M. Duan et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6470–6475
6473
Table 1
Preparation of 2-amido-3-methyl-pyridiyl urea 34 began with a
nitro reduction of commercially available 2-bromo-3-methyl-5-
nitropyridine. Urea 32 was next obtained from the resulting amine
31 and 5, following a similar protocol described in Scheme 1. After
the installation of the cyano moiety using Zn(CN)2/Pd, the resultant
2-cyano-3 methyl pyridyl 33 was then converted into 34 by a
two-step sequence similar to that employed towards 30 (Scheme
5).
Antiviral Potency of 15a–d, 21, and 26 in HOS and PBL cell assays
Compounds
HOSa IC50 (nM)
PBLb IC50 (nM)
15a
15b
15c
15d
21
1.0
1.0
1.2
1.2
2.4
1.0
0.8
0.6
1.1
0.5
0.9
2.2
26
Synthesis of 2-amido-3-fluoro-pyridyl urea 39 proved to be a
challenge (Scheme 6). The nitration of the pyridine N-oxide 35 ob-
tained from an oxidation of 2-cyano-3 fluoropyridine under UHP
condition was highly troublesome and gave the desired nitro com-
pound 36 in only 9% yield under our best conditions. Reduction of
the newly introduced nitro and deoxygenation of pyridine-N-oxide
were achieved in one pot with Fe/HOAc at elevated temperature.
The resulting 5-amino-3-fluoro-2-pyridinecarbonitrile 37 was then
progressed to 39 using the same chemistry as described in Scheme
5. With all ring E modified urea moiety in hand, final compounds
40, 41, and 42 were synthesized by reductive amination with
aldehyde 14 (Fig. 4).
The additional methylation on ring E in compounds 40 and 41
led to a modest potency loss (ꢀ4-fold), while the fluorinated 42
was equipotent to parent 15a (Table 3). These compounds were
thus advanced to rat plasma stability screening. We were very
pleased to find that 40 and 41 demonstrated a substantial
improvement in the rat plasma stability despite that the observed
percentage increase for 41 over the time may suggest a slow disso-
lution of this compound. On the contrary, 42 exhibited a profile
similar to 15a (Table 2). We concluded that while the 6-methyl
moiety was likely to provide the direct steric hindence to enzymes;
the 3-methyl moiety was likely to exert its action by disrupting the
stability of planarity of complex a, Figure 4. On the other hand,
introduction of F in position 3 of ring E restored the five-mem-
bered-ring bident chelating environment due to the weak intramo-
lecular hydrogen bonding or electrostatic interaction between F
and the proton of the amide, hence, resulting in a quick hydrolysis
of the amide.14 Compounds 40, 41, and 42 were further evaluated
in rat PK models. Compound 40 had an excellent PK profile in rat
with low clearance and high bioavailability (49%). Compound 41
was also bioavailable though its Cl was slightly higher (Table 4).
Compound 42 exposure from either iv or p.o. dosing could not be
measured due to its instability in rat plasma.
a
IC50 values for the HOS assay represent the concentration of inhibitor that
prevents 50% of virus entry and subsequent activation of an HIV-LTR-Luciferase
reporter.12
b
IC50 values for the PBL assay represent the concentration of inhibitor that
prevents 50% of virus replication.12
Table 2
Stability of 15a, 40, 41, and 42 in Sprague Dawley rat plasma
Time (min)
Percent remaining (%)
15a
40
41
42
0
5
100
85
67
44
25
11
3
100
94
91
95
84
74
71
100
90
88
100
86
71
42
26
12
4
10
20
30
45
60
94
120
123
134
or no exposure for all these compounds in either iv or p.o. doses.
Detailed in vitro study indicated that the amide moiety on ring E
was rapidly converted into the corresponding carboxylic acid in
rat plasma (Table 2, 15a), a liability for the entire class of com-
pounds. One possible explanation of this unexpected amide liabil-
ity is that both the carbonyl oxygen of the amide group and the
pyridine nitrogen12 may chelate metal ions present in biological
milieu and form a quasi-planar five-membered ring (Fig. 3a). Such
chelation would lead to activation of the amide moiety and
accelerated amide hydrolysis in presence of plasma enzymes.13
To further test this hypothesis and to enable the future rat toxicity
study, we designed analogues that either introduce steric hin-
drance adjacent to the pyridine N (Fig. 3b), attempted to disrupt
the presumed co-planarity between the carbonyl oxygen and pyr-
idine nitrogen through steric interference of neighboring group
(Fig. 3c), or installed a putative intra-molecular H-bond (or electro-
static interaction) to restore the chelating environment (Fig. 3d).
Syntheses of these proposed molecules started with the prepara-
tions of the individual urea portions as shown in Scheme 4. Follow-
ing bis-Boc protection of 2-methyl-3-aminopyridine, an oxidation
with UHP was carried out, yielding pyridine N-oxide 27. Treatment
of 27 with TMSCN in presence of triethylamine installed the cyano
group at position 6. Upon removal of bis-Boc, the resulting 2-
methyl-3-amino-6-cyanopyridine dihydrochloride 28 was sub-
jected to urea formation. Subsequent oxidative hydrolysis of the
resulting 29 followed by a routine Boc deprotection to furnish 30.
In summary, a novel class of pyridyl carboxamides as potent
CCR5 antagonists (15a–d) was discovered. This class of the com-
pounds demonstrated high antiviral activity in both HOS and PBL
assays, but its exposure in rat PK could not be evaluated due to a
rapid metabolic hydrolysis of its amide. To resolve this issue, we
pursued several chemical modifications. While it is unclear
whether these analogues are stable in plasma because of hindrance
to deaminases or due to destabilization of putative five-member
chelate, which would render the amide highly prone to hydrolysis,
resulting in potent inhibitors 40 and 41 and successfully resolving
the amide instability issue in rat plasma. This allowed us to further
progress compounds 40 and 41 into pre-clinical development.
M 2+
O
O
O
O
N
N
N
H
H2N
O
N
N
H2N
O
N
N
H
O
N
H2N
H3C
O
N
N
H
N
N
H
N
F
N
H
N
N
H
N
F
F
F
F
c
b
d
a
Figure 3. Metal ion chelating hypothesis (a) and the proposals to disrupt/restore such chelating environments (b, c, and d).