Discovery of potent and selective succinyl hydroxamate inhibitors of matrix metalloprotease-3 (stromelysin-1)

Article abstract of DOI:10.1016/S0960-894X(00)00720-4

Structure-activity relationships are described for a series of succinyl hydroxamic acids 4a-O as potent and selective inhibitors of matrix metalloprotease-3 (stromelysin-1). Optimisation of P1′ and P3′ groups gave compound 4j (MMP-3 IC₅₀ = 5.9 nM) which was >140-fold less potent against MMP-1 (IC₅₀ = 51,000 nM), MMP-2 (IC₅₀ = 1790 nM), MMP-9 (IC₅₀ = 840 nM) and MMP-14 (IC₅₀ = 1900 nM).

Full text of DOI:10.1016/S0960-894X(00)00720-4

Bioorganic & Medicinal Chemistry Letters 11 (2001) 571±574
Discovery of Potent and Selective Succinyl Hydroxamate
Inhibitors of Matrix Metalloprotease-3 (Stromelysin-1)
M. Jonathan Fray* and Roger P. Dickinson
Department of Discovery Chemistry, P®zer Global Research and Development, Sandwich, Kent CT139NJ, UK
Received 9 June 2000; revised 27 November 2000; accepted 19 December 2000
AbstractÐStructure±activity relationships are described for a series of succinyl hydroxamic acids 4a±o as potent and selective
inhibitors of matrix metalloprotease-3 (stromelysin-1). Optimisation of P10 and P30 groups gave compound 4j (MMP-3
IC50=5.9 nM) which was >140-fold less potent against MMP-1 (IC50=51,000 nM), MMP-2 (IC50=1790 nM), MMP-9
(IC50=840 nM) and MMP-14 (IC50=1900 nM). # 2001 Elsevier Science Ltd. All rights reserved.
The previous paper1 described SAR for a series of suc-
cinyl hydroxamic acids designed to optimise inhibition
of MMP-3 and selectivity over MMP-2 inhibition.
Starting from 1, appropriate choice of P30 group, for
example, 2 and 3, diminished inhibition of MMP-2 by
up to 1000-fold whilst retaining inhibition of MMP-3.
acid derivatives 6±83 to give compounds 9±11. The
t-butyl esters were cleaved and the resulting acids 12±14
converted to hydroxamic acids 4a±c.
A ¯exible route4 (Scheme 2) was employed for com-
pounds 4d±o in which a biarylpropyl group was intro-
duced into P10. The amines 5, 15 and 161 were coupled
to acid 174 as above, a€ording 18±20, and then a biaryl
group was introduced via a Heck reaction. The reaction
partners for the Heck reaction are listed below.5 The
major product (ꢀ70%) in each case was the (E)-3-
(biaryl)-3-propenyl derivative (21), but other alkene
products were evident by NMR. These were tentatively
identi®ed as the (Z)-3-(biaryl)-3-propenyl, (E)- and (Z)-
3-(biaryl)-2-propenyl and 2-(biaryl)-3-propenyl isomers
(ꢀ7% each). The last product (arising from attack of
the biaryl group on the more substituted end of the
alkene) could be removed by (a) careful chromato-
graphy at this stage, (b) following the next step, or (c)
by recrystallisation of carboxylic acid 22. The mixture
of alkenes was then reduced and the t-butyl ester
cleaved to give carboxylic acids 22, which were usually
converted to the hydroxamic acids as described pre-
viously.1 Compounds 4l, 4n and 4o were prepared from
the carboxylic acid using hydroxylamine and the pep-
tide coupling agent HATU.
However, the selectivity of 2 and 3 was modest and we
felt the limit of what could be achieved from the P30 sub-
stituent alone had been reached. We therefore decided to
synthesise P10 analogues 4a±o in the hope of gaining fur-
ther increases in potency and selectivity for MMP-3.
Results and Discussion
Chemistry
Two routes were used for the preparation of 4a±o. In
the ®rst (Scheme 1), amine 52 was coupled to succinic
Inhibition of catalytic domain MMP-3 and MMP-2 was
measured as described previously using the Nagase
¯uorogenic substrate,6 and results are shown in Table 1.
Initially, we surveyed the long chain aliphatic, biphe-
nethyl, cyclohexylpropyl and biphenpropyl derivatives
*Corresponding author. Tel.: +44-1304-640206; fax: +44-1304-
640200; e-mail: jonathan_fray@sandwich.p®zer.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00720-4

572
M. J. Fray, R. P. Dickinson / Bioorg. Med. Chem. Lett. 11 (2001) 571±574
Scheme 1. Reagents and conditions: (a) 6±8, EDC, HOAt, NMM, CH2Cl2, 0±20 ꢁC, 18 h, 80±97%; (b) HCl(g), CH2Cl2 or dioxane, 0 ꢁC, 54±89% or
TFA/H2Cl2 (1:1), 0±20 ꢁC, 4 h, 80±100%; (c) O-allylhydroxylamine, EDC, HOAt, Et3N or i-Pr2NEt, CH2Cl2 or DMF, 0±20 ꢁC, 18 h; or O-allylhy-
+
À
droxylamine, PyAOP, i-Pr2NEt, DMF, 0±20 ꢁC, 18 h, 57±95%; (d) NH4 HCO2 , 5 mol% Pd(OAc)2 2PPh3, EtOH/H2O (4:1), re¯ux, 2 h, 17±88%.
EDC=N-ethyl-N0-dimethylaminopropylcarbodiimide, NMM=N-methylmorpholine, HOAt=7-aza-1-hydroxy-1,2,3-benzotriazole, PyAOP=7-
azabenzotriazol-1-yloxytris(pyrrolidino)phosphonium hexa¯uorophosphate.
.
ꢁ
.
Scheme 2. Reagents and conditions: (a) EDC, HOAt or HOBt, NMM, DMF or CH2Cl2, 0±20 C, 18 h, 87±100%;ꢁ(b) 5 mol% Pd(OAc)2 2P(o-tol)3,
ꢁ
.
Et3N, DMF or MeCN, 100 C, 18±24 h, 58±98%; or 5 mol% Pd(OAc)2 2P(o-tol)3, Na2CO3, n-Bu4NCl, DMF, 80 C, 68±24 h, 68%; (c) H2 (3 bar),
10% Pd/C, EtOH, 20 ꢁC, 6±ꢁ18 h, 62±98%; or, for Ar=chlorobiphenyl, TsNHNH2, toluene, re¯ux, 4 h, 74%. (d)±(f) see steps b±d, Scheme 1; (g)
ꢁ
.
HATU, i-Pr2NEt, DMF, 0 C then NH2OH HCl, 0±20 C, 44±74%. HOBt=1-hydroxy-1H-1,2,3-benzotriazole, HATU=O-(7-aza-1,2,3-benzo-
triazol-1-yl)-N,N,N0,N0-tetramethyluronium hexa¯uorophosphate.
4a±d, since these groups had featured in literature MMP
inhibitors. From this set of compounds it was clear that
biphenpropyl a€orded increased selectivity and potency
compared to 1. The increase in potency can be ascribed
to a favourable interaction between the terminal phenyl
and His-2243b,7 when the biaryl is linked via a propylene
spacer group.
biphenyl moiety is tolerated by MMP-2, whereas an
ortho substituent disrupts this planarity and causes a
steric clash in the MMP-2 binding pocket. Potency for
MMP-3 inhibition was relatively una€ected. We then
sought to optimise selectivity by systematically
increasing the size of the ortho substituent (compounds
4i±m).
Suitably encouraged, we prepared analogue 4e in the
expectation that selectivity for MMP-3 would increase.
We were grati®ed to ®nd it did, although potency
against MMP-3 was reduced by approximately 3-fold.
Examination of molecular models of MMP-3 and
MMP-2 incorporating 4d suggested there might be less
room in MMP-2 to accommodate an ortho substituent
on the biaryl, since the enzyme loop which de®nes S10 is
three amino acid residues shorter in MMP-2.8 We
therefore prepared compounds 4f±h. Surprisingly,
introduction of an ortho ¯uorine into 4d to give 4f
increased selectivity to 47-fold, but 4g was only 4-fold
selective (cf. 35-fold for 4e).
Inhibition of MMP-3 and selectivity over MMP-2 was
remarkably sensitive to the size of the substituent and is
clearly optimal for a methyl group (compound 4j). Lar-
ger groups (compounds 4k±m) led to a signi®cant loss of
inhibition of both MMP-3 and MMP-2.
Having discovered a P10 group that markedly increased
selectivity, we considered whether simpli®cation of the
P30 group would be possible, and therefore prepared
compounds 4n and 4o. Although 4n was very potent,
selectivity was dramatically reduced, thereby emphasis-
ing the vital contributions made by both the P30 and P10
groups in 4j. Replacement of the a-methylbenzyl in 4m by
methyl (as in 4o) led to increased potency for inhibition of
MMP-3 and MMP-2 of approximately 19- and 600-
fold, respectively, a similar e€ect on the selectivity at a
Even more remarkable was the very low selectivity of
the ¯uorenyl derivative 4h.9 Presumably, a planar

M. J. Fray, R. P. Dickinson / Bioorg. Med. Chem. Lett. 11 (2001) 571±574
Table 1. Inhibition of MMP-3 and MMP-2 by succinyl hydroxamic acids 4a±o
573
Compd
R1
R2
MMP-3 IC50 (nM)a
25
MMP-2 IC50 (nM)a
100
Sel.b
4
a
(CH2)8CH3
(R)-CH(CH3)Ph
b
c
d
e
(R)-CH(CH3)Ph
(R)-CH(CH3)Ph
(R)-CH(CH3)Ph
CHPh2
26
120
2.6
8.8
210
920
34
8.1
7.7
13
310
35
f
(R)-CH(CH3)Ph
CHPh2
7.2
10
335
41
47
4.1
2.9
100
300
15
g
h
i
(R)-CH(CH3)Ph
(R)-CH(CH3)Ph
(R)-CH(CH3)Ph
(R)-CH(CH3)Ph
(R)-CH(CH3)Ph
(R)-CH(CH3)Ph
CH3
3.8
13
11
1300
1790
6,800
1700
52,000
1.0
j
5.9
450
110
770
0.5
40
k
l
15
m
n
o
68
2
CH3
86
2.1
aIC50s are an average of two determinations.
bSelectivity for MMP-3=MMP-2 IC50 Ä MMP-3 IC50
.
lower level of potency. The remarkable interdependence
of SARs on these two substituents can be rationalised
qualitatively from the structures of the S30 and S10
binding sites. S30 can more easily accommodate a large
substituent in MMP-3, and thus in MMP-2 a residue
may have to move. If this residue is Tyr426 then the
next amino acid (Thr427), which starts the loop which
de®nes S10, may also shift. Thus, whilst MMP-2 can
adjust its conformation to alleviate steric interactions
with S30 or S10 residues, movement to embrace both
substituents simultaneously is very dicult, and a
large drop in inhibitory potency results. Compound
4j was more widely pro®led against other MMPs (see
Table 2).10

574
M. J. Fray, R. P. Dickinson / Bioorg. Med. Chem. Lett. 11 (2001) 571±574
Table 2. Enzyme inhibition pro®le of 4j
3. (a) Compound 6: Broadhurst, M. J.; Brown, P. A.; John-
son, W. H.; Lawton, G. Eur. Pat. Appl. EP 575844, 1993;
Chem. Abstr. 1994, 121, 57993. (b) Compound 7: prepared
from 40-¯uoro-1,10-biphen-4-ylbutanoic acid (Esser, C. K. et
al. J. Med. Chem. 1997, 40, 1026) by the method of Porter, J.
R.; Beeley, N. R. A.; Boyce, B.; Mason, B.; Millican, A.;
Millar, K.; Leonard, J.; Morphy, J. R.; O'Connell, J. P.
Bioorg. Med. Chem. Lett. 1994, 4, 2741. (c) Compound 8:
prepared by catalytic hydrogenation of the phenpropyl analo-
gue (World Pat. Appl. WO 95/04033, 1995; Chem. Abstr. 1995,
122, 264925) over Rh/Al2O3 (4 bar, HOAc, 20 ꢁC, 100%).
4. Castelhano, A. L.; Bender, S. L.; Deal, J. G.; Horne, S.;
Liak, T. J.; Yuan, Z. World Pat. Appl. WO 96/16027, 1996;
Chem. Abstr. 1996, 125, 143320, and K. N. Dack, 1994,
unpublished results from our laboratories.
5. Reaction partners for the Heck reaction (Scheme 2): Alkene
18 was reacted with 4-bromo-1,10-biphenyl, 4-bromo-2-¯uoro-
1,10-biphenyl, 2-bromo-¯uorene, 2-chloro-1,10-biphen-4-yl tri-
¯ate,12 4-bromo-2-methyl-1,10-biphenyl,13 4-bromo-2-ethyl-
1,10-biphenyl,13 4-bromo-2-methoxy-1,10-biphenyl13 and 4-
bromo-2-tri¯uoromethyl-1,10-biphenyl.13 Alkene 19 was reac-
ted with 4-bromo-1,10-biphenyl and 4-bromo-2-¯uoro-1,10-
biphenyl. Alkene 20 was reacted with 2-methyl-1,10-biphen-4-
yl tri¯ate12 and 4-bromo-2-tri¯uoromethyl-1,10-biphenyl.
6. Nagase, H.; Fields, C. G.; Fields, G. B. J. Biol. Chem. 1994,
269, 20952.
MMP
IC50 (nM)
ÆSEM (n)
1
2
3
51,000
1790
5.9
Æ27,000 (4)
Æ1000 (5)
Æ2.2 (5)
9
13
14
840
73
1900
Æ91 (5)
Æ1.4 (4)
Æ450 (4)
As expected from the large P10 group, inhibition of
MMP-1 was very weak, whereas it was a relatively
potent inhibitor of MMP-13, consistent with the more
open S10 binding site of the latter. Inhibition potency of
MMP-14 was similar to MMP-2. MMP-14 is similar to
MMP-3 in the S30 site, but is more sterically demanding
at S20 (Phe233 replaces Leu222),11 so it is possible the t-
butyl P20 group plays a part in determining the selectiv-
ity over MMP-14. Inhibition potency of MMP-9 was
about 2-fold greater than MMP-2, and the small di€er-
ence was anticipated given the high homology between
the two gelatinases.
7. Cherney, R. J.; Decicco, C. P.; Nelson, D. J.; Wang, L.;
Meyer, D. T.; Hardman, K. D.; Copeland, R. A.; Arner, E. C.
Bioorg. Med. Chem. Lett. 1997, 7, 1757.
8. Massova, I.; Fridman, R.; Mobashery, S. J. Molec. Model
1997, 3, 17.
Conclusions
We have demonstrated that inhibitory potency against
MMP-2 may be signi®cantly reduced by subtle mod-
i®cations to P10 and P30 groups in a series of succinyl
hydroxamic acid inhibitors of MMP-3. Compound 4j
(UK-356,618) is the most potent and selective MMP-3
inhibitor reported to date, and may be a useful tool for
elucidating the contribution of MMP-3 to pathological
conditions in which selectivity may be required over
other MMPs.
9. Babine, R. E.; Bender, S. L. Chem. Rev. 1997, 97, 1359.
10. The assays for MMPs 2, 3, 9, and 14 are based upon the
original protocol described by Knight et al. (Fed. Euro. Bio-
chem. Soc. 1992, 296, 263) with slight modi®cations. Assays
for MMPs 2, 3, and 9 used the Nagase substrate,6 whereas the
assay for MMP-14 (purchased from Prof. Tschesche, Depart-
ment of Biochemistry, Faculty of Chemistry, University of
Bielefeld, Germany) used the substrate Mca-Pro-Leu-Gly-
Leu-Dpa-Ala-Arg-NH2 (Bachem Ltd, Essex, UK) as descri-
bed by Will et al. (J. Biol. Chem. 1996, 271, 17119). The assay
for MMP-1 inhibition used the substrate Dnp-Pro-b-cyclo-
hexyl-Ala-Gly-Cys(Me)-His-Ala-Lys(N-Me-Ala)-NH2 as ori-
ginally described by Bickett et al. (Anal. Biochem. 1993, 212,
58). The assay for MMP-13 inhibition used the substrate Dnp-
Pro-Cha-Gly-Cys(Me)-His-Ala-Lys(NMA)-NH2. Enzyme and
substrate concentrations were typically 1 nM and 5±10 mM,
respectively.
Acknowledgements
The authors wish to thank Mrs. E. J. Fairman, Mrs.
K. E. Holland, Ms. L. M. Reeves and Mr. S. Lewis for
measuring enzyme inhibition potencies, Mrs. K. S.
Mills, Ms. C. A. Loosley and Messrs. D. Ellis, T. J.
Evans and M. Sproates for making the compounds,
Drs. A. Alex and M. J. de Groot for molecular model-
ing studies and sta€ of the Physical Sciences Dept. for
measuring spectroscopic data.
11. Yamamoto, M.; Tsujishita, H.; Hori, N.; Ohishi, Y.;
Inoue, S.; Ikeda, S.; Okada, Y. J. Med. Chem. 1998, 41, 1209.
12. Prepared as follows: (a) 4-bromo-3-chloroanisole or 4-
bromo-3-methylanisole, PhB(OH)2, 3 mol% Pd(PPh3)4, CsF,
DME, re¯ux; (b) BBr3, CH2Cl2, À78 to 0 ꢁC; (c) Tf2O, pyri-
dine, CH2Cl2, 0±20 ꢁC.
References and Notes
13. Prepared by
a
modi®ed Gomberg±Bachmann±Hey
1. Fray, M. J.; Burslem, M. F.; Dickinson, R. P. Bioorg. Med.
Chem. Lett. 2001, 11, 567.
2. For preparation of (S,S) diastereomer, see Broadhurst, M.
J.; Brown, P. A.; Johnson, W. H.; Lawton, G. Eur. Pat. Appl.
EP 497192, 1992; Chem. Abstr. 1993, 118, 169601.
reaction of the appropriate 2-substituted-4-bromoaniline and
i-amyl nitrite in benzene at re¯ux for 2±3 h (yields 25±40%);
cf. Hassanaly, P.; Vernin, G.; Dou, H. J. M.; Metzger, J. Bull.
Soc. Chim. Fr. 1974, 560, and Eur. Pat. Appl. EP 455058,
1991; Chem. Abstr. 1992, 116, 123167.