EC50 values. Moreover,
either using a cutoff of 5 nM (if there were active compounds at
such cutoff) or 1.3 lM (which would be based based on compound
34 which has also demonstrated a working memory performance
in a novel object recognition test) woulde erronously classify com-
pound 4 has inactive, because its EC50 value is greater than 5 nM
Supplementary data
and 1.3 lM.
Supplementary data associated with this article can be found, in
To ensure that our shape-based findings were not restricted by
the size of the chemical library, we performed a similar study in
which we increased the library size by including aromatic bicyclo
[3.3.0]octane series and and bicyclo [3.3.1]nonane series contaning
furoyl or pyridine moiety, which have been recently reported by
Mazurov et al.,6 Using the same program ROCS, the same EC50 cut-
References and notes
off (5 lM) and the same reference (see Fig. S3), we obtained a Tver-
679. Functional data were obtained according to methods described in Ref. 6.
sky shape model with a good roc score of 0.84. This result, shown
in Figure S4, was significantly different from a Lingos-derived mod-
el (roc score of 0.62) with a p-value of 0.001. Besides, due to the
fact that functional values could quite vary depending on the
source and experimental settings, we increased the EC50 cutoff to
10 lM, using this larger and more diverse chemical library. The
ROCS software package still produced a reasonable Tversky shape
model with a roc score of 0.83, whcih was also significantly supe-
rior and different from Lingos-derived model (roc score of 0.63),
with a p-value of 0.002.
In vitro pharmacological profiling and molecular modeling
studies of synthesized compounds indicate that the carbonyl
group, as a hydrogen bond acceptor, along with the cationic phar-
mocophoric element, promote activation of nAChRs, especially, the
9. 1H NMR (CDCl3, 300 MHz): d 3.69 (s, 3H), 3.60–3.42 (m, 2H), 3.36–3.05 (m, 3H),
3.18 (s, 3H), 2.82–2.57 (m, 2H), 2.08–2.02 (m, 2H), 1.82–1.65 (m, 2H), 1.46 (d,
9H); MS (m/z: 299 (M+1), 243 (M+1-56). 39, 1H NMR (CD3OD, 300 MHz): d
3.59–3.40 (m, 2H), 3.37–3.29 (m, 1H), 3.02–2.90 (m, 4H), 2.16–2.03 (m, 3H),
1.87–1.80 (m, 2H), 0.96–0.91 (m, 4H); MS (m/z): 180 (M+1). 41, 1H NMR
(CD3OD, 300 MHz): d 3.74 (s, 3H), 3.57–3.32 (m, 2H), 3.22–3.16 (m, 2H), 3.05–
2.88 (m, 3H), 2.38–2.03 (m, 2H), 1.83–1.61 (m, 2H); MS (m/z): 170 (M+1). 47,
1H NMR (CD3OD, 300 MHz): d 3.25–3.10 (m, 4H), 2.85–2.75 (m, 1H), 2.40–2.18
(m, 4H), 1.85–1.42 (m, 8H), 0.95 (t, 3H); MS (m/z): 211 (M+1).
a4b2 subtype. Moreover, it appears that pyridine, which is a com-
mon moiety in nicotinic ligands, or another heterocycle in the mol-
ecule can be equivalently replaced by both an exocyclic carbonyl
group and a hydrophobic aliphatic group to successfully confer
agonism towards the
4b2 nAChR is associated with a cognitive-enhancing effect,11 we
tested aliphatic amide 34 in a novel object recognition (NOR) task
a4b2 nAChR subtype. Since activation of
a
Mazurov, Anatoly A.
