α,β-Cyclic-β-benzamido hydroxamic acids: Novel oxaspiro[4.4]nonane templates for the discovery of potent, selective, orally bioavailable inhibitors of tumor necrosis factor-α converting enzyme (TACE)

Article abstract of DOI:10.1016/j.bmcl.2008.01.030

Two novel oxaspiro[4.4]nonane β-benzamido hydroxamic scaffolds have been synthesized in enantio- and diasteriomerically pure form. These templates proved to be exceptional platforms that have led to the discovery of potent inhibitors of TACE that are acti

Full text of DOI:10.1016/j.bmcl.2008.01.030

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1288–1292
a,b-Cyclic-b-benzamido hydroxamic acids: Novel
oxaspiro[4.4]nonane templates for the discovery of potent,
selective, orally bioavailable inhibitors of tumor necrosis
factor-a converting enzyme (TACE)
Gregory R. Ott,* Naoyuki Asakawa, Rui-Qin Liu, Maryanne B. Covington, Mingxin Qian,
Krishna Vaddi, Robert C. Newton, James M. Trzaskos, David D. Christ, Laurine Galya,
Thomas Scholz, Will Marshall and James J.-W. Duan
Departments of Discovery Chemistry and Discovery Biology, Bristol-Myers Squibb Research and Development,
Rte 206 and Province Line Road, Princeton, NJ 08543, USA
Received 11 December 2007; revised 4 January 2008; accepted 8 January 2008
Available online 11 January 2008
Abstract—Two novel oxaspiro[4.4]nonane b-benzamido hydroxamic scaffolds have been synthesized in enantio- and diasteriomer-
ically pure form. These templates proved to be exceptional platforms that have led to the discovery of potent inhibitors of TACE
that are active in a cellular assay measuring suppression of LPS-induced TNF-a. Furthermore, these inhibitors are selective against
related MMPs, demonstrate permeability in a Caco-2 assay, and display good oral bioavailability.
ꢀ 2008 Elsevier Ltd. All rights reserved.
Tumor Necrosis Factor-a (TNF-a), a pro-inflammatory
cytokine, has been the subject of intense interest for its
purported role in numerous pathological diseases.1
The modulation of TNF-a as a drug discovery target
has been validated by the clinical success of the anti-
TNF-a biologics entaneracept, infliximab, and ada-
limumab.2 From the point of view of small molecule
drug discovery, an alternative target to mitigate TNF-
a production is the metalloprotease TNF-a converting
enzyme (TACE).3
is shown in Fig. 1. In accordance with homology model-
ing, the active conformation presents the ring into sol-
vent.6 Interestingly, data generated from the initial
carbocyclic leads indicated that for oral bioavailability,
a certain degree of lipophilicity was required in the cyclic
b-amino acid portion of the inhibitor, but was deleteri-
ous to obtaining the necessary cellular potency. It was
demonstrated that certain cyclic b-amino acid deriva-
tives including tetrahydropyran, tetrahydrofuran, and
substituted pyrrolidine b-amino hydroxamic acids were
preferred.4,5
The search for non-peptidic inhibitors of TACE has led
to the discovery of novel a,b-cyclic-b-benzamido
hydroxamic acids.4,5 These studies demonstrated that
the active site tolerates a wide variety of rings including
5- and 6-membered carbocyclic and heterocyclic rings
including cyclopentane, pyrrolidine, tetrahydrofuran,
cyclohexane, tetrahydropyran, and piperidine systems.
A representation of the binding motif of these inhibitors
During the examination of substituent effects on the 4-
position of the (1R,2S)-cyclopentane-b-benzamido
Keywords: TACE; TNF-a converting enzyme; MMP; Matrix metallo-
protease; b-Amino acid.
*
Corresponding author at present address: Cephalon, Inc. 145
Brandywine Parkway, West Chester, PA 19380, USA. Tel.: +1 610
738 6861; fax: +1 610 738 6558; e-mail: gott@cephalon.com
Figure 1. TACE inhibitors, cis-a,b-cyclic-b-amino acid scaffolds.
0960-894X/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.01.030

G. R. Ott et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1288–1292
1289
Figure 2. Design of spiroether-b-benzamido hydroxamic acid template.
hydroxamic acid scaffold,5a it was shown that the 1,3-
dioxolane moiety (1, Fig. 2) is not only tolerated in
the active site (pTACE IC50 = 1.0 nM), it imparts a dra-
matic boost to potency in the cellular suppression of
LPS-induced TNF-a in human whole blood (WBA,
IC50 = 24 nM), as compared to the unsubstituted parent
cyclopentane (WBA IC50 = 475 nM).4,7 Importantly,
in vitro screening parameters predictive of oral absorp-
tion (Caco-2) illustrated that 1 had promising perme-
ability (Papp = 1.8 · 10ꢀ6 cm/s). Unfortunately, the acid
labile nature of the dioxolane precluded advancing a
compound with that moiety. Furthermore, the resultant
hydrolysis product (i.e. the ketone), though potent,
showed poor Caco-2 permeability.5a
Since 1 appeared to have a desirable in vitro profile
(WBA IC50 = 24 nM, Caco-2 Papp = 1.8 · 10ꢀ6 cm/s),
we felt that the structural motif represented by the
Scheme 1. Reagents and conditions: (a) Allyltrimethylsilane, TiCl4, CH2Cl2 (57%); (b) BH3 Æ THF; NaOH, H2O2; (c) MsCl, TEA, CH2Cl2, 0 ꢁC to
reflux; (d) H2, Pd(OH)2/C, MeOH; (e) i-PrO2CCl, TEA; NaN3, H2O; benzene, reflux; BnOH, p-TsOH, reflux; (f) H2, Pd(OH)2/C, MeOH; (g) BOP
reagent, DIPEA, DMF; (h) NH2OH, NaOMe, MeOH.

1290
G. R. Ott et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1288–1292
dioxolane system may be transformed into a viable
alternative which would impart both the necessary
polarity to achieve cellular potency and the desired lipo-
philicity to impart bioavailability. To exploit this unique
architecture we proposed exchange of one of the oxy-
gens of the dioxolane species with a carbon, thus reveal-
ing a novel spiroether scaffold (Fig. 2). It was expected
that this oxaspiro[4.4]nonane system would retain the
in vitro properties of the dioxolane system, while
remaining stable to hydrolytic conditions.
Figure 3. Observed nOe’s in 18 and 21.
tion, the phenol derivative 21 displayed similar nOe’s
from H1 to H5a and H5a to H8 using a GOESY
experiment.11
Installation of the b-amino-acid portion of the inhibitor
would be achieved using a Curtius rearrangement of an
appropriately substituted dicarboxylate (Scheme 1).8
The spiroether portion of the b-amino acid template
was envisioned to arise from addition of an allyl
nucleophile to a ketone to give a quaternary carbinol
followed by hydroboration of the olefin and intramolec-
ular ether formation. Thus, Lewis-acid catalyzed addi-
tion of allyltrimethylsilane to ketone 25a,9 provided the
desired allyl addition products in >10:1 ratio of (4R)
to (4S). Compounds 3 and 4 were separated using silica
gel chromatography and taken on independently; the
structure was assigned by correlation with later products
(vide infra). The olefin of the allyl group was hydrobo-
rated which provided the desired hydroxyl derivatives
in modest isolated yield. The primary alcohols were
mesylated, then heated at reflux to effect cyclization.
The oxaspiro[4.4]nonane systems10 7 and 8 were isolated
in good yield. Debenzylation followed by Curtius rear-
rangement gave the Cbz-protected amines 11 and 12.
The free amines 13 and 14 were delivered by hydrogena-
tion. Coupling to carboxylic acid 154 afforded the com-
plete carbon skeleton of the inhibitors. Treating the
esters 16 and 17 with a solution of hydroxylamine
hydrochloride and sodium methoxide in methanol that
had been premixed and filtered provided the inhibitors
18 and 19 in 57% and 60% isolated yields, respectively.
The stereochemistry determined by NMR analysis was
confirmed by single crystal X-ray of the phenol deriva-
tive 21.12 A 3-D representation of the crystal structure
of 21 is shown in Fig. 4.
The independently synthesized spiroether b-amino
hydroxamates 18 and 19 were tested in vitro using
semi-purified porcine TACE (pTACE)13 since there is
high sequence homology between the human and por-
cine forms of TACE. Selectivity against matrix metallo-
NMR analysis of both 18 and the phenol derivative 21
revealed the relative stereochemistry of the spiroether
system. 21 was prepared by reductive cleavage of the
quinolinyl moiety from 16 and conversion of the derived
ester 20 to the hydroxamic acid 21 (Scheme 2).
The results from the NMR experiments are shown in
Fig. 3. For 18, nOe’s were observed for H1 to H5a and
H5a to H8. Furthermore, using a 2D-ROESY experi-
ment, nOe was observed from H1 to H8, thus indicating
that the spiroether has the (R)-stereochemistry. In addi-
Figure 4. 3-D representation of 21 from X-ray coordinates.
Scheme 2. Reagents: (a) Zn, AcOH, CH2Cl2 (83%); (b) NH2OH, NaOMe, MeOH (55%).

G. R. Ott et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1288–1292
Table 2. Pharmacokinetic parameters of 18 and 19a,b
1291
proteases (MMPs) was determined using MMP-1 as a
representative with a shallow S1’ pocket and MMP-2
and MMP-9 with deep S1’ pockets. Cellular inhibition
of LPS-stimulated TNF-a was measured in human
whole blood (WBA). We desired potency in a cellular as-
say since there is evidence that suggests proteolytic
activity of TACE occurs both in an intra- as well as
extracellular fashion.14 In vivo pharmacokinetics were
determined following iv and oral dosing using LC/MS/
MS.15 In vivo efficacy was determined in an acute model
of endotoxemia in mouse.13
PK parameters
18
19
IV
dose (mg/kg)
t1/2 (h)
5.0
4.2
5.0
6.9
0.8
1.4
Cl (L/h/kg)
Vss (L/kg)
AUC (nM•h)
0.7
0.3
12,628
10,947
PO
dose (mg/kg)
tmax (h)
t1/2 (h)
5.0
0.1
1.3
3930
31
5.0
0.25
1.9
AUC (nM•h)
F%
3848
35
Shown in Table 1 are the in vitro results for 18 and 19.
Both isomers were extremely potent against pTACE
(1.0 nM, respectively). Importantly, 18 and 19 displayed
potency in the cellular suppression of TNF-a. As com-
pared to the dioxolane 1, several-fold loss in activity
was observed and potency still remained in the desired
range (WBA IC50 < 200 nM). Following the previously
established paradigm, the 4-[(2-methylquinolin-4-
yl)methoxy]benzamide P1’ moiety imparted the desired
selectivity;16 18 and 19 displayed >1000-fold selectivity
for pTACE over MMP ꢀ 1, ꢀ2, and ꢀ 9. We were grat-
ified to find that both examples gave Caco-2 values
indicative of good oral absorption (i.e. Papp > 1.0 ·
10ꢀ6 cm/s) with 18 demonstrating Papp = 5.7 · 10ꢀ6 cm/
s and 19 having Papp = 1.5 · 10ꢀ6 cm/s. Consistent with
the Caco-2 values, both 18 and 19 proved to be orally
bioavailable in Sprague–Dawley rats.
a Determination of 3 for each dosing group, avg. value.
b Data from n-in-1 cassette dosing in Sprague–Dawley rats.
with tmax = 0.1 and 0.25 h for 18 and 19, respectively.
The oral half-lives were also quite similar, for 18
(1.3 h) and 19 (1.9 h). The oral exposure as measured
by AUC was 3930 nMh for 18 and 3848 nMh for 19.
Oral bioavailability proved to be 31% for 18 and 35%
for 19. The overall oral pharmacokinetic profiles for
the two diastereomeric inhibitors were remarkably alike.
Furthermore, 18 displayed oral bioavailability in beagle
dogs (F = 68%).
Since the overall PK profiles were similar and the
in vitro potencies were also comparable, 18, available
in larger quantities since it was derived from major dia-
stereomer of the allyl addition, was analyzed for efficacy
in a murine model of endotoxemia.13 Lipopolysaccha-
ride was administered along with different doses of 18
by oral gavage administration at t = 0 h. One hour post
dose, the mice were euthanized, whole blood collected
by cardiac puncture, and the plasma used for the mea-
surement of TNF-a concentration. 18 displayed a dose
dependent suppression of TNF-a with a calculated
ED50 of between 1 and 3 mg/kg.
Shown in Table 2 are the pharmacokinetic profiles of 18
and 19 after iv (5 mg/kg) and oral dosing (5 mg/kg) in
Sprague–Dawley rats. Data obtained for 18 and 19 were
obtained from n-in-1 cassette dosing studies. Both 18
and 19 displayed low clearance following iv dosing
(0.7 and 0.8 L/h/kg, respectively). 18 had a much smaller
volume of distribution (0.3 L/kg) as compared to 19
(1.4 L/kg). The iv half-life for 18 was 4.2 h and for 19
was 6.9 h. The overall iv exposure for each compound
was quite high as measured by the total AUC,
12,628 nMh for 18 and 10,947 nMh for 19. After oral
administration, both compounds were rapidly absorbed
In summary, two novel oxaspiro[4.4]nonane b-benz-
amido hydroxamic scaffolds have been synthesized. These
novel b-amino acid derivatives proved to be exceptional
Table 1. In vitro properties of inhibitors 1, 18 and 19
Compound pTACEa (IC50, nM) WBAb (IC50, nM) MMP-1 Ki (nM) MMP-2 Ki (nM) MMP-9 Ki (nM) Caco-2c F% rat F% dog
d
1
18
19
1.0
1.0
1.0
24
108
143
>4948
>4948
>4948
2800
1357
1920
>2128
1447
1246
1.8
5.7
1.5
—
—
31%
35%
68%
—
a pTACE IC50 and MMP Ki values are from single determination.
b Inhibition of TNF-a release in WBA was determined with three donors.
c Papp (A ! B) · 10ꢀ6 cm/s.
d
—, Not determined.

1292
G. R. Ott et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1288–1292
Decicco, C. P.; Duan, J. J.-W. Bioorg. Med. Chem. Lett.
2007, 17, 1865.
6. Wasserman, Z. R.; Duan, J. J.-W.; Voss, M. E.; Xue,
C.-B.; Cherney, R. J.; Nelson, D. J.; Hardman, K. D.;
Decicco, C. P. Chem. Biol. 2003, 10, 215.
7. Consistent with the observed paradigm that increased
polarity drives cell potency, the dithiolane congener of 1,
though displaying equipotent activity in the enzymatic
assay and similar MMP selectivity, does not impart a
similar boost to cellular potency (WBA IC50 = 263 nM).
8. For examples of the preparation of cyclic b-amino acids
using a Curtius rearrangement, see: Bolm, C.; Schiffers, I.;
Atodiresei, I.; Hackenberger, C. P. R. Tetrahedron:
Asymmetry 2003, 14, 3455.
platforms that, in conjunction with an optimized P1’
group, have led to the discovery of potent inhibitors of
pTACE. Inhibitors 18 and 19 are active in a cellular assay
measuring suppression of LPS-induced TNF-a. These
inhibitors are selective against MMPs and displayed good
oral bioavailability in rats. Furthermore, 18 showed
excellent oral bioavailability in beagle dogs. In a rodent
model of endotoxemia, 18 displayed an oral ED50 of 1–
3 mg/kg in the suppression of TNF-a.
Acknowledgment
9. All new compounds were consistent with 1H NMR and
low resolution mass spectrometry data.
10. The numbering system of the oxaspiro[4.4]nonane system
is reflected in the structure.
11. The GOESY is a double-PFG-spin-echo, transient nOe
experiment using shape selective pulses and pulse field
gradients in which the results appear similar to a steady-
state nOe difference experiment. Since excitation is done
with shaped pulses, a much narrower region of the
spectrum is affected so it is possible to see weaker nOes
than with a steady-state nOe difference experiment, see
Stott, K.; Keeler, J.; Van, Q. N.; Shaka, A. J. J. Magn.
Reson. 1997, 125, 302.
We thank John Giannaras, Sherrill Nurnberg, and Paul
Strzemienski for running in vitro assays, Robert J. Col-
lins for LPS study in animals, Bernice Brogdon, Bing
Lu, Hang Zheng, and Jingtao Wu for assistance in phar-
macokinetic studies, Maria Ribadeneira for Caco-2
assay, and Gerry Everlof for running stability screens.
We also thank Mary Malley for help in the preparation
of Fig. 4.
References and notes
12. The coordinates for 21 have been placed in the Cambridge
Crystallographic Data Centre, deposition number: CCDC
670062.
1. Newton, R. C.; Decicco, C. P. J. Med. Chem. 1999, 42,
2295.
2. (a) Moreland, L. W.; Baumgartner, S. W.; Schiff, M. H.;
Tindall, E. A.; Fleischmann, R. M.; Weaver, A. L.;
Ettlinger, R. E.; Cohen, S.; Koopman, W. J.; Mohler, K.;
Widmer, M. B.; Blosch, C. M. N. Engl. J. Med. 1997, 337,
141; (b) Lipsky, P. E.; van der Heijde, D. M. F. M. E.; St.
Clair, W.; Furst, D. E.; Breedveld, F. C.; Kalden, J. R.;
Smolen, J. S.; Weisman, M.; Emery, P.; Feldmann, M.;
Harriman, G. R.; Maini, R. N. N. Engl. J. Med. 2000, 343,
1594; (c) Machold, K. P.; Smolen, J. S. Exp. Opin. Bio.
Ther. 2003, 3, 351.
13. Details of the pTACE (porcine TACE), MMPs, human
whole blood assay (WBA) as well as LPS-mouse efficacy
model are reported in previous publication, see: Xue,
C.-B.; Voss, M. E.; Nelson, D. J.; Duan, J. J.-W.;
Cherney, R. J.; Jacobson, I. C.; He, X.; Roderick, J.;
Chen, L.; Corbett, R. L.; Wang, L.; Meyer, D. T.;
Kennedy, K.; DeGrado, W. F.; Hardman, K. D.;
Teleha, C. A.; Jaffee, B. D.; Liu, R.-Q.; Copeland, R.
A.; Covington, M. B.; Christ, D. D.; Trzaskos, J. M.;
Newton, R. C.; Magolda, R. L.; Wexler, R. R.; Decicco,
C. P. J. Med. Chem 2001, 44, 2636.
3. For a recent review, see: Skotnicki, J. S.; Levin, J. I. Ann.
Rep. Med. Chem. 2003, 38, 153.
14. (a) Schlondorff, J.; Becherer, J. D.; Blobel, C. P. Biochem.
J. 2000, 347, 131; (b) Solomon, K. A.; Covington, M. B.;
Decicco, C. P.; Newton, R. C. J. Immunol. 1997, 159,
4524.
15. Wu, J.-T.; Zeng, H.; Qian, M.; Brogdon, B. L.; Unger, S.
E. Anal. Chem. 2000, 72, 61.
16. For related analogs in the a,b-cyclic-b-benzamido
hydroxamic acid series, replacement of the 4-[(2-methyl-
quinolin-4-yl)methoxy]benzamide P1’ substituent with 4-
[(2,6-dimethylpyridin-4-yl)methoxy]benzamide resulted in
maintenance of pTACE activity but loss in both WBA
activity and selectivity against MMPs-2 and -13. Ott, G.
R.; Asakawa, N, unpublished results.
4. Duan, J. J.-W.; Chen, L.; Lu, Z.; Xue, C.-B.; Liu,
R.-Q.; Covington, M. B.; Qian, M.; Wasserman, Z.
R.; Vaddi, K.; Christ, D. D.; Trzaskos, J. M.;
Newton, R. C.; Decicco, C. P. Bioorg. Med. Chem.
Lett. 2008, 18, 241.
5. (a) Ott, G. R.; Asakawa, N.; Lu, Z.; Liu, R.-Q.; Covington,
M. B.; Vaddi, K.; Qian, M.; Newton, R. C.; Traskos, J. M.;
Christ, D. D.; Decicco, C. P.; Duan, J. J.-W. Bioorg. Med.
Chem. Lett. 2008, 18, 694; (b) Chen, X.-T.; Ghavimi, B.;
Corbett, R. L.; Xue, C.-B.; Liu, R.-Q.; Covington, M. B.;
Qian, M.; Vaddi, K. G.; Christ, D. D.; Hartman, K. D.;
Ribadeneira, M. D.; Trzaskos, J. M.; Newton, R. C.;