Potent and selective mechanism-based inhibition of gelafinases [18]

Full text of DOI:10.1021/ja001461n

J. Am. Chem. Soc. 2000, 122, 6799-6800
6799
Scheme 1
Potent and Selective Mechanism-Based Inhibition of
Gelatinases
Stephen Brown,§ M. Margarida Bernardo,‡ Zhi-Hong Li,§
Lakshmi P. Kotra,§ Yasuhiro Tanaka,§ Rafael Fridman,‡ and
Shahriar Mobashery*,§
Departments of Chemistry and Pathology
Wayne State UniVersity, Detroit, Michigan 48202-3489
ReceiVed April 27, 2000
Specific interactions of cells within the extracellular matrix are
critical for the normal function of the organism. Alterations of
the extracellular matrix are carried out by a family of zinc-
dependent endopeptidases called matrix metalloproteinases (MMPs)
in various cellular processes such as organ development, ovula-
tion, fetus implantation in the uterus, embriogenesis, wound
healing, and angiogenesis.1,2 Gelatinases, collagenases, strome-
lysins, membrane-type MMPs, and matrilysin comprise the five
major groups of MMPs, of which at least 26 members have been
identified in humans to date. The activities of MMPs in physi-
ological conditions are strictly regulated by a series of complicated
zymogen activation processes and inhibition by the protein tissue
inhibitors of metalloprotainases (“TIMPs”).2,3 Excessive MMP
activity has been implicated in cancer growth, tumor metastasis
and angiogenesis, arthritis, connective tissue diseases, inflamma-
tion, and cardiovascular and autoimmune diseases.1,2,4 Due to the
potential therapeutic value of MMP inhibitors for these conditions,
synthetic inhibitors of MMPs are highly sought.5,6 All the known
inhibitors for MMPs take advantage of chelation to the active
site zinc ion for inhibition of activity. The known MMP inhibitors
usually suffer from toxicity to hosts.5a,c,7 Besides the issue of
undesirable side effects, the design of MMP inhibitors has been
complicated by only low levels of specificity among members
of the MMP family, which hampers our ability to target specific
MMPs in each pathological condition. We describe herein the
first mechanism-based inhibitor for MMPs, a novel concept for
the selective inhibition of these enzymes. We show that our
inhibitor rivals the action of TIMPs in its efficacy in inhibition
of MMPs.
(also known as “kcat inhibitors” or “suicide substrates”). This type
of inhibitor has the potential to impart high selectivity in inhibition
of closely related enzymes, such as MMPs.9 Our strategy for
mechanism-based inhibition of MMPs by compound 1 is depicted
in Scheme 1. The strategy envisions that coordination of the
thiirane with the active-site zinc ion would activate it for
modification by a nucleophile in the enzyme active site. The
biphenyl moiety in compounds 1-6 would fit in the P1′ subsite
of gelatinases, which is a deep hydrophobic pocket.10,11 Energy-
minimized complexes of MMP-2 and MMP-912 with compound
1 indicated that the biphenyl group would fit in the active site
analogously to the same group in reversible inhibitors of MMP-2
and MMP-9.6 This binding mode would bring the sulfur of the
thiirane in 1 into the coordination sphere of the zinc ion. The
models indicated that the thiirane moiety in compounds 2 and 3,
with longer carbon backbones, would not be able to coordinate
with the zinc ion, but would fit in an extended conformation in
the active site.
Scheme 2 shows the synthetic route for compounds 1-6.
4-Phenoxythiophenol 10 was prepared from the commercially
available 4-phenoxyphenol 7 via a three-step procedure described
by Newman and Karnes for a related system.13 Subsequent
alkylation of 10 with allyl bromide, 4-bromo-1-butene, and
5-bromo-1-pentene, respectively, led to the sulfanyl compounds
11-13 in good yields. Epoxidation of 12 and 13 with mCPBA
proceeded in 2-3 days, but that for 11 took 7 days and required
an excess of mCPBA. Finally, conversion of the epoxides 4-6
to their corresponding thiirane derivatives 1-3, respectively, was
accomplished by the treatment of each epoxide with ammonium
thiocyanate. Although the thiiranes 2 and 3 were isolated in high
yields (93 and 85%, respectively), thiirane 1 could only be
recovered in a poor 14% yield.
Increased level of activity for human gelatinases, MMP-2 and
MMP-9, has been implicated in the process of tumor metastasis
and angiogenesis.8 As a result, we have been interested in the
selective inhibition of these two key MMPs. For this purpose,
we have resorted to the design of mechanism-based inhibitors
* Address correspondence to this author: Phone: 313-577-3924. FAX:
313-577-8822. E-mail:s om@chem.wayne.edu.
Compounds 1-6 were evaluated with MMPs.14 Whereas
inhibitors 2-6 showed either no inhibition or relatively poor
§ Department of Chemistry.
‡ Department of Pathology.
(1) Massova, I.; Kotra, L. P.; Fridman, R.; Mobashery, S. FASEB J. 1998,
12, 1075-1095.
(8) Dalberg, K.; Eriksson, E.; Enberg, U.; Kjellman, M.; Backdahl, M.
World J. Surg. 2000, 24, 334-340. Salo, T.; Liotta, L. A.; Tryggvason, K. J.
Biol. Chem. 1983, 258, 3058-3063. Pyke, C.; Ralfkiaer, E.; Huhtala, P.;
Hurskainen, T.; Dano, K.; Tryggvason, K. Cancer Res. 1992, 52, 1336-1341.
Dumas, V.; Kanitakis, J.; Charvat, S.; Euvrard, S.; Faure, M.; Claudy, A.
Anticancer Res. 1999, 19, 2929-2938.
(9) The high specificity for a targeted enzyme with these inhibitors arises
predominantly from a combination of the differentials in noncovalent
interactions and those for the influences of microscopic rate constants for the
incremental steps in the process of enzyme inhibition (Silverman, R. In
Mechanism-Based Enzyme InactiVation: Chemistry and Enzymology; CRC
Press: Boca Raton, FL, 1988).
(10) (a) Morgunova, E.; Tuuttila, A.; Bergmann, U.; Isupov, M.; Lindqvist,
Y.; Schneider, G.; Tryggvason, K. Science 1999, 284, 1667-1670. (b)
Massova, I.; Fridman, R.; Mobashery, S. J. Mol. Mod. 1997, 3, 17-34.
(11) Olson, M. W.; Bernardo, M. M.; Pietila, M.; Gervasi, D. C.; Toth,
M.; Kotra, L. P.; Massova, I.; Mobashery, S.; Fridman, R. J. Biol. Chem.
2000, 275, 2661-2668.
(12) This enzyme has not been crystallized to date. However, a computa-
tional model based on three-dimensional homology modeling for this enzyme
is at hand in our laboratory (see citations 10b and 11).
(13) Newman M. S.; Karnes H. A. J. Org. Chem. 1996, 31, 3980-3984.
(2) Forget, M.-A.; Desrosier, R. R.; Be´liveau, R. Can. J. Physiol.
Pharmacol. 1999, 77, 465-480.
(3) Brew, K.; Dinakarpandian, D.; Nagase, H. Biochim. Biophys. Acta 2000,
1477, 267-283. Westermarck, J.; Kahari, V. M. FASEB J. 1999, 13, 781-
792.
(4) Nelson, A. R.; Fingleton, B.; Rothenberg, M. L.; Matrisian, L. M. J.
Clin. Oncol. 2000, 18, 1135-1149.
(5) (a) Michaelides, M. R.; Curtin, M. L. Curr. Pharm. Des. 1999, 5, 787-
819. (b) Beckett, R. P.; Davidson, A. H.; Drummond, A. H.; Huxley, P.;
Whittaker, M. Drug Discuss. Today 1996, 1, 16-26. (c) Woessner, J. F., Jr.
Ann. N.Y. Acad. Sci. 1999, 878, 388-403. (d) Skotnicki, J. S.; Zask, A.;
Nelson, F. C.; Albright, J. D.; Levin, J. I. Ann. N.Y. Acad. Sci. 1999, 878,
61-72.
(6) Freskos, J. N.; Mischke B. V.; DeCrescenzo, G. A.; Heintz, R.; Getman,
D. P.; Howard, S. C.; Kishore, N. N.; McDonald, J. J.; Munie, G. E.; Rangwala,
S.; Swearingen, C. A.; Voliva, C.; Welsch, D. J. Bioorg. Med. Chem. Lett.
1999, 9, 943-948. Tamura, Y.; Watanabe, F.; Nakatani, T.; Yasui, K.; Fuji,
M.; Komurasaki, T.; Tsuzuki, H.; Maekawa, R.; Yoshioka, T.; Kawada, K.;
Sugita, K.; Ohtani, M. J. Med. Chem. 1998, 41, 640-649.
(7) Greenwald, R. A. Ann. N.Y. Acad. Sci. 1999, 878, 413-419.
10.1021/ja001461n CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/30/2000

6800 J. Am. Chem. Soc., Vol. 122, No. 28, 2000
Communications to the Editor
Scheme 2
MMP-2 to the point that less than 5% activity remained. This
inhibitor-enzyme complex was dialyzed over 3 days, which
resulted in recovery of approximately 50% of the activity. This
observation is consistent with modification of the active site Glu-
404, via the formation of an ester bond, which is a relatively
labile covalent linkage.17
We observe selectivity in inhibition of gelatinases by inhibitor
1. The Ki values are 13.9 ( 0.4 and 600 ( 200 nM for MMP-2
and MMP-9, respectively. In contrast, the corresponding Ki values
for the other MMPs tested, including MMP-3, which does show
the slow-binding mechanism-based inhibition profile, are in the
micromolar range. Interestingly, the values for kon are 611- and
78-fold larger for MMP-2 and MMP-9, respectively, than that
for MMP-3. Collectively, these kinetic parameters make inhibitor
1 a potent and selective inhibitor for both MMP-2 and MMP-9,
more so for MMP-2. We have determined previously that two
molecules of either TIMP-1 or TIMP-2 bind to activated MMP-2
and MMP-9.18 One binding event is high affinity and would
appear physiologically relevant, whereas the second binding event
takes place with relatively lower affinity (micromolar).18 Inhibition
of MMP-2 and MMP-9 by TIMP-2 and TIMP-1, respectively,
also follows slow-binding kinetics. The kinetic parameters for
these interactions at the high affinity site are listed in Table 1.
We find it noteworthy that the kinetic parameters for the slow-
binding component of inhibition of MMP-2 and MMP-9 by
inhibitor 1 (kon and koff) approach closely the same parameters
for those of the TIMPs.18
We have outlined in this paper a novel example for potent
inhibition of human gelatinases by the small-molecule inhibitor
1, which follows both slow-binding and mechanism-based inhibi-
tion in its kinetic profile. This compound appears to behave
similarly to TIMP-2 and TIMP-1 in the slow-binding component
of inhibition. Furthermore, the inhibitor also exhibits a covalent
mechanism-based behavior in inhibition of these enzymes. The
selectivity that inhibitor 1 displays (in both affinities and the
modes of inhibition) among the other structurally similar MMPs
is noteworthy and should serve as a paradigm in the design of
inhibitors for other closely related enzymes in the future.
Table 1. Kinetic Parameters for Inhibition of MMPs by the
Synthetic Inhibitor
10-4kon (M-1 s-1
)
103koff (s-1
)
Ki (µM)
Inhibitor 1
MMP-2
MMP-9
MMP-3
MMP-7
MMP-1
TIMP-118
MMP-2
MMP-9
TIMP-218
MMP-2
MMP-9
11 ( 1
1.8 ( 0.1
7.1 ( 0.5
5.5 ( 0.4
0.0139 ( 0.0004
0.6 ( 0.2
15 ( 6
1.4 ( 0.3
0.018 ( 0.004
96 ( 41
206 ( 60
4.4 ( 0.1
5.2 ( 0.1
1.3 ( 0.2
1.2 ( 0.2
0.029 ( 0.005
0.024 ( 0.004
3.3 ( 0.1
2.2 ( 0.1
0.8 ( 0.1
1.3 ( 0.2
0.023 ( 0.004
0.058 ( 0.007
inhibition of the MMPs (Ki values of micromolar at best; see
Supporting Information), the behavior of inhibitor 1 was different.
Inhibitor 1 showed a dual behavior. It served as a mechanism-
based inhibitor with a partition ratio of 79 ( 10 (i.e., kcat/kinact
)
for MMP-2 and of 416 ( 63 for MMP-9.15 Furthermore, it also
behaved as a slow-binding inhibitor, for which the rate constants
for the on-set of inhibition (kon) and recovery of activity from
inhibition (koff) were evaluated (Table 1). It would appear that
coordination of the thiirane with the zinc ion (as seen in the
energy-minimized computational models; Scheme 1) would set
in motion a conformational change, which is presumed from the
slow-binding kinetic behavior. The kinetic data fit the model for
slow-binding inhibition.16 Covalent modification of the enzymes
ensued this conformational change. We incubated inhibitor 1 with
Acknowledgment. This work was supported by grants DAMD17-
97-1-7174 from the US Army (to S.M.) and CA-61986 and CA-82298
from the NIH (to R.F.). We are indebted to Dr. Paul Cannon and Dr. W.
Parks for the gifts of human MMP-3 and MMP-1, respectively.
Supporting Information Available: Detailed procedures for syn-
theses and kinetic determinations are provided (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA001461N
(14) Homogeneous preparations of MMPs were used in our studies.
Recombinant human MMP-2 and MMP-9 were prepared as described
previously (see Supporting Information). Representative members of the other
classes of MMPs, such as stromelysin 1 (MMP-3), matrilysin (MMP-7), and
collagenase-1 (MMP-1), were used in our studies.
(15) The partition ratio indicates that there is turnover of the thiirane for
each covalent inhibition of the enzyme. The partition ratios were relatively
low, such that given the quantities of the enzymes available to us, we were
not able to isolate and characterize the product of this turnover.
(16) Morrison, J. F. AdV. Enzymol. 1988, 61, 201-301.
(17) The time-dependent loss of activity is not merely due to the slow-
binding behavior. For instance, for a koff of 2 × 10-3 s-1 (the values are not
very different from one another in Table 1) the half-time for recovery of
activity (t1/2) is calculated at just under 6 min. The fact that 50% of activity
still did not recover after dialysis over 3 days strongly argues for the covalency
of enzyme modification.
(18) Olson, M. W.; Gervasi, D. C.; Mobashery, S.; Fridman, R. J. Biol.
Chem. 1997, 272, 29975-29983.