Potent, selective pyrone-based inhibitors of stromelysin-1

Article abstract of DOI:10.1021/ja054558o

In an effort to develop alternatives to hydroxamate-based matrix metalloproteinase inhibitors (MPIs), we have utilized the drug discovery program LUDI enhanced with the structural coordinates of a bioinorganic model complex. This method has yielded the fi

Full text of DOI:10.1021/ja054558o

Published on Web 09/22/2005
Potent, Selective Pyrone-Based Inhibitors of Stromelysin-1
David T. Puerta, John Mongan, Ba L. Tran, J. Andrew McCammon, and Seth M. Cohen*
Department of Chemistry and Biochemistry, Howard Hughes Medical Institute, Department of Pharmacology,
Bioinformatics Program and Center for Theoretical Biological Physics, UniVersity of California, San Diego,
La Jolla, California 92093
Received July 9, 2005; E-mail: scohen@ucsd.edu
This communication describes the design and synthesis of
pyrone-based inhibitors of matrix metalloproteinases (MMPs) that
show superior potency over their hydroxamate analogues. The zinc-
(II)-dependent MMPs have been pursued as chemotherapeutic
targets for the treatment of illnesses, such as cancer, arthritis, and
heart disease. Consequently, over the past two decades, attempts
to interfere with MMP activity have yielded numerous inhibitors.1
MMP inhibitors (MPIs) are based on a two-part strategy: chelation
of the catalytic zinc(II) ion combined with noncovalent interactions
within “subsite” pockets in the MMP active site.1,2
The majority of MPIs synthesized to date contain a hydroxamic
acid as the chelating or zinc-binding group (ZBG).1,2 Despite recent
improvements, hydroxamate-based MPIs have not yet succeeded
in clinical trials.3 This has prompted the investigation of a limited
number of non-hydroxamate-based MPIs.4-6 Herein, we describe
inhibitors that utilize a pyrone ZBG, which results in improved
Figure 1. LUDI docking image of backbone fragment (green, in S1′ subsite)
with pyrone ZBG (colored by element) in the active site of MMP-3 (gray).
This fragment combination leads to the compound designated AM-5 (see
Scheme 1). The zinc(II) ion is shown as a gold sphere; a hydrogen bond
between the ZBG and L164 is shown as a dashed line.
potency and novel selectivity relative to similar hydroxamate-based
MPIs.4
Pyrones were selected for this study due to their synthetic
versatility,7 known biocompatibility,8 and good aqueous solubility.
An earlier study examining the use of maltol (3-hydroxy-2-methyl-
4-pyrone) as a ZBG indicated that the 2-methyl substituent was
favorably oriented toward the hydrophobic S1′ pocket of stromel-
ysin-1 (MMP-3).9,10 Several studies show that targeting the S1′
pocket of MMPs yields potent and selective MPIs.1,2 Therefore,
we sought to attach simple aryl groups to the 2-position of maltol
in order to exploit this interaction.
Scheme 1 a
To design pyrone-based inhibitors, we used the drug discovery
program LUDI (Accelrys) augmented with parameters from a
bioinorganic model complex. LUDI uses a constrained docking
approach that identifies optimal fragments to link to the pyrone
moiety at a specified point of attachment. Structural data of maltol
bound to a tris(pyrazolyl)borate model complex9 were integrated
into a known MMP crystal structure to generate the initial receptor
complex (Figure S1).11 The point of attachment to the ZBG was
defined as a N-H bond from an amide moiety on the 2-position
of the maltol ring (the amide group was built in silico on the ZBG).
Fragments were screened and ranked using a LUDI scoring
function.12 The results from an initial screen with MMP-3 using
the LUDI link library yielded modest scores for several compounds.
Consequently, we created a custom library primarily based on the
work of Hadjuk et al.,13 which generated three high, one moderate,
and two low scoring fragments. The result of the LUDI docking
for one of the high scoring compounds (AM-5, vide infra) is shown
in Figure 1. The fragment in Figure 1 was found to reside in the
S1′ pocket of MMP-3. The high and low scoring fragments from
the custom library were similar in structure; therefore, all six
compounds were synthesized to test the accuracy of the LUDI
docking and scoring function.
a Key: (I) (a) NHS, DCC, dry THF; (b) “amine”, dry THF, 88%; (c)
10% Pd/C, H2 35 psi, MeOH or 1:1 HCl:CH3COOH, 60-89%. (II) (d)
ArB(OH)2, 2 M K2CO3, Pd(C2H3O2)2, PPh3, toluene, 135 °C, 40-85%;
(e) 10% Pd/C, H2 35 psi, MeOH or 1:1 HCl:CH3COOH, 60-91%.
Synthesis of the pyrone-based MPIs was performed according
to Scheme 1. Two synthetic routes were utilized, based on the
commercial availability of the desired amine backbones. 2-Carboxy-
3-benzyloxy-6-methylpyran-4(1H)-one (1) was prepared by a
literature method.7 Compound 1 was then activated with NHS,
followed by coupling to the desired amine, and removal of the
benzyl protecting group to yield compounds AM-1, AM-2, AM-
3, and AM-4. The synthesis of AM-5 and AM-6 was accomplished
similarly, but required the Suzuki coupling of 3-benzyloxy-6-
methylpyran-4(1H)-one-2-carboxy-N-(4-iodobenzylamide) (2) with
9
14148
J. AM. CHEM. SOC. 2005, 127, 14148-14149
10.1021/ja054558o CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
generally more potent for MMP-3 than for MMP-2,4,17 which is
attributed to the difference in the optimal pH for the two enzymes.
MMP-3 prefers a more acidic environment (pH ∼6.0) compared
with other MMPs (including MMP-2), which favor a higher pH
(∼7.5).18 By analogy, we propose that the selectivity of the MPIs
reported here is due to the greater acidity of the pyrone versus
hydroxamate chelator (∆pKa ∼ 1).19 These results suggest that the
ZBG, and not only the MPI backbone, can provide selectivity
between different MMPs without compromising potency. The
investigation of additional MPIs, with a range of pKa’s against a
wider range of MMPs is ongoing, to interrogate the aforementioned
hypothesis more rigorously.
Table 1. IC50 Values (µM) for MPIs Against MMP-1, MMP-2, and
MMP-3: LUDI Scores for MMP-3 (PDB code 1G4K) are Shown
inhibitor
MMP-1
MMP-2
MMP-3
LUDI score
AM-1
AM-2
AM-3
AM-4
AM-5
AM-6
>50
>50
>50
>50
>50
>50
36(5)
9.3(0.5)
27(2)
>50
0.61(0.01)
>50
>50
NSa
600
NSa
440
640
700
0.24(0.01)
36(1)
2.4(0.2)
0.010(0.002)
0.019(0.002)
a NS ) no score; no acceptable conformations were found.
4-cyanophenylboronic acid and 4-biphenylboronic acid, respec-
tively, as an intermediate step.
The ability of AM-5 and AM-6 to inhibit invasion of neonatal
rat cardiac fibroblasts through a collagen membrane was examined,
as a gauge of the in vivo potential of these MPIs. At a concentration
of 250 nM, the two inhibitors were found to reduce invasion by
67% (AM-5) and 55% (AM-6) (Figure S3). In summary, we have
demonstrated that the use of pyrone ZBGs results in more potent
inhibitors than those produced with the widely employed hydrox-
amate group. Our results also indicate that the use of a non-
hydroxamate ZBG reveals a novel route to MMP inhibitor
selectivity. Overall, the findings reported here suggest that a
chelator-driven approach to metalloprotein drug design can produce
potent and selective metalloprotein inhibitors.
The inhibitory activity of compounds AM-1 through AM-6 was
evaluated using a fluorescence-based assay;14 the IC50 values are
listed in Table 1. AM-2, AM-5, and AM-6 were the most potent
compounds against MMP-3, with IC50 values in the nanomolar
range. The IC50 values against MMP-3 correlate well with the scores
obtained for each fragment using the program LUDI. Although the
LUDI scores do not perfectly parallel the relative inhibitory activity,
the approach used here does clearly distinguish between poor,
moderate, and exceptional MPIs.
Interestingly, the pyrone-based MPIs presented here are more
potent than the analogous hydroxamate-based inhibitors,13 which
is contrary to the accepted dogma that hydroxamic acids are the
best ZBGs.15 As expected, the effects of linker length (compare
AM-1, AM-2, and AM-3) and backbone substituents (AM-5
relative to AM-2) are consistent with analogous hydroxamate-based
MPIs.13 These results strongly support the concept that ZBGs equal
or superior to hydroxamates can be identified and utilized in novel
MPI designs.6,16
Acknowledgment. We thank Prof. F. Villarreal (U.C.S.D.) for
supplying us with neonatal rat cardiac fibroblasts. This work was
supported by the U.C.S.D., a Hellman Faculty Scholar award, a
Cottrell Scholar award, and the American Heart Association
(0430009N) to S.M.C. Other support was provided by NIH Grant
GM-72129-01 (D.T.P.), the LJIS program (J.M.), and NIH, NSF,
NBCR, and Accelrys (J.A.M.).
The observed trends in the IC50 values of the MPIs described
here against MMP-3 suggest that the large aromatic backbone
substituents of these compounds occupy the S1′ subsite. This
hypothesis was further examined by determining the selectivity of
these compounds against different MMPs. Traditionally, the
incorporation of bulky groups directed toward the S1′ pocket results
in selectivity over MMP-1, which has a shallow S1′ pocket.1 All
six MPIs were found to be poor inhibitors of MMP-1 (Table 1).
The poor activity of these compounds against MMP-1 is wholly
consistent with the aryl backbone groups occupying the S1′ pocket,
which supports the LUDI results (Figure 1) and ZBG orientation
predicted by our bioinorganic modeling studies.9
The inhibitors were also tested for potency against MMP-2. Like
MMP-3, MMP-2 has a deep S1′ pocket, and potency against these
two enzymes is expected to be comparable, as found with
hydroxamate-based MPIs.1,2 Interestingly, although AM-2, AM-4,
AM-5, and AM-6 showed a range of potencies against MMP-3,
all four compounds were substantially less potent against MMP-2.
Indeed, AM-6 showed >2500-fold selectivity for MMP-3, which,
to the best of our knowledge, is the highest selectivity reported for
an MPI for MMP-3 over MMP-2.
The observed selectivity of these compounds for MMP-3 over
MMP-2 is in contrast to the selectivity observed for most deep S1′
pocket MPIs. Hydroxamate-based MPIs that occupy the S1′ pocket
are almost exclusively more potent for MMP-2 than for MMP-3,
with few exceptions.1,2,17 MPIs reported to be selective for MMP-3
over MMP-2 generally target the S3′ subsite;17 however, on the
basis of the LUDI docking, the MPIs presented here have no
significant interactions in the S3′ subsite and, indeed, give similar
LUDI scores when docked to MMP-2 or MMP-3 (Table S1, Figure
S2). Therefore, it is plausible that the observed selectivity originates
from the pyrone ZBG. It has been reported that more acidic ZBGs,
such as carboxylates (a weaker ZBG than the hydroxamate),15 are
Supporting Information Available: Complete refs 5 and 13,
Figures S1-S3, Table S1, and experimental details for syntheses,
assays, and computational work. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Skiles, J. W.; Gonnella, N. C.; Jeng, A. Y. Curr. Med. Chem. 2004, 11,
2911-2977.
(2) Whittaker, M.; Floyd, C. D.; Brown, P.; Gearing, A. J. H. Chem. ReV.
1999, 99, 2735-2776.
(3) Coussens, L. M.; Fingleton, B.; Matrisian, L. M. Science 2002, 295, 2387-
2392.
(4) Breuer, E.; Frant, J.; Reich, R. Expert Opin. Ther. Pat. 2005, 15, 253-
269.
(5) Hajduk, P. J. et al. J. Med. Chem. 2002, 45, 5628-5639.
(6) Puerta, D. T.; Cohen, S. M. Curr. Top. Med. Chem. 2004, 4, 1551-1573.
(7) Liu, Z. D.; Piyamongkol, S.; Liu, D. Y.; Khodr, H. H.; Lu, S. L.; Hider,
R. C. Bioorg. Med. Chem. 2001, 9, 563-573.
(8) Finnegan, M. M.; Rettig, S. J.; Orvig, C. J. Am. Chem. Soc. 1986, 108,
5033-5035.
(9) Puerta, D. T.; Cohen, S. M. Inorg. Chem. 2003, 42, 3423-3430.
(10) Chen, L.; Rydel, T. J.; Gu, F.; Dunaway, C. M.; Pikul, S.; McDow
Dunham, K.; Barnett, B. L. J. Mol. Biol. 1999, 293, 545-557.
(11) Puerta, D. T.; Schames, J. R.; Henchman, R. H.; McCammon, J. A.; Cohen,
S. M. Angew. Chem., Int. Ed. 2003, 42, 3772-3774.
(12) Bo¨hm, H. J. J. Comput. Aided Mol. Des. 1994, 8, 623-632.
(13) Hajduk, P. J. et al. J. Am. Chem. Soc. 1997, 119, 5818-5827.
(14) Knight, C. G.; Willenbrock, F.; Murphy, G. FEBS Lett. 1992, 296, 263-
266.
(15) Castelhano, A. L.; Billedeau, R.; Dewdney, N.; Donnelly, S.; Horne, S.;
Kurz, L. J.; Liak, T. J.; Martin, R.; Uppington, R.; Yuan, Z.; Krantz, A.
Bioorg. Med. Chem. Lett. 1995, 5, 1415-1420.
(16) Puerta, D. T.; Lewis, J. A.; Cohen, S. M. J. Am. Chem. Soc. 2004, 126,
8388-8389.
(17) Fray, M. J.; Dickinson, R. P.; Huggins, J. P.; Occleston, N. L. J. Med.
Chem. 2003, 46, 3514-3525.
(18) Johnson, L. L.; Pavlovsky, A. G.; Johnson, A. R.; Janowicz, J. A.; Man,
C.-F.; Ortwine, D. F.; Purchase, C. F., II; White, A. D.; Hupe, D. J. J.
Biol. Chem. 2000, 275, 11026-11033.
(19) Gorden, A. E. V.; Xu, J.; Raymond, K. N.; Durbin, P. Chem. ReV. 2003,
103, 4207-4282.
JA054558O
9
J. AM. CHEM. SOC. VOL. 127, NO. 41, 2005 14149