Discovery and characterization of the potent, selective and orally bioavailable MMP inhibitor ABT-770

Article abstract of DOI:10.1016/S0960-894X(01)00032-4

Modification of the biphenyl portion of MMP inhibitor 2a gave analogue 2i which is greater than 1000-fold selective against MMP-2 versus MMP-1. The stereospecific synthesis of both enantiomers of 2i was achieved beginning with (S)- or (R)-benzyl glycidyl ether. The (S)-enantiomer, 11 (ABT-770), is orally bioavailable and efficacious in an in vivo model of tumor growth. Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(01)00032-4

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1557±1560
Discovery and Characterization of the Potent, Selective
and Orally Bioavailable MMP Inhibitor ABT-770
Michael L. Curtin,* Alan S. Florjancic, H. Robin Heyman, Michael R. Michaelides,
Robert B. Garland, James H. Holms, Douglas H. Steinman, Joseph F. Dellaria,
Jane Gong, Carol K. Wada, Yan Guo, Ildiko B. Elmore, Paul Tapang,
Daniel H. Albert, Terrance J. Magoc, Patrick A. Marcotte, Jennifer J. Bouska,
Carole L. Goodfellow, Joy L. Bauch, Kennan C. Marsh, 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
AbstractÐModi®cation of the biphenyl portion of MMP inhibitor 2a gave analogue 2i which is greater than 1000-fold selective
against MMP-2 versus MMP-1. The stereospeci®c synthesis of both enantiomers of 2i was achieved beginning with (S)- or (R)-
benzyl glycidyl ether. The (S)-enantiomer, 11 (ABT-770), is orally bioavailable and ecacious in an in vivo model of tumor growth.
# 2001 Elsevier Science Ltd. All rights reserved.
High expression of gelatinase A (MMP-2) has been
observed in solid tumors and is associated with their
ability to grow, invade, develop new blood vessels, and
metastasize.1 Chronic dosing of broad spectrum MMP
inhibitors in clinical studies causes dose-limiting joint
pain and sti€ness and in some cases necessitates an
interruption in therapy.2 Given the prominent role of
MMP-1 in tissue remodeling, an inhibitor of MMP-2
that spares MMP-1 may avoid the side e€ects seen with
the broad spectrum agents.
Because the hydroxamate moiety of 1a/2a was essential
for MMP inhibition and the hydantoin functionality
a€orded high bioavailability (F>85%) and long half-
life after intravenous dosing (t1/2 >3 h) in rodent, the
biphenyl portion was modi®ed to improve the MMP-1/
MMP-2 selectivity. A series of retrohydroxamates with
modi®ed biphenyls and 5,5-dimethylhydantoins was
prepared using the sequence in Scheme 1. In the key
step, an a-bromoketone bearing a di- or trimethylhy-
dantoin (e.g., 3) was alkylated with a variety of phenols.
In the preceding paper,3 a novel and potent family of
MMP inhibitors bearing a retrohydroxamate zinc-bind-
ing group (i.e., 1a and 2a) was introduced. While several
of the disclosed compounds a€orded acceptable phar-
macokinetic pro®les and signi®cant selectivity for
MMP-2 over MMP-1 (40- to 300-fold), an even greater
selectivity ratio was desired. This communication dis-
closes the re®nements of 1a/2a that led to the discovery
of MMP inhibitor 11 (ABT-770) and summarizes this
compound's pre-clinical characterization.
For the example shown, commercially available 2-
(chloromethyl)-3,5-dioxahex-1-ene was treated with
K2CO3 and trimethylhydantoin to give an enol ether
which was deprotected and brominated in one step with
NBS/K2CO3 (63% two-step yield). The resulting a-bro-
moketone 3 was treated with 4-phenylphenol and
Cs2CO3 to a€ord ether 4 in 75% yield, which was then
*Corresponding author. Tel.: +1-847-938-0463; fax: +1-847-938-5034;
e-mail: mike.curtin@abbott.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00032-4

1558
M. L. Curtin et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1557±1560
Following the discovery of the metabolic N-demethyla-
tion of 1a, which was disclosed in the preceding paper,
focus was shifted to the dimethylhydantoin 2. In light of
the SAR work mentioned above, compounds bearing a
40-substituted biphenyl were prepared and examples are
given in Table 2 along with MMP inhibition and intra-
venous half-lives in monkey.
Replacement of the cyano group of 2a gave compounds
with equal potency against MMP-2, but with less or
equal selectivity against MMP-1 (2b and 2c). The
exception to this trend was the alkoxy group which
a€orded potent MMP-2 inhibition along with unex-
pected selectivity. This e€ect was seen with ethoxy (2d)
but was most dramatic for n-propoxy (2e), n-pentoxy
(2f) and methoxyethoxy (2g), all of which had no mea-
surable activity against MMP-1 up to 10 mM. However,
pharmacokinetic analysis of these compounds in mon-
key indicated a uniformly short half-life and low oral
bioavailability.
Speculation that the poor pharmacokinetic pro®le of
the alkoxy analogues was due to facile metabolic clea-
vage of the arylether moiety suggested the preparation
of n-butyl analogue 2h which, despite having a very
attractive in vitro pro®le, also had a short half-life.
Further work along these lines led to tri¯uoromethoxy
analogue 2i, which has excellent potency for MMP-2,
weak inhibition of MMP-1 and an unexpectedly long
half-life and oral activity (see below) in monkey. While
no structural work was done to elucidate the nature of
the binding of these inhibitors with the MMPs, it is
presumed based on previous NMR studies at Abbott6
that the biphenyl moiety interacts with the S10 pocket of
the MMPs and is the origin of the enzyme selectivity.
Scheme 1. (a) 1,5,5-Trimethylhydantoin, K2CO3, DMF; (b) NBS,
K2CO3, acetone; (c) 4-phenylphenol, Cs2CO3, DMF; (d) NH2OH±
HCl, THF/EtOH; (e) BH3±pyridine, 6 M HCl, THF/EtOH; (f)
HCO2Ac, THF.
converted to retrohydroxamate 1b using the three-step
procedure shown (NH2OH; BH3±pyridine, 6 M HCl;
HCO2Ac) in 50% yield. Examples of other tri-
methylhydantoin analogues along with their MMP
inhibition data are shown in Table 1.4,5
While the electronic nature of the biphenyl substituent
was not important for MMP-2 activity (1a±d), alternate
regiochemistry gave dramatic losses in activity (1e and
1f). The use of basic heterocycles (1g) was not allowed
whereas other groups such as furyl (1h) gave equipotent
analogues when compared to 1a. Other acceptable con-
jugated aromatic systems included phenylthienyl (1i)
and styrenyl (1j).
The potency and selectivity of 2i prompted development
of a stereoselective synthesis of both enantiomers and
the preparation of the (S)-enantiomer (11) is shown in
Table 2. Biphenyl modi®cations of 2 and MMP inhibition
Table 1. Biphenyl modi®cations of 1 and MMP inhibition
Compd
X
MMP-1a
MMP-2a
t1/2 (h)b
2a
2b
2c
2d
2e
2f
2g
2h
2i
CN
CF3
SO2Me
OEt
2100Æ590 (4)
840
4900
8Æ3 (4)
8
ndc
5
nd
1
nd
1
1
20
Compd
X
MMP-1a
MMP-2a
20
48
1a
1b
1c
1d
1e
1f
1g
1h
1i
4-(40-CNPh)
4-Ph
2300
1100
1600
440
58
89
20
4
1000
1
4
4-(40-MePh)
4-(40-MeOPh)
4-(30-CNPh)
3-(40-CNPh)
4-(3-Pyridyl)
4-(2-Furyl)
4-(2-Thienyl)
4-Styrenyl
O(n-Pr)
O(n-Pent)
OCH2CH2OCH3
n-Bu
>10,000
>10,000
>10,000
>10,000
8600Æ2700 (5)
26
4
16
6Æ2 (7)
>10,000
>10,000
>10,000
1600
3600
>10,000
1300
>10,000
>10,000
85
OCF3
aIC50 (nM)Æstandard deviation with the number of determinations in
130
61
1j
parentheses.
bAfter a 3 mg/kg iv dose in monkey.
cnd=not determined.
aIC50 (nM).

M. L. Curtin et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1557±1560
1559
Scheme 2. (a) 4-Methoxyphenylboronic acid, PdCl2(dppf), Na2CO3, DME; (b) BBr3, CH2Cl2; (c) (S)-benzyl glycidyl ether, K2CO3, DMF; (d) tert-
butyl N-(tert-butoxycarbonyloxy)carbamate, DEAD, PPh3, THF; (e) H2, Pd/C, THF; (f) 5,5-dimethylhydantoin, DIAD, PPh3, THF; (g) TFA,
CH2Cl2; (h) acetoformic anhydride, THF.
Scheme 2. Suzuki coupling of 1-bromo-4-(tri¯uoro-
methoxy)benzene and 4-methoxyphenylboronic acid
using PdCl2(dppf) and Na2CO3 followed by BBr3
demethylation gave biphenyl alcohol 6 in 89% yield.
This phenol was added to (S)-benzyl glycidyl ether
under basic conditions to give alcohol 7 which was
treated with DEAD/PPh3 and tert-butyl N-(tert-
butoxycarbonyloxy)carbamate7 to a€ord carbamate 8
(81% for two steps). Debenzylation by hydrogenolysis
and addition of 5,5-dimethylhydantoin using DIAD/
PPh3 gave hydantoin 9 in high yield (79%). Finally,
hydroxylamine deprotection (TFA) and N-formylation
yielded retrohydroxamate 11 as a crystalline powder
(50±70% for last two steps, >98% ee by chiral HPLC).
Assessment of both enantiomers (Table 3) not supris-
ingly indicated that both were poorly active against
MMP-1 and MMP-7; more importantly, (S)-enantiomer
11 (ABT-770) was over 40-fold more potent against
MMP-2 than (R)-enantiomer 12 and was chosen for
further characterization.
Single dose pharmacokinetic analysis of 11 after a 3 mg/
kg intravenous dose indicated half-lives of 7.7, 8.7, and
21.0 h in rat, dog, and monkey, respectively, with very
low plasma clearance values (Table 4). Oral adminis-
tration of 11 at 3 mg/kg as a solution gave peak plasma
concentrations of approximately 1.5 mg/mL within 3 h
in the three species tested. Bioavailabilities ranged from
83±95% (Fig. 1).
Table 3. MMP inhibition of enantiomers 11 and 12
Compd
MMP-1a
MMP-2a MMP-3a MMP-7a MMP-9a
11
12
4600Æ830 (3)
4Æ2 (5)
42
3100
>10,000 120Æ36 (3)
>10,000 1200
>10,000
180Æ37 (2)
aIC50 (nM)Æstandard deviation with number of determinations in
parentheses.
Table 4. Pharmacokinetics of 11
Species
t1/2
(h)a
Vc
(L/kg)a
Clp
(l/h kg)
AUC
(mg h/mL)
F
(%)c
a
b
.
.
Ratd
8
9
21
1.2
0.9
1.2
0.12
0.11
0.068
24
23
43
93
83
95
Doge
Monkeye
aAfter a 3 mg/kg iv dose.
bAfter a 3 mg/kg oral dose.
cAbsolute bioavailability.
dn=4.
Figure 1. Plasma concentration of 11 after a 3 mg/kg iv and oral dose
in monkey (n=6).
en=6.

1560
M. L. Curtin et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1557±1560
Given the likelihood that MMP inhibitors will be used
in combination with conventional cytotoxic therapies in
the treatment of solid tumors, a study with compound
11 and paclitaxel, both at sub-optimal doses, was con-
ducted. Figure 3 shows that inhibitor 11 acts in an
additive manner with paclitaxel a€ording a 41% ILS for
the combination versus 16% for paclitaxel alone at that
dose. No overt toxicity such as weight loss was observed
during this study.
Substitution of the biphenyl portion of retro-
hydroxamate 2a gave compound 11, which was shown
to be highly selective for MMP-2 versus MMP-1, long-
lived in vivo with high oral bioavailability in multiple
species. Compound 11 was found to inhibit the growth
of B16 tumors implanted subcutaneously in mice at
doses of 3 mg/kg, bid, or higher and exhibit an additive
e€ect against this tumor cell line when dosed with
paclitaxel. Due to its potency, selectivity, bioavailability
and ecacy, the MMP inhibitor 11 (ABT-770) was
chosen for further preclinical development.
Figure 2. E€ect of 11 on the growth of B16 melanoma cells implanted
subcutaneously in the ¯ank of mice. Compound 11 was administered
orally twice daily on days 7±14 (n=10 per group).
References
1. Stetler-Stevenson, W. G. J. Clin. Invest. 1999, 103, 1237.
2. Steward, W. P. Cancer Chemother. Pharmacol. 1999, 43,
S56.
3. Michaelides, M. R.; Dellaria, J. F.; Gong, J.; Holms, J. H.;
Stacey, J. R.; Wada, C. K.; Heyman, H. R.; Curtin, M. L.;
Guo, Y.; Goodfellow, C. L.; Bouska, J. J.; Elmore, I. B.;
Albert, D. H.; Magoc, T. J.; Marcotte, P. A.; Morgan, D. W.;
Davidsen, S. K. Bioorg. Med. Chem. Lett. 2001, 101, 1553.
4. All compounds were characterized by 1H and 13C NMR,
mass spectroscopy, and elemental analysis.
5. 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.
6. Hajduk, P. J.; Sheppard, G.; Nettesheim, D. G.; Olejni-
czak, 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.
Figure 3. E€ect of 11 and paclitaxel on the growth of B16 melanoma
cells implanted subcutaneously in the ¯ank of mice. Compound 11
(30 mg/kg) was administered orally twice daily on days 7±14. Pacli-
taxel (12.5 mg/kg) was administered ip every 4th day starting on day 7
(n=10 per group).
7. Malamas, M. S.; Palka, C. L. Heterocycl. Chem. 1996, 33,
475. For an example of a stereospeci®c Mitsunobu displace-
ment involving an O,N-bis-protected hydroxylamine, see:
Kolasa, T.; Stewart, A. O.; Brooks, C. D. W. Tetrahedron:
Asymmetry 1996, 7, 729.
8. (a) For the experimental procedure for the tumor ¯ank
model, see: Chirivi, R. G.; Garofalo, A.; Crimmin, M. J.;
Bawden, L. J.; Stoppacciaro, A.; Brown, P. D.; Giavazzi, R.
Int. J. Cancer 1994, 58, 460. (b) Albert, D. H.; Marcotte, P.
A.; Davidsen, S. K. In Current Protocols in Pharmacology;
Enna, S. J., Williams, M., Ferkany, J. W., Kenakin, T., Por-
solt, R. D., Sullivan, J. P., Eds.; John Wiley & Sons: New
York, 2000; in press.
The ecacy of 11 was assessed in several ¯ank models
of tumor growth using B16 murine melanoma cells.8
Oral administration of a solution of 11, starting seven
days after tumor cell implantation, resulted in dose-
dependent inhibition of tumor growth during the 21 day
study (Fig. 2). Statistically signi®cant inhibition was
observed with doses at or above 10 mg/kg, bid, with the
maximal e€ective dose between 100 and 300 mg/kg, bid.
The 100 mg/kg, bid, dose produced a 6.5 day delay to
1 g tumor weight which corresponded with a 46%
increase in life span (ILS).