J. I. Levin et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2189–2192
2191
a possible source of the musculoskeletal side effects seen
in clinical trials of broad spectrum MMP inhibitors.11
Although benzyl ether 4l was only a weakly active
TACE inhibitor, it was the first member of this series
that was more potent against TACE than against
MMP-1, MMP-9, and MMP-13. The optimization of
this lead into a potent and selective TACE inhibitor is
represented with compounds 4u–4y. The 5-bromo ana-
logue 4u provides a slight increase in TACE potency
over 4l and substantially improves selectivity for TACE.
Next, a 5-fold increase in potency was realized in going
from N-benzyl analogue 4u to the corresponding N-
methyl sulfonamide 4v. Replacement of the terminal
phenyl ring of 4v with a 3-thienyl group (4w) offered no
improvement.
The in vitro potencies against the MMPs and TACE for
a series of anthranilate hydroxamic acid analogues, 4a–
4i, in which substitution at the 5-position of the
anthranilate phenyl ring was explored are shown in
Table 1. All of these compounds, with the exception of
analogue 4f, are potent inhibitors of both MMP-9 and
MMP-13, comparable to the 5-unsubstituted derivative
1. Aryl or heteroaryl groups at the anthranilate 3- (4h
and 4i) or 5-positions (4c–4e) provide excellent activity
versus MMP-13. Furthermore, the 5-aryl compounds,
4c–4e, display high levels of selectivity over MMP-1,
from almost 50- to over 700-fold. Interestingly, incor-
porating basic alkylamino substituents at the 5-position,
as in diethylaminomethyl derivative 4f and dimethyl-
aniline 4g, results in compounds that are potent against
MMP-9 and selective over MMP-1 and MMP-13.
Thiazole analogue 4x, however, is over twice as potent
as 4v and is approximately 400-, 40-, and 30-fold selec-
tive over MMP-1, MMP-9, and MMP-13, respectively.
Still greater TACE selectivity is obtained with 3-picolyl
ether 4y, while the analogous 2- and 4-picolyl deriva-
tives (not shown) displayed a 6- to 8-fold diminution of
TACE activity.
TACE activity is also affected by the choice of the
anthranilate 5-substituent. Thus, the 5-bromo derivative
4a is 5 times more potent against TACE than the 5-
unsubstituted parent, 1. Although a 5-bromo substituent
(4a) provided the greatest TACE potency in the series 4a–
4i, all of the 5-substituents in this series which were not
excessively bulky (4d and 4e) or basic (4f and 4g) had
enhanced TACE activity relative to compound 1. Unfor-
tunately, despite their potency against isolated enzyme,
compounds 4a–4i did not display significant TNF-a
inhibitory activity in a THP-1 cellular assay at 3 mM.12
Examination of the X-ray structure of TACE suggests
that the terminal phenyl or heteroaryl ring of 4u–4y is
positioned in the channel connecting the S11 and S31
pockets.14 The TNF-a inhibitory activity of compounds
4u–4y in a THP-1 cellular assay was disappointingly
poor, however. Thiophene 4w was the most potent
derivative in cells, affording only 22% inhibition of
TNF-a at 3 mM.
The in vivo bioactivity against MMP-13 for some of the
anthranilate-hydroxamates (4c–4e, 4h–4i, and 4p–4t)
after oral dosing was assessed through the use of a dia-
lysis tubing implant assay.15 All of the compounds tes-
ted were compared to Novartis’ sulfonamide-
hydroxamate clinical lead, CGS-27023A,16 in the same
experiment. Despite the fact that the compounds tested
had in vitro potencies comparable to CGS-27023A
against MMP-13, only 4h was as potent as CGS-27023A
in vivo. Compounds 4c, 4e, 4p, and 4r were approxi-
mately 80% as potent CGS-27023A.
Variations of the P11 substituent of the anthranilate
hydroxamic acids are shown for compounds 4j–4y in
Table 1. NMR studies have shown that the R1 group of
compound 4 occupies the S11 pocket of MMP-13.13
Lengthening of P11 alkoxy moieties (4j–4l) results in a
loss of MMP-1 activity, as expected from the shallow
nature of the S11 pocket of this enzyme. However,
MMP-9 and MMP-13 activity for 4k and benzyl ether 4l
also diminish beyond useful levels. Inserting an addi-
tional methylene spacer in 4l leads to the phenethyl
analogue 4m and restores some MMP-9 and MMP-13
activity. On the other hand, the more rigid P11 biaryl
ethers, 4n–4q, and thioether 4r retain or improve their
potency against MMP-9 and MMP-13 relative to methyl
ether 1. Surprisingly, comparison of the neutral phenyl
ethers, 4n and 4o, shows that a bulky para-substituent is
required for these to realize greater than 100-fold selectiv-
ity over MMP-1. It is possible that the arginine residue
that normally forms the bottom of the MMP-1 S11 pocket
is pushed aside to some degree by the P11 biphenyl ether
substituent of 4n, extending the depth of the pocket.
Thioether 4r and the more polar 4-pyridyl ethers, 4p and
4q, do not need additional substitution and are among
the most selective members of the series (>250-fold). It
is important to note that compound 4q, with its lengthy
P11 moiety, no longer requires the 3-substituent on the
anthranilate ring (R3=H) that our initial series needed
to attain acceptable potency.4 The biaryl derivatives 4s
and 4t are also potent and selective MMP-13 inhibitors,
with 4t possessing over 300- and 35-fold selectivity over
MMP-1 and MMP-9, respectively.
Anthranilate-hydroxamates 4c, 4e, and 4p were also
tested side by side with CGS-27023A in a bovine
articular cartilage explant assay.17 At a concentration of
1 mM, compound 4e provided a level of inhibition of
cartilage degradation slightly superior to CGS-27023A
(4e: 83%/CGS-27023A: 70%, n=2). Compounds 4c
and 4p were roughly equivalent to CGS-27023A at 1 mM
in this assay.
Hydroxamates 4c, 4e, and 4p were then evaluated in an
in vivo rat sponge-wrapped cartilage model.18 Only
pyridyl ether 4p demonstrated significant inhibition of
collagen degradation in this model. Oral dosing at
50 mg/kg/bid provided a 35% inhibition (n=2) of col-
lagen degradation compared to a 51% inhibition by
CGS-27023A at the same dose.
In conclusion, we have expanded upon our initial series
of anthranilate-hydroxamic acid MMP inhibitors. We
have found that through the proper choice of
Levin
