A novel series of highly selective inhibitors of MMP-3

Article abstract of DOI:10.1016/j.bmcl.2007.10.042

The design and synthesis of a series of highly selective hydroxamate inhibitors of stromelysin-1 (MMP-3) is described. Substitution of a 4-biaryl piperidine sulfonamide core, which binds at the S1′ subsite of MMP-3, was optimised to give potent inhibitors of MMP-3, with greater than 300-fold selectivity over MMP-1, MMP-2, MMP-9 and MMP-14. Compounds 26 and 27 were identified as having the best balance of pharmacology and properties required for topical drug delivery.

Full text of DOI:10.1016/j.bmcl.2007.10.042

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 17 (2007) 6750–6753
A novel series of highly selective inhibitors of MMP-3
Gavin A. Whitlock,* Kevin N. Dack, Roger P. Dickinson and Mark L. Lewis
Sandwich Chemistry Department, Pfizer Global Research and Development, Sandwich Labs, Ramsgate Road,
Sandwich, Kent CT13 9NJ, UK
Received 20 September 2007; revised 11 October 2007; accepted 13 October 2007
Available online 17 October 2007
Abstract—The design and synthesis of a series of highly selective hydroxamate inhibitors of stromelysin-1 (MMP-3) is described.
Substitution of a 4-biaryl piperidine sulfonamide core, which binds at the S10 subsite of MMP-3, was optimised to give potent inhib-
itors of MMP-3, with greater than 300-fold selectivity over MMP-1, MMP-2, MMP-9 and MMP-14. Compounds 26 and 27 were
identified as having the best balance of pharmacology and properties required for topical drug delivery.
Ó 2007 Elsevier Ltd. All rights reserved.
Matrix metalloproteinases (MMPs) are a family of zinc-
dependant endopeptidases which are involved in normal
tissue remodelling and extracellular matrix degrada-
tion.1 The overexpression of MMPs has been implicated
in a number of pathological conditions including can-
cer,2,3 arthritis4 and chronic non-healing wounds.5 In
particular, MMP-3 (stromelysin-1) over-activity has
been implicated in the pathology of chronic non-healing
wounds,6 and therefore selective MMP-3 inhibition rep-
resents an attractive target for the treatment of this
condition.
Me
1 UK-370106
MMP-3 IC50 23nM
MMP-2 IC50 34.2μM
MMP-1 IC50 > 100μM
MMP-9 IC50 30.4μM
MMP-13 IC50 2.3μM
MMP-14 IC50 66.9μM
OMe
O
O
H
N
HO
N
H
O
Me
Me
Me
Figure 1. MMP pharmacology of UK-370106.
posed a new sulfonamide template where the nitrogen
and sulfur were reversed (Fig. 2).
Co-workers at Pfizer recently described the discovery of
UK-370106 1 (Fig. 1),7 a highly selective peptidic MMP-
3 inhibitor, which was identified as a clinical candidate
for the topical treatment of chronic dermal ulcers. As
part of a multi-template approach to selective MMP-3
inhibitors, we now wish to describe our efforts to dis-
cover non-peptidic MMP-3 inhibitors with good selec-
tivity over other MMPs, which may have utility in the
topical treatment of chronic non-healing wounds.8 In
this paper we describe the SAR for MMP-3 potency,
selectivity over other MMPs along with strategies to de-
liver good physicochemical properties for topical drug
candidates.
This strategy could deliver a template which was
novel,10 synthetically simple to make and lacked chiral
centres. By using the S10 SAR for selectivity from the
UK-370106 series combined with in silico modelling,
we hoped to rapidly identify a series of reversed sulfon-
amides with good selectivity for MMP-3 inhibition.
The synthesis of analogues 4–30 was accomplished by
coupling of the appropriate amine with the sulfonyl
chloride 211 to give sulfonamides 3. When the P1 group
remained as hydrogen, direct condensation of the ester
with hydroxylamine then afforded the desired hydroxa-
mic acids 4–23 (Scheme 1). In the case where P1 was
gem-dimethyl, dialkylation of 3 was then followed by
hydrolysis and coupling with hydroxylamine to give
analogues 24–30.
With the advent of non-selective MMP inhibitors based
on sulfonamides derived for a-amino acids,9 we pro-
Keywords: MMP-3; Stromelysin; Selectivity; Topical delivery; Wound
healing.
Our initial medicinal chemistry strategy was to find the
optimal sulfonamide linker to extend a lipophilic biphe-
*
Corresponding author. Tel.: +44 1304 649174; fax: +44 1304
651987; e-mail: gavin.whitlock@pfizer.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.10.042

G. A. Whitlock et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6750–6753
6751
S1' Pocket
S1' Pocket
P1'
S1' Pocket
R1
OMe
Zn
HO
Zn
HO
Zn
HO
O
O
H
N
O
O
N
O
P3'
O
S
O
N
N
H
N
H
O
S
R2
N
H
R
N
H
O
P1
O
P2'
H
N
P1
P1
H
N
H
N
α-Amino acid sulfonamide template
'Peptidic' Succinyl Template
New Inhibitor Template (4-30)
Figure 2. Evolution of ‘Reversed’ sulfonamide template.
Table 1. In vitro inhibition of MMP-3 activitya,b for compounds 5–12
R1
O
O
R1
O
O
N
O
Cl
O
(a) 30-50%
S
R2
S
O
O
N
MeO
MeO
R1
R2
R2
HO
S
2
3
N
H
N
H
O
(b) 40-60%
R1
Compound NR1R2
MMP-3 IC50 (nM)
378
(c), (d), (e)
20-50%
R1
O
O
O
O
N
O
N
O
4
HO
S
R2
HO
S
R2
N
N
N
H
Me
H
Me Me
4-23
24-30
5
164
Scheme 1. Reagents and conditions: (a) DBU, CH2Cl2, 0 °C to rt; (b)
HONH2ÆHCl, K2CO3, MeOH, reflux; (c) MeI, K2CO3, DMSO, rt; (d)
NaOH, MeOH, H2O, reflux; (e) HATU, N-ethyl-diisopropylamine,
HONH2ÆHCl, NMP, rt.
N
Me
N
6
174
26
Me
nyl unit into the S10 pocket (Table 1), as this region of
UK-370106 delivered both MMP-3 potency and
MMP-2 selectivity. Flexible alkyl linkers in compounds
4, 5 and 6 afforded modest MMP-3 potency, with the
chain length having no significant effect on MMP-3 po-
tency. Cyclisation and conformational restriction was
then investigated. Piperidine 7 was significantly more
potent than the acyclic analogue 6, and a further in-
crease in potency was achieved with dihydropiperidine
8. The azetidines 9 and 10 lost MMP-3 potency, indicat-
ing that this linker was either too short, or conforma-
tionally suboptimal, to extend the biphenyl into the
S10 pocket. Interestingly a significant degree of potency
was also lost when a phenoxyphenyl substituent was
incorporated (compound 11), indicating the enzyme is
also sensitive to the linearity of the P10 group.
7
N
N
8
6
9
161
84
N
10
O
N
O
11
205
N
a See Ref. 12 for description of assay conditions.
b MMP IC50 values are geometric means of at least three experiments.
Compounds 4–11 were also assessed for their MMP-2
activity,13 however we found no separation of activity.
We then investigated substitution on the biphenyl as a
strategy for delivering selectivity for MMP-3 over
MMP-2 (Table 2). This approach has been successfully
employed during the discovery of UK370106 1. Incor-
poration of an R3 substituent in the 3-position of the
A-ring was initially investigated whilst keeping the distal
B-ring unsubstituted.
creased, but activity against MMP-2 was even more sen-
sitive. Based on its good balance of MMP-3 potency and
MMP-2 selectivity, the methyl derivative 15 was selected
for further investigation of substitution of the distal phe-
nyl, ring B.
Modelling of 15 into the published MMP-2 catalytic do-
main crystal structure14 suggested substitution in the 3-
position of ring B could induce unfavourable interac-
tions with the loop forming the S10 pocket and therefore
The MMP-3 inhibitory potency of compounds 12–17
tended to decrease as the size of the R3-group was in-

6752
G. A. Whitlock et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6750–6753
Table 2. In vitro inhibition of MMP-3 amd MMP-2 activitya,b for
compounds 12–23
formulations and therefore aqueous solubility and auto-
clave stability were important parameters to measure.16
Compound 21 was found to have very low aqueous sol-
ubility (ꢀ1 lg/ml) and poor autoclave solution stability
across a range of pH. Hydrolysis of the hydroxamic acid
was found to be a major degradation pathway, as was
oxidation of the tetrahydropiperidine ring, followed by
fragmentation of the N–S bond of the sulfonamide.
We reasoned that by introducing steric bulk next to
the hydroxamic acid, and by reducing the dihydropiperi-
dine ring to a piperidine we may improve stability. To
increase solubility, we focused on adding polar solubilis-
ing groups onto the 3-OEt group of compound 21.
These design principles led to a series of further ana-
logues (Table 3).
R3
Ring B
R4
Ring A
O
O
N
O
HO
S
N
H
Compound R3
R4
MMP-3 IC50 MMP-2 IC50
(nM)
(nM)
12
13
14
15
16
17
H
H
H
H
H
H
H
6
3
9
17
F
Cl
55
16
620
320
Me
Et
420
960
4380
3080
Incorporation of steric bulk adjacent to the hydroxamic
acid increased MMP-3 potency (examples 21 vs 24), and
retained MMP-2 selectivity. Encouragingly this addi-
tional steric bulk also improved solution autoclave sta-
bility. We then investigated polar additions onto the
R4 substituent. Methoxyethoxy, hydroxyethoxy and
aminoethoxy substitutions also increased potency with-
out compromising MMP-2 selectivity. Examples 26
and 27 had improved aqueous solubility (1–2 lg/ml)
and excellent solution stability. Incorporation of a pyr-
idyl ring in example 27 also retained excellent MMP-3
activity and slightly improved MMP-2 selectivity. Solu-
bility was increased significantly for the basic analogues
28–30, however these compounds suffered from poor
CF3
18
19
20
21
22
23
Me
Me
Me
Me
Me
Me
Me
4
5
776
222
OMe
Et
OEt
31
4
1208
998
CH2OMe
OCF3
3
51
196
173
a See Ref. 12 for description of assay conditions.
b MMP IC50 values are geometric means of at least three experiments.
reduce MMP-2 enzyme inhibition (Fig. 3). This loop re-
gion is three residues shorter in MMP-2 compared to
MMP-3.
This strategy gave a number of compounds with im-
proved potency and selectivity, that is, substitution not
only reduced inhibition of MMP-2 but also increased
MMP-3 enzyme inhibition. In particular 3-OEt ana-
logue 21 was selected for further evaluation.
Table 3. In vitro inhibition of MMP-3 amd MMP-2 activitya,b for
compounds 24–30
Me
R4
X
Having identified 21 as a highly potent and selective
MMP-3 inhibitor, we used this compound to assess
the series against key topical drug-like properties. We
envisaged that MMP-3 inhibitor drug candidates would
be delivered topically as sterile solution or suspension
O
O
N
O
HO
S
N
H
Me Me
Compound R4
X
MMP-3IC50 MMP-2IC50
(nM)
(nM)
24
25
26
27
28
29
30
OEt
CH
CH
CH
N
2
3
1
1
457
853
262
529
188
196
534
OCH2CH2OMe
OCH2CH2OH
OCH2CH2OH
OCH2CH2NH2
CH 0.4
OCH2CH2NHMe CH 0.3
OCH2CH2NMe2 CH
1
a See Ref. 12 for description of assay conditions.
b MMP IC50 values are geometric means of at least three experiments.
Table 4. In vitro inhibition of MMP-1, MMP-9 and MMP-14 activitya,b
for compounds 26 and 27
Compound
MMP-1 IC50
(nM)
MMP-9 IC50
(nM)
MMP-14 IC50
(nM)
26
27
3230
14,000
406
2420
1710
20,100
a See Ref. 7b for description of assay conditions.
b MMP IC50 values are geometric means of at least two experiments.
Figure 3. Docking of compound 15 (gold) into X-ray structures of
catalytic domain MMP-315 (green) and MMP-2 (blue).

G. A. Whitlock et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6750–6753
6753
8. Dack, K. N.; Dickinson, R. P.; Lewis, M. L.; Whitlock, G.
A. Abstracts of Papers, 222nd ACS National Meeting,
2001, MEDI-260.
aqueous stability. Further MMP selectivity screening for
examples 26 and 27 demonstrated their excellent selec-
tivity for MMP-3 over MMP-1, MMP-9 and MMP-14
(Table 4).
9. MacPherson, L. J.; Bayburt, E. K.; Capparanelli, M. P.;
Carroll, B. J.; Goldstein, R.; Justice, M. R.; Zhu, L.; Hu,
S.; Melton, R. A.; Fryer, L.; Goldberg, R. L.; Doughty, J.
R.; Spirito, S.; Blancuzzi, V.; Wilson, D.; O’Byrne, E. M.;
Ganu, V.; Parker, D. T. J. Med. Chem. 1997, 40, 2525.
10. (a) Dack, K. N.; Whitlock, G. A. WO9929667; (b) Dack,
K. N.; Whitlock, G. A. EP931788; (c) Dack, K. N.; Fray,
M. J.; Whitlock, G. A.; Lewis, M. L.; Thomson, N. M.
WO2000074681.
In summary, we have described the discovery of a no-
vel series of highly selective inhibitors of MMP-3. By
reversing the nitrogen and sulfur of previous non-
selective MMP inhibitor series, we were able to work
in a novel area of chemical space and in a simplified
template without chiral centres. Identification of the
optimal linker for the key pharmacophoric biaryl S10
substituent was rapidly achieved, and using learning
and selectivity SAR from an unrelated peptidic series,
we were able to deliver compounds which were highly
selective for MMP-3. Drug-like properties for topical
delivery, namely solution autoclave stability and solu-
bility, were also measured during the evolution of the
series. It was found that steric bulk adjacent to the
hydroxamic acid, and reduction of a tetrahydropiperi-
dine to a piperidine significantly improved aqueous
autoclave stability. Polar groups could be tolerated
on the distal aryl ring of the biaryl which improved
solubility. The combination of these features led to
compounds 26 and 27 which had the best balance of
pharmacological and physicochemical properties as
potential candidates for the topical treatment of
chronic dermal ulcers.
11. Oliver, J. E.; DeMilo, A. B. Synthesis 1975, 5, 321.
12. The MMP-2 and MMP-3 enzyme inhibition assays are
based on the original protocol described by Knight et al.17
with the modifications described below. Catalytic domain
MMP-2 and MMP-3 were prepared at Pfizer Global
R&D. A stock solution of MMP-2 or MMP-3 (1 lM) was
activated by the addition of aminophenylmercuric acetate
(APMA, 1 mM for MMP-2, 2 mM for MMP-3) followed
by incubation at 37 °C (1 h for MMP-2, 3 h for MMP-3).
The enzymes were then diluted in Tris–HCl assay budder
to a concentration of 10 nM. The final assay concentration
of enzyme used in the assays was 0.1 nM. The fluorogenic
substrate used in these assays was Mca-Arg-Pro-Lys_Tyr-
Ala-Nva-Trp-Met-Lys(Dnp)-NH216 (Bachem Ltd). This
substrate was selected because it had balanced hydrolysis
rates against MMP-2 and MMP-3 (kcat/km 54,000 and
59,400 sꢁ1Mꢁ1, respectively). The final substrate concen-
tration used in the assay was 5 lM. Test compounds were
dissolved in DMSO then diluted with test buffer solution
(as above) so that not more than 1% DMSO was present.
Test compound and enzyme were added to each well of a
96-well plate and allowed to equilibrate for 15 min at
37 °C prior to addition of substrate. Plates were then
incubated for 1 h at 37 °C prior to determination of
fluorescence using a fluorimeter (Fluostar, BMG Lab-
Technologies) at an excitation wavelength of 328 nm and
emission wavelength of 393 nm. The potency of inhibitors
was measured from the amount of substrate cleavage
obtained from a range of test compound concentrations,
and IC50 values were calculated from the resulting
concentration–response curves.
Acknowledgments
We thank Drs. Nick Occleston, John Huggins and Mike
Collis and their teams (Discovery Biology Department)
for screening data and useful discussions, and Debbie
Lovering, Mike Closier, Paul Bradley, Paul Allen and
Kerry Paradowski for compound synthesis. We are also
grateful to Charlotte Reed and Dawn McCleverty of the
Pharmaceutical Sciences Department for solubility and
autoclave stability data.
13. MMP-2 inhibition data for compounds 4–11 demon-
strated that selectivity for MMP-3 over MMP-2 was <2-
fold.
14. (a) Dhanaraj, V.; Williams, M. G.; Ye, Q.-Z.; Molina, F.;
Johnson, L. L.; Ortwine, D. F.; Pavlovsky, A.; Rubin, J.
R.; Skeean, R. W.; White, A. D.; Humblet, C.; Hupe, D.
J.; Blundell, T. L. Croat. Chem. Acta 1999, 72, 575; The
crystal structure of full length MMP-2 has also been
described (b) Morgunova, E.; Tuuttila, A.; Bergmann, U.;
Isupov, M.; Lindqvist, Y.; Schneider, G.; Tryggvason, K.
Science 1999, 284, 1667.
References and notes
1. Woessner, J. F. FASEB J. 1991, 5, 2145.
2. Coussens, L. M.; Werb, Z. Chem. Biol. 1996, 3, 895.
3. Yu, A. E.; Hewitt, R. E.; Conner, E. W.; Stetler-
Stevenson, W. G. Drugs Ageing 1997, 11, 229.
4. Walakovitis, L. A.; Bahardwaj, N.; Gallick, G. S.; Lark,
M. W. Arthritis Rheum. 1992, 35, 35.
5. Tarnuzzer, R. W.; Macauley, S. P.; Mast, B. A.; Gibson, J.
S.; Stacey, M. C.; Trengrove, N.; Moldawer, L. L.;
Burslem, F.; Schultz, G. S. Growth Factors and Wound;
Chapter 12, pp 206–228.
6. (a) Mast, B. A.; Schultz, G. S. Wound Rep. Reg. 4, 411; (b)
Saarialho-Kere, U. K.; Pentland, A. P.; Birkedal-Hansen,
H.; Parks, W. C.; Welgus, H. G. . J. Clin. Invest. 1994, 94,
79; (c) Vaalamo, M.; Weckroth, M.; Puolakkainen, P.;
Kere, J.; Saarinen, P.; Lauharanta, J.; Saarialho-Kere, U.
K. Br. J. Dermatol. 1996, 135, 52.
7. (a) Fray, M. J.; Dickinson, R. P. Bioorg. Med. Chem.
Lett. 2001, 11, 571; (b) Fray, M. J.; Dickinson, R. P.;
Huggins, J. P.; Occleston, N. L. J. Med. Chem. 2003,
46, 3514.
15. Docking of compound 15 into the MMP-3 active site was
based on the X-ray structure of a close analogue bound to
catalytic domain MMP-3 (unpublished data).
16. Aqueous solubility was measured at pH 1, 5, 7, 9 and 11.
Solution stability was measured after 15 min autoclave at
120 °C across the same range of pH. For a compound to
be suitable as a candidate for topical delivery, it was felt
that aqueous solubility needed to be >1 lg/ml to enable
sufficient exposure within the site of action. In addition
<5% degradation of parent after autoclave stability
assessment was required for this sterilisation technique
to be viable.
17. Knight, C. G.; Wilenbrock, F.; Murphy, G. FEBS Lett.
1992, 296, 263.