Identification of 7 Nicotinic Acetylcholine Receptor Agonists for their Assessment in Improving Cognition in Schizophrenia - Accounts in Drug Discovery Case Studies in Medicinal Chemistry - page 343

portion modifications (see Figure 15.10) showed a7 nAChR binding and
functional activity similar to that of PHA-543613. Metabolic stabilities in RLM
(expressed as a percent remaining) for analogs 18a-e ranged from 10% to 77%,
which met our initial metabolic stability criteria of Z10%. 26 Early in the
program, we had observed a rank-order correlation between RLM and in vivo
rat clearance, affording a key attribute to drive our ADME (absorption, dis-
tribution, metabolism, and excretion) SAR based on in vitro properties. We had
also observed, however, that RLM routinely underpredicted compound sta-
bility in HLM; 20,27 hence the low RLM threshold for compound advancement.
The rationale for using RLM instead of HLM for routine screening was two-
fold: (1) our primary in vivo ecacy and safety models were in rats (vide infra),
and (2) high-throughput RLM data were readily available to us, whereas HLM
data were not.
Focusing on the azabicyclic portion of PHA-543613, we prepared four new
ring systems as quinuclidine isosteres (see Figure 15.10). 27,43,44 For each new
system prepared (19-22), all possible stereoisomers were tested individually as
pure stereoisomers. Once the most biologically relevant stereoisomer with regard
to a7 nAChR agonist activity and selectivity was identified, we also identified its
absolute configuration (see Figure 15.10). For some azabicyclic ring systems, a7
nAChR agonist activity resided in a single stereoisomer {e.g. 1-azabicy-
clo[3.2.1]octan-3-yl (21) and 7-azabicyclo[2.2.1]heptan-3-yl (22)}, whereas for
others {e.g. 1-azabicyclo[2.2.1]heptan-3-yl (19)}, significant a7 nAChR activity
was seen in more than one isomer. In the case of azabicyclic ring 19, two ste-
reoisomers with the same aryl fragment [i.e.,(3R,4S)- and (3S,4S)-isomers]
typically showed significant a7 nAChR activity; however, the (3S,4S)-isomer
also usually displayed significant 5-HT 3 activity. Each new azabicyclic amine
19-22 was coupled to the aryl fragments a-f shown in Figure 15.10.
Analogs derived from azabicyclic ring 19 were on average two to three times
less potent than their corresponding quinuclidine analogs. The RLM stability
of analogs derived from azabicyclic ring 19 was significantly enhanced with
respect to the corresponding quinuclidine analogs, however, which may be due
to their lower log D values. 27 Analogs with azabicyclic ring 20 were 8-20 times
less potent than their corresponding quinuclidine analogs; however, analog 20a
still met our potency criteria, and it was stable in RLM. Aryl amides derived
from azabicyclic ring 21 were equipotent to the quinuclidine analogs, and the
RLM stability of type 21 analogs was also comparable to that of quinuclidine
analogs. Compounds derived from azabicyclic ring 22 were in general two to
three times less potent than the corresponding quinuclidine analogs. Surpris-
ingly, even though type 22 analogs have an epibatidine-like 7-azabicy-
clo[2.2.1]heptan-3-yl skeleton, they did not exhibit significant off-target activity
at other nicotinic receptors such as a4b2 and a3b4. 27 In fact, all of the analogs
derived from azabicyclic rings 19-22 showed excellent selectivity against off-
target nicotinic receptors such as a4b2, a3b4, and a1b1gd. 27 Based on their
excellent a7 nAChR agonist activity and acceptable metabolic stability in
RLM, a number of amides derived from azabicyclic rings 19-22 were selected
for further profiling 27 in a set of in vitro and in vivo assays that included 5-HT 3