Heterocycle-based MMP inhibitors with P2′ substituents

Article abstract of DOI:10.1016/S0960-894X(01)00137-8

Potent and selective inhibition of matrix metalloproteinases was demonstrated for a series of sulfonamide-based hydroxamic acids. The design of the heterocyclic sulfonamides incorporates a six- or seven-member central ring with a P2′ substituent that can be modified. Binding interactions of this substituent at the S2′ site are believed to contribute to high inhibitory potency against stromelysin, collagenase-3 and gelatinases A and B, and to provide selectivity against collagenase-1 and matrilysin. An X-ray structure of a stromelysin-inhibitor complex was obtained to provide insights into the SAR and selectivity trends observed for the series.

Full text of DOI:10.1016/S0960-894X(01)00137-8

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1009–1013
Heterocycle-Based MMP Inhibitors with P20 Substituents
Stanislaw Pikul,* Kelly McDow Dunham, Neil G. Almstead, Biswanath De,
Michael G. Natchus, Yetunde O. Taiwo, Lisa E. Williams, Barbara A. Hynd,
Lily C. Hsieh, Michael J. Janusz, Fei Gu and Glen E. Mieling
Procter and Gamble Pharmaceuticals, Health Care Research Center, 8700 Mason-Montgomery Road, Mason, OH 45040, USA
Received 27 July 2000; accepted 16 February 2001
Abstract—Potent and selective inhibition of matrix metalloproteinases was demonstrated for a series of sulfonamide-based hydro-
xamic acids. The design of the heterocyclic sulfonamides incorporates a six- or seven-member central ring with a P20 substituent
that can be modified. Binding interactions of this substituent at the S20 site are believed to contribute to high inhibitory potency
against stromelysin, collagenase-3 and gelatinases A and B, and to provide selectivity against collagenase-1 and matrilysin. An
X-ray structure of a stromelysin–inhibitor complex was obtained to provide insights into the SAR and selectivity trends observed
for the series. # 2001 Elsevier Science Ltd. All rights reserved.
Matrix metalloproteinases (MMPs) are a family of zinc-
containing enzymes that are capable of degrading many
proteinaceous components of the extracellular matrix.1
Under normal physiological conditions, the proteolytic
activities of the enzymes are controlled by tissue inhibi-
tors of matrix metalloproteinases (TIMPs).2 In patho-
logical conditions this balance is shifted towards
overactivation of MMPs leading to excessive degrada-
tion of the matrix components. Excessive MMP activity
has been implicated in numerous disease states invol-
ving matrix degradation, which include arthritis,3 can-
cer,4 and periodontal diseases.5 Consequently, there has
been a significant interest in developing MMP inhibitors
that may control the aberrant activity of MMP
production.6
with the S20 binding site could provide an opportunity
to modify not only the potency but also the selectivity of
such inhibitors against various MMPs. Since inhibition
of collagenase-1 (MMP-1) may be responsible for the
musculoskeletal side effects observed clinically with
broad-spectrum inhibitors, we were especially interested
in molecules that would spare this enzyme. Modification
of the P20 substituent could provide such selectivity due
to the hydrophilic nature of the S20 site of MMP-1
containing Ser-222 as compared to the hydrophobic
nature of our target enzymes: stromelysin (MMP-3) and
collagenaase-3 (MMP-13), which contain Leu-222 and
Ile-222, respectively. Similar effects were expected for
matrilysin (MMP-7), which has a hydrophilic threonine
in the 222 position. In an effort to provide a heterocyclic
template capable of additional P20 binding interactions,
we designed a new series of MMP inhibitors based on a
1,5-piperazine/diazepine framework as represented by
structure B. In this paper, we wish to report our early
SAR work that produced a series of potent inhibitors
with a unique range of selectivities for various MMPs.
The Design Concept ofConstrained Inhibitors
Earlier work at P&GP7,8 and other laboratories had
demonstrated that a central, heterocyclic ring could be
effectively used in the design of potent MMP inhibitors
(see structure A below). For many of these inhibitors
the zinc binding group (ZBG) and the P10 substituent
appear to be primarily responsible for the binding
interactions with the enzymatic target. We believed that
an additional ring substituent capable of interacting
*Corresponding author at current address: Avalon Pharmaceuticals,
19 Firstfield Road, Gaithersburg, MD 20878, USA. Fax: +301-556-
1100; e-mail: spikul@avalonrx.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00137-8

1010
S. Pikul et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1009–1013
Chemistry
derivatized using methoxybenzene sulfonyl chloride.
The resulting intermediate was treated with tri-
fluoroacetic acid to produce the key intermediate amino
acid 8. The P20 substituent was introduced by acylation
or sulfonylation of the amino group under standard
Schotten–Baumann conditions. Finally, the carboxylic
group was converted into the hydroxamic functionality
through intermediate formation of the corresponding acid
chloride, which, upon treatment with hydroxylamine,
provided the target inhibitor 9.10
Compounds discussed below were prepared using syn-
thetic routes shown in Schemes 1 and 2. An orthogon-
ally protected d-aspartic acid 1 served as a common
starting material for the synthesis of the 6-oxohexa-
hydropyrimidines 4 and the 7-oxo-[1,4]diazepine 6.
First, the free amino group of 1 was derivatized using
4-methoxybenzene sulfonyl chloride, and then the tert-
butyl ester was cleaved with trifluoroacetic acid. The
resulting carboxylic acid 2 was converted to amide 3
using the requisite amine component in the EDAC
mediated coupling reaction. Compounds 3a–f, upon
treatment with trioxane and catalytic amounts of sul-
furic acid in methylene chloride, cyclized cleanly to form
the 6-oxohexahydropyrimidine ring. Exposure of the
resulting intermediates to hydroxylamine in basic
methanol provided the target hydroxamic acids 410 in
good overall yield.11,12
Results and Discussion
All compounds were tested for the inhibition of various
MMP enzymes including collagenases-1, -2, and -3
(MMP-1, -8, and -13, respectively), gelatinases A and B
(MMP-2 and -9, respectively), stromelysin (MMP-3)
and matrilysin (MMP-7) and the data are summarized
in Tables 1 and 2. First, we explored a series of 6-oxo-
hexahydropyrimidines containing a six-member central
ring. We chose this ring size after observing potent
inhibition of MMPs in a related series of achiral com-
pounds.7 Compound 4a was prepared to test the intrin-
sic inhibitory activity associated with its novel ring
structure. Its inhibitory activity against MMP-13 and
MMP-2 was found to be comparable to that observed
for the reference compound CGS27023A. However, it
was interesting to see a significant improvement in
selectivity against MMP-1 and MMP-7. Next, the
N-methyl group of 4a was replaced by increasingly
bulky alkyl substituents (compounds 4b–d) in order to
improve contact with the S20 pocket. As expected, we
observed progressively more potent inhibition of all
tested enzymes with the most significant response from
Alternatively, carboxylic acid 2 was coupled with an
appropriate aminoalcohol derived from the corre-
sponding l-amino acid to form hydroxysulfonamide 5.
Formation of the desired seven-member ring was
cleanly accomplished via a Mitsunobu reaction. The
intermediate methyl esters were converted into their
corresponding hydroxamic acids10 using standard con-
ditions as described above.
A known racemic [1,4]diazepane 713 was used as the
starting material in the synthesis of compounds 9a–c
(Scheme 2). The two amino groups of 7 were differen-
tially functionalized in a two-step, one-pot sequence.
First, the distal amino group was selectively protected
using Boc2O and then the proximal amino group was
Scheme 1. Reagents and conditions: (a) MeOC6H4SO2Cl, Et3N, CH2Cl2; (b) TFA, CH2Cl2; (c) RNH2, EDAC, HOBT, DMF; (d) trioxane, H2SO4,
CH2Cl2; (e) NH2OH, KOH, MeOH; (f) R1NHCHR2CH2OH, EDAC, HOBT, DMF; (g) DEAD, Ph3P, CH2Cl2.
Scheme 2. Reagents and conditions: (a) Boc2O, NaOH, dioxane–water, then Et3N followed by MeOC6H4SO2Cl; (b) TFA, CH2Cl2; (c) CbzCl, Et3N,
dioxane–water or ArSO2Cl, Et3N, dioxane–water; (d) (COCl)2, DMF, CH2Cl2, then Et3N followed by NH2OH.

S. Pikul et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1009–1013
1011
stromelysin. Compound 4d containing a tert-butyl P20
group was found to have 1 nM potency for MMP-2, -3,
-8, -9, and subnanomolar potency for MMP-13. At the
same time the inhibition of MMPs-1 and -7 was not
highly affected, producing a considerable improvement
in selectivity (e.g., MMP-1/MMP-3 selectivity is 5 for 4a
vs 246 for 4d; MMP-1/MMP-13 selectivity is 278 for 4a
vs 941 for 4d). Introduction of the benzyl group in 4e
brought a noticeable improvement in inhibition of
MMP-1, while inhibition of MMP-7 decreased mark-
edly. This result and significantly lower overall potency
of 4f, containing P20 substituent with a heteroatom,
pointed to relatively low tolerance for structural varia-
tion of the P20 substituent.
MMP-7 increased significantly compared to the six-
member homologue 4b. A marked improvement in
potency against MMP-7 was also observed for com-
pound 9b with an N-benzyl substituent, while the inhi-
bition of other MMPs was only slightly improved. The
broad-spectrum character of these seven-member ring-
based inhibitors was confirmed by compound 6c derived
from l-prolinol. Potent inhibition of all MMP enzymes
observed for these seven-member ring-based molecules
was in stark contrast to the high selectivity displayed by
the six-member ring-based analogues.
To investigate the effect of increased flexibility of the
central ring on inhibitory potency we decided to remove
the ring oxo substituent. The distal amino group was
then functionalized to form carbamate 9a or sulfona-
mides 9b and 9c. Interestingly, all three analogues
showed increased inhibition of MMP-1 and decreased
inhibition of MMP-7. The inhibition of other enzymes
Next, we turned our attention to inhibitors containing a
[1,4]diazepine central ring and prepared compound 6a.
It was interesting to notice that while inhibition of
MMP-13 was not affected, the inhibition of MMP-1 and
Table 1. SAR of 6-oxohexahydropyrimidines
Compound
R
IC50 (nM)14
MMP-1
MMP-2
MMP-3
MMP-7
MMP-8
MMP-9
MMP-13a
4a
4b
4c
4d
Me
iso-Pr
cyclo-Hex
tert-Bu
CH2Ph
2200
643
465
8.8
5.3
0.7
1.0
1.0
3.4
9.1
404
5.3
3.5
1.3
3.1
>10,000
3460
776
47.3
6.3
2.8
1.5
6.4
102
4.4
24.1
4.2
3.3
1.1
6.8
52.5
4.3
7.9
2.9
0.75
0.34
2.56
nd
320
909
4e
4f
CGS27023A
22.2
696
49.5
>10,000
>10,000
106
CH2CH2OMe
180
16.9
4.3
aData obtained using rat MMP-13; nd, not determined.
Table 2. SAR of [1,4]diazepanes
Compound
R1
R2
X
IC50 (nM)14
MMP-1
MMP-2
MMP-3
MMP-7
MMP-8
MMP-9
MMP-13a
6a
6b
iso-Pr
CH2Ph
CH2CH2CH2
Cbz
SO2C6H4OMe
H
H
O
O
O
H,H
H,H
22
13.6
1.2
97
28.9
2.4
0.68
5.2
1.6
0.5
6.5
3.3
3.4
14
369
88
nd
737
905
3.7
1.5
nd
3.3
1.8
4.1
1.0
nd
1.8
0.5
1.9
nd
2.0
2.1
0.84
6c
9ab
9bb
H
H
6.3
9cb
H
H,H
346
2.7
39.2
2780
6.6
5.5
2.4
aData obtained using rat MMP-13.
bRacemic material; nd, not determined.

1012
S. Pikul et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1009–1013
position. For some reasons, such a presence does not
seem to be well tolerated by selected MMPs. This pro-
vides additional evidence that specific geometry
requirements must be met by MMP inhibitors if high
selectivity is to be achieved through modulation of
interactions at the S20 site.
Conclusion
A new series of MMP inhibitors was designed based on
a scaffold containing a central ring of six-member
hexahydropyrimidines or seven-member [1,4]diazepines.
The distal nitrogen present in both ring systems was
used to introduce the P20 substituent which is believed
to influence inhibitory selectivity through interactions
with the S20 pocket. Compounds with hexahydropyri-
midine ring with short and rigid alkyl P20 substituents
were found to be highly potent against MMP-2, -3, -8,
-9 and -13, and very selective against MMP-1 and -7
(e.g., MMP-1/MMP-13 IC50 ratio was found to be
>200-fold). A considerable loss of selectivity was
observed for compounds containing a seven-member
[1,4]diazepine ring. An X-ray structure of a stromelysin-
inhibitor complex confirmed the predicted arrangement
of the inhibitor in the enzyme active site and helped
rationalize the observed SAR.
Figure 1. X-ray structure of a complex of 4e with truncated stromelysin.
was not significantly altered except for somewhat
weaker inhibition of MMP-2 and MMP-3 by compound
9c.
X-ray Crystal Structure
In order to better understand the interactions of this
new series of MMP inhibitors with stromelysin we have
obtained X-ray crystal data of compound 4e with trun-
cated stromelysin (Fig. 1).15 All interactions involving
the hydroxamic acid group, the P10 methoxyphenyl
group and the sulfonamide system linking them were
found to be similar to those previously described.16 The
6-oxohexahydropyrimidine ring in its half-chair con-
formation provided the desired conformation placing
the hydroxamic acid in the pseudo-axial position and
the sulfone group in pseudo-equatorial position. The
ring oxo group was found to point towards the solvent
and did not develop any characteristic interactions with
the enzyme. As intended in the original design of this
series of MMP inhibitors, the alkyl group attached to
the distal nitrogen atom of the hexahydropyrimidine
ring was found to extend towards the S20 pocket. This
explains the increased potency and selectivity observed
for six-member ring-based inhibitors with substituents
rigidly attached to the ring (compounds 4a–d). The
benzylic phenyl group of 4e developed good van der
Waals contacts (within 3 A) with Leu-222 of stromely-
sin. Flexibility of the benzyl group may allow alter-
native conformations supported by favorable
interactions with neighboring amino acid residues (e.g.,
Pro-221 or Val-163). This arrangement could also save
4e from unfavorable interactions with the S20 pocket of
MMP-1, which may help explain the relatively low
selectivity of this inhibitor. Similar reasoning can be
used to explain the very good inhibitory potency, but
fairly low selectivity observed for the [1,4]diazepine-
based series. In this case, different ring size and con-
formational flexibility allow the P20 substituents to ori-
ent themselves away from the S20 pocket if necessary.
The unusual inhibitory profile of 4f containing a flexible
methoxyethyl substituent is difficult to explain and may
be attributed to the presence of a heteroatom in the P20
Acknowledgements
The authors wish to thank a number of P&GP scientists
associated with the MMP project whose hard work and
dedication over many years have contributed towards
the completion of this work.
References and Notes
1. Woessner, J. F., Jr. FASEB 1991, 5, 2145.
2. Matrisian, L. M. Trends Genet. 1990, 6, 121.
3. (a) Firestein, G. S.; Paine, M. M.; Littman, B. H. Arthritis
Rheum. 1991, 34, 1094. (b) Walakovits, L. A.; Bhardwaj, N.;
Gallick, G. S.; Lark, M. W. Arthritis Rheum. 1992, 35, 35.
4. Pyke, C.; Ralfkiaer, E.; Huhtala, P.; Hurskeinen, T.; Dano,
K.; Tryggvason, K. Cancer Res. 1992, 52, 1336.
5. Overall, C. M.; Wiebkin, O. W.; Thonard, J. C. J. Perio-
dontal Res. 1987, 22, 81.
6. (a) Whittaker, M.; Floyd, C. D.; Brown, P.; Gearing,
A. J. H. Chem. Rev. 1999, 99, 2735. (b) Greenwald, R. A.;
Zucker, S.; Golub, L. M., Eds. Ann. N.Y. Acad. Sci. 1999, 87.
7. Pikul, S.; McDow Dunham, K. L.; Almstead, N. G.; De,
B.; Natchus, M. G.; Anastasio, M. B.; McPhail, S. J.; Snider,
C. E.; Taiwo, Y. O.; Rydel, T.; Dunaway, C. M.; Gu, F.;
Meiling, G. E. J. Med. Chem. 1998, 41, 3568.
8. Almstead, N. G.; Bradley, R. S.; Pikul, S.; De, B.; Natchus,
M. G.; Taiwo, Y. O.; Gu, F.; Williams, L. E.; Hynd, B. A.;
Janusz, M. J.; Dunaway, C. M.; Meiling, G. E. J. Med. Chem.
1999, 42, 4547.
9. Levin, J. I.; DiJoseph, J. F.; Killar, L. M.; Sung, A.; Wal-
ter, T.; Sharr, M. A.; Roth, C. E.; Skotnicki, J. S.; Albright,
J. D. Bioorg. Med. Chem. Lett. 1998, 8, 2657.

S. Pikul et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1009–1013
1013
10. Satisfactory analytical data (1H NMR, 13C NMR, MS,
EA and/or HRMS) were obtained for all target hydroxamic
acids described in this paper.
11. Compounds analogous to 4 with R=H and aryloxy-
phenylsulfonyl group in the P10 position were also disclosed by
Agouron: Bender, S. L.; Melnick, M. J. US Patent 5753653,
1998; Chem. Abstr. 1998, 324822.
12. Although we have no direct data on the optical purity of
compounds 4a–f, we assume little racemization occurred dur-
ing their synthesis as we observed no sign of formation of the
second possible diastereomer during the synthesis of com-
pounds 6a–c.
13. Barraclough, P.; Farrant, R. D.; Kettle, D.; Smith, S. J.
Chem. Res., Synop. 1991, 306.
14. All in vitro inhibition assays were performed as described
previously: Pikul, S.; McDow Dunham, K. L.; Almstead,
N. G.; De, B.; Natchus, M. G.; Anastasio, M. V.; McPhail,
S. J.; Snider, C. E.; Taiwo, Y. O.; Chen, L.; Dunaway, C. M.;
Gu, F.; Mieling, G. E. J. Med. Chem. 1999, 42, 87.
15. The atomic coordinates of compound 4e bound to trun-
cated stromelysin have been deposited to the Protein Data
Bank, and the file name is 1G05.
16. Babine, E. R.; Bender, S. L. Chem. Rev. 1997, 97,
1359.