Biaryl ether retrohydroxamates as potent, long-lived, orally bioavailable MMP inhibitors

Article abstract of DOI:10.1016/S0960-894X(01)00031-2

A novel series of biaryl ether reverse hydroxamate MMP inhibitors has been developed. These compounds are potent MMP-2 inhibitors with limited activity against MMP-1. Select members of this series exhibit excellent pharmacokinetic properties with long elimination half-lives (7 h) and high oral bioavailability (100%).

Full text of DOI:10.1016/S0960-894X(01)00031-2

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1553±1556
Biaryl Ether Retrohydroxamates as Potent, Long-Lived, Orally
Bioavailable MMP Inhibitors
Michael R. Michaelides,* Joseph F. Dellaria, Jane Gong, James H. Holms,
Jennifer J. Bouska, Jamie Stacey, Carol K. Wada, H. Robin Heyman, Michael L. Curtin,
Yan Guo, Carole L. Goodfellow, Ildiko B. Elmore, Daniel H. Albert, Terrance J. Magoc,
Patrick A. Marcotte, Douglas W. Morgan and Steven K. Davidsen
Cancer Research Area, Abbott Laboratories, Dept. 47J, Bldg. AP10, 100 Abbott Park Road, Abbott Park, IL 60064, USA
Received 18 October 2000; accepted 5 January 2001
AbstractÐA novel series of biaryl ether reverse hydroxamate MMP inhibitors has been developed. These compounds are potent
MMP-2 inhibitors with limited activity against MMP-1. Select members of this series exhibit excellent pharmacokinetic properties
with long elimination half-lives (7 h) and high oral bioavailability (100%). # 2001 Elsevier Science Ltd. All rights reserved.
The matrix metalloproteinases (MMPs) are a family of
zinc endopeptidases that collectively have the ability to
degrade all the components of the extracellular matrix.
Under physiological conditions MMPs play an impor-
tant role in normal tissue turnover and development.
However, increased MMP activity has been associated
with a variety of pathological conditions such as arthri-
tis, cancer, multiple sclerosis, and arteriosclerosis.1
Therefore, inhibition of MMP activity represents an
attractive therapeutic target and has been the subject of
intense research in both industrial and academic
laboratories, resulting in several structurally diverse
compounds advancing to clinical trials.2
We have previously disclosed a series of biaryl ether
hydroxamates (1) as potent stromelysin and gelatinase
inhibitors, discovered through the use of SARs by
NMR.5 Compounds such as 1 are selective MMP inhi-
bitors since they possess limited activity against MMP-1
and MMP-7. These two enzymes di€er from other
MMP family members in that they have a relatively
small S10 pocket, which presumably cannot accom-
modate the large biaryl substituent. However, poor
pharmacokinetic properties due to rapid hydroxamate
hydrolysis precluded further development of this series.
We now wish to report that the replacement of the
hydroxamate of 1 with an N-formylhydroxylamine
(retrohydroxamate) 2a leads to potent, long-lived and
orally bioavailable MMP inhibitors.
Our program has focused on the discovery of selective
MMP inhibitors as antitumor agents. Although the role
of each of the MMPs in the various pathologies is not
known with certainty, an increasing body of evidence sug-
gests a critical role for the gelatinases in tumor progression
and metastasis.3 Furthermore, selective agents may avoid
the musculoskeletal side e€ects observed in patients treated
chronically with broad spectrum inhibitors.4
The retrohydroxamate 2a was prepared as shown in
Scheme 1. Reaction of 1,2-dibromoethane with the
biaryl phenol 3 gave the bromide 4 (25%), which was
reacted with N,O-bis-Boc protected hydroxylamine to
give 5 in 85% yield. Acid mediated deprotection (75%)
*Corresponding author. Tel.:+1-847-937-9874; fax; +1-847-935-
5165; e-mail: michael.michaelides@abbott.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00031-2

1554
M. R. Michaelides et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1553±1556
followed by formylation with formic acetic anhydride
gave 2a in 25% recrystallized yield. Alternatively
acetylation or reaction with TMSNCO gave the N-acetyl
hydroxylamine 2b and the N-hydroxyurea 2c, respectively.
Thus, alkylation of hydantoin 15 followed by mCPBA
epoxidation, epoxide opening with iodine and Dess±
Martin oxidation gave the key iodo ketone intermediate
16 (45% overall). Alkylation of 16 with phenol 3, and
further elaboration as in the previous schemes provided
17.
The a-alkyl and aryl substituted compounds were pre-
pared according to Scheme 2. The key step was the
addition of an alkyl or aryl Grignard to the Weinreb
amide intermediate 7. Alkylation of ethyl bromoacetate
with phenol 3, followed by ester hydrolysis and amide
coupling a€orded the intermediate 7 in 60% overall
yield. Reaction of 7 with Grignard reagents gave the
ketone 8, which was then converted to the retro-
hydroxamate through a three-step procedure involving
oxime formation, reduction and formylation (40±50%).
The biaryl ether hydroxamates, although potent MMP
inhibitors in vitro, had very poor pharmacokinetic
properties with elimination half-lives of less than 30 min
due to rapid hydrolysis to the corresponding carboxylic
acid. The acids are generally two orders of magnitude
weaker inhibitors (cf. 2a vs 2e). This led to an investi-
gation of alternate zinc binding groups that would
The a-heteroatom methylene substituted compounds
were synthesized as in Scheme 3, through the alkylation
of the key iodoketone intermediate 12. Thus, reaction of
phenol 3 with epibromohydrin gave the epoxide 11.
Epoxide opening of 11 with iodine and triphenyl phos-
phine (91%) followed by Dess±Martin alcohol oxida-
tion (75%) gave 12. Iodide displacement with a variety
of heteroatom nucleophiles gave the ketone 13, which
was converted to 14 following the same three-step
sequence of reactions described in Scheme 2.
The a-ethyl linked hydantoin compounds were prepared
following the same general procedures used for the pre-
paration of 14, except reversing the order of introduction
of the a-substituent and the biaryl phenol (Scheme 4).
Scheme 3. (a) 3, K2CO3, THF; (b) I2, Ph3P, CH2Cl2; (c) Dess±Martin,
CH2Cl2; (d) HX-R, K2CO3, DMF; (e) same as e, f, and g in Scheme 2.
Scheme 4. (a) 3-Buten-1-ol, Ph3P, DEAD; (b) mCPBA; (c) I2, Ph3P;
(d) Dess±Martin; (e) 2, K2CO3; (f) same as e, f, and g in Scheme 2.
Scheme 1. (a) Br(CH2)2Br, K2CO3, DMF; (b) BocNHOBoc, NaH,
DMF; (c) 4 N HCl, dioxane; (d) HCOOAc, THF (for 2a); Ac2O (for
2b): TMSNCO, Et3N, THF (for 2c).
Table 1. Alternate chelators: MMP enzyme inhibition potency7
Compd
ZBG
IC50 (mM)
MMP-2
MMP-3
MMP-1
1a
2a
2b
2c
2d
2e
CONHOH
N(OH)CHO
N(OH)COMe
N(OH)CONH2
NHCONHOH
COOH
0.032
0.4
5.9
35
>100
8.0
0.057
0.62
4.0
>100
80
>10
3.3
>50
nda
>50
>50
nda
Scheme 2. (a) 3, K2CO3, THF; (b) LiOH, dioxane, H2O; (c)
HNMe(OMe), Et3N, BOPCl; (d) RMgBr, THF, À78 ꢀC; (e)
.
.
NH2OH HCl, pyridine, THF; (f) BH3 Pyr., 6 N HCl, EtOH; (g)
HCOOAc, THF.
and=Not determined.

M. R. Michaelides et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1553±1556
1555
potentially be more stable in vivo (Table 1). Replace-
ment of the hydroxamate with either an N-term-
inal (2d) or internal N-hydroxyurea (2c) led to a
dramatic loss of activity. Reversing the regiochemistry
of the hydroxamate moiety led to the N-formyl and
N-acetyl hydroxylamines (retrohydroxamates) 2a and
2b, respectively.6 The N-formyl hydroxylamine 2a was
only 10-fold weaker than the `normal' hydroxamate 1a
whereas the N-acetyl derivative 2b was signi®cantly
weaker. Consequently, attention was focused on the N-
formyl analogues. The a-methyl derivative 9a was more
potent than 2a and only 3-fold less potent than the cor-
responding hydroxamate 1b (Table 2). More impor-
tantly 9a had signi®cantly better pharmacokinetic
properties than 1b when dosed intravenously and orally
in rat. This compound exhibited a half-life of 1.6 h and
oral bioavailability of 33% as opposed to 0.3 h and 7%,
respectively, for 1b. Thus, a systematic investigation of
the e€ect of a-substitution on potency and pharmaco-
kinetics was undertaken. Introduction of larger alkyl
and aryl groups (9b±9d) or aryl ether and thioether (14a
and 14b) had a deleterious e€ect on MMP-2 potency.
Heteroalkyl groups such as succinimide, phthalimide
and pyridazinone gave a boost in MMP-2 inhibitory
potency, however they all exhibited poor pharmacoki-
netic properties as evidenced by the short (<1 h) half-
lives after intravenous administration in cynomolgous
monkeys. Interestingly, introduction of a glutarimide
led to a 10-fold increase in MMP-1 potency.
Table 2. MMP enzyme inhibition potency and elimination half-lives:
e€ect of a-substitution
Compd
R
IC50 (nM)
Half-life, iv
MMP-2 MMP-3 MMP-1
1b
9a
9b
9c
9d
14a
14b
Ð
Me
i-Bu
pF-Ph
19
76
45
620
120
300
180
130
36
2500
>10,000
>10,000
>10,000
14,000
0.3 ha
1.6 ha
nd
nd
nd
950
480
650
260
160
CH2-(p-F-Ph)
CH2OPh
CH2SPh
>10,000
4100
nd
0.3 hb
14c
14d
14e
5.2
11
21
37
37
>10,000
>10,000
4200
0.6 hb
0.3 hb
0.5 hb
nd
One of the heterocycle substituents that has been exten-
sively investigated in several series of MMP inhibitors is
the hydantoin moiety. In particular two MMP inhibi-
tors that advanced into clinical studies, RO 32-3555 and
D-2163, contain a 4,4-dimethyl-N-methyl hydantoin.8
The analogous 4,4-dimethyl-N-methyl hydantoin sub-
stituted retrohydroxamate 14g was prepared and found
to exhibit approximately the same potency against
MMP-2 (IC50=58 nM) as the a-methyl compound 9a
(IC50=76 nM), although signi®cantly more potent
against MMP-3 (IC50=9.1 nM vs 620 nM). The N-ethyl
analogue 14h showed improved potency against MMP-
2 (IC50=12 nM) and over 100-fold selectivity over
MMP-1 (Table 3). Pharmacokinetic evaluation of these
two compounds in cynomolgous monkey showed that
37
3.4
14f
25
360
aRat.
bCynomolgous monkey.
Table 3. Hydantoin substituted retrohydroxamates: MMP enzyme inhibition potency and pharmacokinetics
Compd
n
R
R0
IC50 (nM)
MMP-3
t1/2 iva
(h)
AUC, iva
mmol h/L
AUCpo (dose) b
mmol h/L (mg/kg)
F (%)
.
.
MMP-2
MMP-1
14g
14h
14i
14j
17a
17b
1
1
1
1
2
2
Me
Me
Me
H
Me
Me
Me
Et
H
H
Me
H
58
12
7.8
8.2
45
9.1
27
24
14
17
35
2300
2400
2100
2400
2.5
1.8
7.7
7.3
0.5
12.7
15.2
12.3
24
59.8
8
11.8 (10)
ndc
27.2 (3)
68 (3)
1.8 (10)
2.4 (3)
23%
ndc
110%
100%
6%
>10,000
14,000
6.9
47.5
5%
aDosed at 3 mg/kg, cynomolgous monkey.
bDosed as a solution in 10% ethanol:0.2% hydroxypropyl-methylcellulose, n=2.
cnd=Not determined.

1556
M. R. Michaelides et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1553±1556
MMP-1. However, despite the close structural similarity
to 14i these compounds su€er from low oral
biovailability (F=5±6%).
In summary, a series of biaryl ether retrohydroxamates
has been identi®ed that are potent MMP-2 inhibitors.
Select members of this series possess long elimination
half-lives and show excellent oral bioavailability. Fur-
ther work to optimize the in vitro pro®le of these com-
pounds and in vivo ecacy is presented in the following
paper.10
Figure 1. Blood levels after iv administration of 14h.
References
1. White, A. D.; Bocan, T. M. A.; Boxer, P. A.; Peterson, J. T.;
Schrier, D. Curr. Pharm. Des. 1997, 3, 45.
2. Michaelides, M. R.; Curtin, M. L. Curr. Pharm. Des. 1999,
5, 787.
3. Summers, J. B.; Davidsen, S. K. Annu. Rep. Med. Chem.
1998, 33, 131.
4. Wojtowich-Praga, S.; Torri, J.; Johnson, M.; Steen, V.;
Marshall, J.; Ness, E.; Dickson, R.; Sale, M.; Rasmussen,
H. S.; Avanzado-Chiode, T.; Hawkins, M. J. J. Clin. Oncol.
1998, 16, 2150.
5. Hajduk, P. J.; Sheppard, G.; Nettesheim, D. G.; Olejnic-
zak, E. T.; Shuker, S. B.; Meadows, R. P.; Steinman, D. H.;
Carrera, G. M.; Marcotte, P. A., Jr.; Severin, J.; Walter, K.;
Smith, E.; Gubbins, H. E.; Simmer, R.; Holzman, T. F.;
Morgan, D. W.; Davidsen, S. K.; Summers, J. B.; Fesik, S. W.
J. Am. Chem. Soc. 1997, 119, 5818.
6. Robl, J. A.; Simpkins, L. M.; Asaad, M. Bioorg. Med.
Chem. Lett. 2000, 10, 257.
7. For the experimental procedure of the MMP inhibition
assay, see: Curtin, M. L.; Garland, R. B.; Davidsen, S. K.;
Marcotte, P. A.; Albert, D. H.; Magoc, T. J.; Hutchins, C.
Bioorg. Med. Chem. Lett. 1998, 8, 1443.
these compounds have half-lives of 1.8 and 2.5 h,
respectively. Compound 14g was also administered
orally and found to possess 23% oral bioavailability.
However, HPLCanalysis of blood samples revealed the
presence of a long-lived metabolite that is common to
both 14g and 14h. Analysis of the same samples via a
stromelysin bioassay indicated that the bioactivity cor-
related with the concentrations of metabolite as mea-
sured by HPLCat the time points when no detectable
parent was present (Fig. 1).
HPLC±mass spectrometry analysis indicated a mass of
422 for the metabolite corresponding to the loss of
methyl and ethyl for 14g and 14h, respectively. Thus,
the putative bioactive metabolite 14i was prepared and
shown to co-elute with the metabolite seen in vivo.
Gratifyingly this compound is a potent MMP-2 inhi-
bitor (IC50=7.8 nM) with only weak potency against
MMP-1. As expected this compound is long-lived with
a half-life after intravenous administration of 7.7 and
6 h in cynomolgous monkey and marmoset, respec-
tively. Remarkably 14i also possesses excellent oral
bioavailabilty in both cynomolgous monkey and mar-
moset (F=110 and 85%, respectively).9 The unsub-
stituted hydantoin analogue 14j has a similar in vitro
pro®le and is also long-lived and orally bioavailable
indicating that the gem-dimethyl moiety does not e€ect
the pharmacokinetic pro®le. The homologated hydan-
toin analogues 17a and 17b have similar MMP-2
potency, but are signi®cantly more selective against
8. Skiles, J. W.; Monovich, L. G.; Jeng, A. Y. Annu. Rep.
Med. Chem. 2000, 35, 167.
9. The greater than 100% bioavailability is likely due to
enterohepatic recirculation and/or interanimal variability.
10. Curtin, M. L.; Florjancic, A. S.; Heyman, H. R.; Michae-
lides, M. R.; Garland, R. B.; Holms, J. H.; Steinman, D. H.;
Dellaria, J. F.; Gong, J.; Wada, C. K.; Guo, Y.; Elmore, I. B.;
Tapang, T.; Albert, D. H.; Magoc, T. J.; Patrick, A.; Mar-
cotte, P. A.; Bouska, J. J.; Goodfellow, C. L.; Bauch, J. L.;
Marsh, K. C.; Morgan, D. W.; Davidsen, S. K. Bioorg. Med.
Chem. Lett. 2001, 11, 1557.