E. M. Smith et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4602–4606
4603
R3CHO
H2N R2
The respective aryl urea azetidinone 21 showed a significant
improvement in the PK profile. However, this modification gave a
significant loss in CaV3.2 potency.
+
a
R2
R3
O
R2
R3
N
N
R2
To remove a potential site of metabolism and improve the PK,
the aliphatic groups at R2 in the azetidinones 9 azetidines 10 were
replaced with an aryl group as shown in Table 2. When R2 equals
3-fluorophenyl, these compounds retain activity provided that
the alkyl (isoleucine methyl ester) and aralkyl ureas are used. In
this series the branched chain alkyl ureas and aryl ureas are less
potent. As seen before, isomer A is still more potent than isomer
B. In this case, the azetidinone series 23a and 23b is slightly more
potent than the corresponding azetidine 22a and 22b (CaV3.2).
Compound 23a was shown to have >92-fold selectivity for blocking
hCaV3.2 compared to hCaV2.2 and hNaV1.5. The R2 3-fluorophenyl
series is slightly more potent than 2-fluorophenyl series.16
In the azetidinone series, the SAR of the isoleucine methyl ester
was explored showing that replacement of the methyl ester 23a
with cyano 24a or isopropyl ester 25a maintained activity. Methyl
amide 26a resulted in slight loss of activity as did the use of
t-butylserine 27a. Replacement of isoleucine with alanine 28a
and serine 29a resulted in a significant decrease in activity. It
was interesting to see that the isoleucine could be replaced with
CO2Et
R3
N
c
6
b
N
N
N
Boc
d, e
7
Boc
Boc
8
5
d, e
R2
O
R2
N
N
R3
R3
N
R1HN
N
O
R1HN
9
10
O
Scheme 1. Reagents: (a) toluene, molecular sieves; (b) LiN(i-Pr)2, THF;
(c) Rh(CO)(PPh3)3, Ph2SiH2, THF; (d) TFA, CH2Cl2; (e) R1N@C@O, ClCH2CH2Cl, CH3CN.
Table 1
O
N
N
N
a
a-methylbenzyl amine with only a slight decrease in the ability
H
N
H
N
N
N
to block the CaV3.2 channel. As shown in Table 2 and 4-fluoro-
a
H3CO2C
Cl
Cl
O
O
2
(R)-methylbenzyl urea 32a was the most potent aralkyl urea
against CaV3.2. Importantly, we found that this modification pro-
vided very good TRPV1 selectivity. In Table 2, for R2 aryl series,
the aryl ureas are less potent. However, the R2 aryl azetidinones
36a and 36b are more potent (CaV3.2) than the respective azeti-
dines 35a and 35b.
1
3
N
H
N
H
N
N
N
F
O
Gratifyingly, we found that replacement of the R3 aryl with a
2-pyridyl group, further improved the potency against the CaV3.2
channel (Table 2). As we had demonstrated before, for both azeti-
dines and azetidinones, the CaV3.2 potency resides in isomer A for
this series as well. As seen before, azetidinones 39a and 39b are
more potent than azetidines 38a and 38b. Incorporation of the
2-pyridyl group provided a dramatic improvement in the rat PK
Cl
O
4
F
F
F
Compds
IW (VC) hCaV3
2 IC50 nM
Rapid rat po
AUC nM h
TRPV1a,11
2b,12
312
412
2ab
2b
104 (123)
174 (340)
138 (328)
115 (124)
733 (3197)
0.70
87.10
64.00
19.10
0
0
profile with 0–6 h AUC’s from 1 to 18
dose of 10 mpk.
lM h observed after an oral
With this overall improvement, we then turned our attention to
the in vivo evaluation of these compounds in the rat SNL model of
neuropathic pain.17 Three compounds were selected: (1) best
hCaV3.2 potency: 39a (isomer A.VC hCaV3.2 IC50 42 nM, rat PPB
99.6%); (2) best AUC: 40a (isomer A; rat AUC (0–6 h, po) @
10 mpk 18,055 nM h, rat PB 99.6%); and (3) good hCaV3.2 and
AUC: 42a (isomer A). Compounds 39a and 42a were inactive in this
model at 100 and 30 mpk, respectively. We speculated that the
inactivity of these compounds in the SNL was either due to their
poor TRPV1 selectivity or their high protein binding. There was a
trend toward efficacy with compound 40a at 2 h (50% PWT:
9.2 2.6 Gm; vehicle À4.9 2.1 Gm) and 3 h (50% PWT: 8.9
2.8 Gm; vehicle À5.0 1.7 Gm) after dosing 30 mpk but statistical
significance was not achieved.
Two isomer B compounds that displayed better selectivity over
TRPV1: 41b (R1 is 3-trifluoromethylphenyl) with moderate CaV3.2
activity, good AUC and clog P, and better TRPV1 selectivity; and for
44b (R1 is 3,5-difluorophenyl) with moderate CaV3.2 activity, good
AUC and clog P and moderate TRPV1 activity were also tested in
the SNL. These compounds were found to be inactive in the rat
SNL at 30 mpk, but 44b, a positive trend was seen at 1 h 50%
PWT 7.4 2.6 Gm, vehicle 3.0 1.4 Gm) which suggested that the
high protein binding was the problem. In support of this we found
that the ability of these compounds to block the CaV3.2 channel in
the voltage clamp assay was considerably right shifted by 40- to
>469-fold 24, 40, 44, 46, and 49 when tested in whole cell patch
clamp the presence of 100% rat serum (Table 3). The example of
a
TRPV1 capsaicin activation, % inhibition at 3.3 l
g/mL.11
Compound 2 was separated on HPLC Chiralpak AS column to
give 2a, isomer A, faster eluting and 2b isomer B.
b
After confirming that the CaV3.2 activity resides predominately
in 2a (isomer A), we developed the SAR using the modular syn-
thetic routes shown in Scheme 1 to improve the potency and opti-
mize the DMPK profiles for these compounds.
Using the methodology in Scheme 1, we were able to explore
the SAR of the azetidinones 9 and azetidines 10. First, we investi-
gated azetidine series by varying the azetidine nitrogen substitu-
ent (R2) from methyl, isopropyl, and isobutyl; and also the C4
substituent (R3) being either 4-chlorophenyl and 4-fluorophenyl
(Table 2).
When R3 4-chlorophenyl azetidine series, the R2 isopropyl 2–4
is equipotent with isobutyl 15 and more potent than methyl ana-
logs 11 and 12. The respective R2 methyl azetidinones 13 and 14
are less potent than the azetidines. The R3 4-chlorophenyl azeti-
dines 2–4, 11, 12, and 15 are more potent than the respective R2
methyl and R2 isopropyl, R3 4-fluorophenyl azetidines 16–18.
To improve the PK profile, we explored introducing a cyclopro-
pyl group15 in place of isopropyl at R2 (Table 2). In the azetidine
series, this modification was made in both the alkyl 19 and aryl
urea 20 analogs which both showed a modest improvement in
the PK profile, however, a loss in CaV3.2 potency was observed.