M. D. Bailey et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4447–4452
4451
Br
O
O
O
O
N
N
O
21
22
23
N
H
0.8
0.8
0.4
8.4
6.5
2.5
>300
>300
91
nd
37/83
44/56
34/45
7.1
3.9
2.4
O
Cl
O
O
N
H
160
>300
O
O
Br
N
O
N
H
improved cellular potency of 8.4 nM. Finally, replacement of the
N-terminal tert-butyl group by a phenyl group as found in com-
pound 22 showed a 10-fold improvement in cellular potency while
maintaining the intrinsic potency as compared to compound 20. It
appears that this later substitution provides an additional hydro-
phobic interaction with the enzyme to improve potency several
fold. Finally, with this same N-substitution our most beneficial
substitutions on the quinoline unit resulted in our most potent
compound in this succinamide series, compound 23 (IC50 = 0.4 nM,
EC50 = 2.5 nM).
Having accomplished our goal of producing potent analogs,
some ADME and physical properties were also screened. It appears
that the nature of both the P2 and the N-terminal group both sep-
arately affect the HLM stability as well as the Caco-2 permeability
values (compound 20 vs 22, and 21 vs 23). In general the solubility
of this series was found to be good, however the compounds suf-
fered from moderate permeability (Table 4). Unfortunately, when
compound 23 was evaluated further in rats at a 5 mg/kg dose,
the plasma level was found to be low at 1 h (0.1 lM) and not
detectable at 2 h. Discouraged by these results, no further work
was conducted on this series of C2-alkoxy quinoline derivatives.
Interestingly however, Achillion have advanced a clinical candidate
into phase 2 trials (ACH-1625, Sovaprevir) with a very similar suc-
cinate backbone modification. In this case the terminal amide is
disubstituted with a cyclohexylamine and lacks the NH interaction
found in our inhibitors. Likely this capping group binds in a differ-
ent fashion to take advantage of the lyophilic nature of this region.
Sovaprevir was shown to have an EC50 value of 11 nM with a mod-
est HLM stability and an excellent safety profile. When dosed in
animals this compound gave low plasma levels but was found to
be rapidly and extensively partitioned into the liver to give a high
Figure 2. Comparison of bound succinamide and carbamate inhibitors. Superpo-
sition of the bound model of compound 19 (thick white sticks) and a cyclopentyl
carbamate capped analog (thick grey sticks) with focus on the P3 and capping group
region. Protein residues are represented in thin sticks and are colored according to
the model from which they originated. Key hydrogen bonds are indicated by blue
lines while key distances are indicated by double headed arrows.
succinamide NH and the aliphatic carbon of the succinamide are
displaced approximately 0.8 and 0.9 Å respectively. These shifts
in the succinamide backbone provide more space and an improved
geometry to allow for the hydrogen bonding interaction with
Ala157. Table 1 can be nicely explained by the clear correlation be-
tween potency and hydrogen bond donor potential at this position.
The capping group also plays a role in potency such that the t-butyl
group in compound 18 is fivefold less potent than the correspond-
ing phenyl group found in compound 19. It seems likely that this
terminal aryl group is in a better position to favourably interact
with the hydrophobic P4 region of the enzyme.
Drawing on past experience, we replaced the P2 isoquinoline
with a quinoline and incorporated a C2-ethoxy group as a replace-
ment for the thiazole ring, a common feature found in our previous
inhibitors. As well we introduced the beneficial halogen atom in
the C8 position to furnish compound 20 with much improved
potency. Introduction of the more optimal C8-Br with an added
C7-OMe group as found in our previous series provided compound
1
9
liver/plasma ratio.
In summary, replacement of the P3 amino acid back bone of
substrate-based inhibitors with a succinamide moiety was suc-
cessfully accomplished in two distinct series. In designing these
new series of inhibitors, we needed to optimize the succinamide
moiety, as well as introduce critically important substitutions on
the P2 hydroxyquinoline fragments. In particular, introduction of
a C8-Br into the hydroxyquinoline fragment delivered potent
nanomolar dipeptide inhibitor of the NS3 serine protease. One
member of the C-terminal acid inhibitors compound 9 was profiled
and was shown to have both good in vitro and in vivo properties,
comparable to the more peptidic analog. In the case of the acyl sul-
fonamide series, somewhat smaller P2 groups could be installed
which produced analogs having interesting properties. These ana-
logs supporting novel P2 fragments exhibited sub-nanomolar
2
1 with good intrinsic potency (IC50 = 0.8 nM) but also a much