Discovery of novel spirocyclopropyl hydroxamate and carboxylate compounds as TACE inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.11.034

We have discovered nanomolar inhibitors of TNF-α convertase (TACE) comprised of a novel spirocyclic scaffold and either a carboxylate or hydroxamate zinc binding moiety. X-ray crystal structures and computer models of selected compounds binding to TACE ex

Full text of DOI:10.1016/j.bmcl.2008.11.034

Bioorganic & Medicinal Chemistry Letters 19 (2009) 54–57
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of novel spirocyclopropyl hydroxamate and carboxylate
compounds as TACE inhibitors
a
a
a
a
b
Zhuyan Guo a, , Peter Orth , Shing-Chun Wong , Brian J. Lavey , Neng-Yang Shih , Xiaoda Niu ,
*
Daniel J. Lundell b, Vincent Madison a, Joseph A. Kozlowski a
a Department of Chemistry, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033-0539, USA
b Department of Inflammation, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033-0539, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
We have discovered nanomolar inhibitors of TNF-
a convertase (TACE) comprised of a novel spirocyclic
Received 17 September 2008
Revised 5 November 2008
Accepted 10 November 2008
Available online 14 November 2008
scaffold and either a carboxylate or hydroxamate zinc binding moiety. X-ray crystal structures and com-
puter models of selected compounds binding to TACE explain the observed SAR. We report the first TACE
X-ray crystal structure for an inhibitor with a carboxylate zinc ligand.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
TACE
TACE inhibitors
Hydroxamates
Carboxylates
Tumor necrosis factor-
a
(TNF-
a
) is a major immunomodulatory
Based on crystal structure information on TACE-inhibitor inter-
actions, we have carried out a structure-based design approach to
discover novel, potent TACE inhibitors. Key features in the design
of TACE inhibitors include a ligand binding to the active site zinc,
hydrogen bonding to the catalytic Glu406 carboxylate, a hydrogen
bond donor that coordinates to the backbone C@O of Gly349, a
hydrogen bond acceptor that coordinates to the backbone NH of
Gly349 and/or Leu348, and finally a large hydrophobic P10/P30
group for binding to the combined S10/S30 pocket. Each of these
components is placed on a rigid scaffold that locates each func-
tional group in the appropriate position in the enzyme. In separate
papers, we have reported hydroxamate inhibitors based on a cyclo-
propyl scaffold, which led to the discovery of inhibitors with low
nanomolar potency.8,9 As demonstrated on IK682 and also applied
to these compounds, the quinoline group was found to bind to the
S30 subsite thereby conferring excellent selectivity against several
MMPs.10,11 Crystal structures of these inhibitor complexes with
TACE reveal an induced fit to the large quinoline S30 group as a
key feature of the potency and selectivity of these inhibitors.
The spirocyclopropyl scaffold disclosed here is a ring con-
strained analog of the earlier cyclopropyl compounds (Table 1).
The synthesis is described in Scheme 1. It was hoped that in addi-
tion to projecting the zinc binding group and the hydrogen bond-
ing acceptors/donors into the desired positions, this scaffold
would project the P10 phenyl ring into the S10 pocket in an orienta-
tion similar to that of the potent and selective TACE inhibitor
IK682. On the other hand, modeling indicates that cyclization
would make the P10 group move further toward the P30, which
and proinflammatory cytokine that has been implicated in various
autoimmune disorders such as rheumatoid arthritis, Crohn’s dis-
ease, and psoriasis. The success of anti-TNF biologics such as En-
brelÒ and RemicadeÒ has led to an intensive search for orally
active small molecules that suppress TNF-
of doing so is to block release of TNF- through the use of small
molecule inhibitors of TNF- converting enzyme (TACE/ADAM17),
the metalloprotease that processes the 26 kDa membrane-bound
pro-TNF-
to its 17 kDa soluble component.3
a
activity.1,2 One way
a
a
a
An early crystallographic structure of a peptidic hydroxamate
inhibitor bound to TACE shows that the active site cleft of TACE
shares many features with the matrix metalloproteinases.4 These
similarities include an active site zinc coordinated to three histi-
dine side chains, a hydrophobic S10 pocket adjacent to it and
hydrogen bond donors/acceptors such as the carboxylate group
of E406 and the backbone amide group of G349. However, one dis-
tinguishing feature of TACE is a tunnel interconnecting S10 and S30
into a single large cavity. Therefore, TACE selectivity may be
achieved by incorporating large substituents on P10 that can not
be accommodated in the smaller S10 of MMPs. A selective TACE
inhibitor is desirable because it may avoid the musculoskeletal
side effects that have been reported for broad-spectrum and par-
tially selective MMP inhibitors.5–7
* Corresponding author. Tel.: +1 908 740 3796.
E-mail address: zhuyan.guo@spcorp.com (Z. Guo).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.11.034

Z. Guo et al. / Bioorg. Med. Chem. Lett. 19 (2009) 54–57
55
Table 1
Table 2
SAR of hydroxamate-cyclopropyl compounds and IK682
SAR for spirocyclopropyl series
O
O
NH2
O
H
R
N
T
N
N
R1
HO
O
O
R1
O
HN
HO
O
O
O
R2
R2
N
1-3
4 (IK682)
N
N
Compound
T
R
R1
R2
TACE Ki (nM)
Compound
R1
R2
TACE Ki (nM)
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
NHOH
NHOH
NHOH
NHOH
NHOH
NHOH
NHOH
OH
OH
OH
OH
OH
H
H
H
H
H
H
CH3
H
CH3
H
F
Phenyl
Phenyl
3-Pyridinyl
CH3
4-Pyridinyl
CF3
Phenyl
Phenyl
Phenyl
Phenyl
4-Pyridinyl
CF3
6.7
30
32
32
33
53
137
40
H
H
H
H
H
F
1
2
3
4
F
H
H
Phenyl
Phenyl
CH3
0.14
0.18
3
0.06
The compounds were tested in a FRET assay using the catalytic domain of TACE.
F
F
94
H
H
H
H
H
F
143
673
2140
3070
3220
>50,000
H
H
H
H
would weaken the pi–pi interactions between the P10 phenyl and
the side chain of His 405, as well as causing steric crowding of
the phenyl group with Ala439. Since the quinoline group has been
found to render excellent TACE selectivity, it was used throughout
the spirocyclopropyl series.
OH
OH
OCH3
CH3
3-Pyridinyl
Phenyl
CH3
All the analogs were prepared and tested as racemic mixtures.
The synthesis of the spirocyclopropyl hydroxamate uses chem-
istry worked out previously for N-Boc-2-pyrrolidinones and ap-
plies it to succinamide in order to prepare compounds 12–18.12
Accordingly, compound 5 was converted to the corresponding
Gringard reagent then reacted with succinamide. Reduction with
sodium cyanoborohydride gave 6. Alkylation with Eschenmoser’s
100-fold. For example, compound 12 (Ki = 6.7 nM) is 50-fold less
potent than compound (Ki = 0.14 nM), and compound 13
1
(Ki = 30 nM) 100-fold less potent than 2 (Ki = 0.18 nM). While in
the cyclopropyl series, 2-phenyl-quinoline increases potency by
>10-fold compared with 2-methyl-quinoline (e.g., compounds 3
vs 2), the potency remains the same in the hydroxamate-spirocy-
clopropyl series (e.g., compounds 13 vs 15). Replacing the phenyl
with a pyridine does not affect potency, as shown in compounds
14 and 16. In addition, compound 12 with R1 as m-fluoro-phenyl
increases potency by 5-fold, although in the cyclopropyl series
the potency remains the same. Due to the limited space in S10, sub-
stitutions on the phenyl ring with a group larger than fluorine are
expected to cause steric crowding and to lose potency, as has been
observed in the cyclopropyl series.9 N-methyl substitution of com-
salt followed by elimination gives a,b-unsaturated lactam 8 which
can be converted to the corresponding spirocyclopropane 9 using
sulfur ylide chemistry. Debenzylation of the phenol, followed by
alkylation with a quinoline chloride affords 10. The ester and
amide functionalities are then unmasked with TFA and the result-
ing compound is converted to the hydroxamate 11 using standard
coupling conditions and acid deprotection.
As shown in Table 2, the hydroxamate-spirocyclopropyl com-
pounds 12–18 are potent TACE inhibitors with Kis in the nanomo-
lar range. However, these compounds are less potent than the
corresponding hydroxamate-cyclopropyl compounds by 10- to
O
O
Br
O
O
N
Boc
R1
Boc
R1
N
a
N
d
NH
b,c
R1
R1
Bn
O
O
O
8
Bn
e, f
Bn
7
6
5
O
O
S
O
Br-
O
O
Boc
R1
O
N
g
HOHN
NH
h, i, j
O
O
O
Cl
Boc
N
O
R1
R3
N
O
O
R1
O
H
R2
R2
N
N
9
10
11
Scheme 1. Reagents and conditions: Yields shown are for R2 = H, R2 = phenyl. (a) i—Mg/THF; ii—succinamide; iii—NaCNBH3 (5% over three steps); (b) i—LDA/THF; ii—Boc2O
(68%); (c) i—LiHMDS; ii—Eschenmoser’s salt (21%); (d) i—CH3I/MeOH; ii—NaHCO3 (97%); (e) sulfur ylide/DBU/CH3CN (40%); (f) Pd(OH)2/H2 (98%); (g) Cs2CO3; aryl chloride
(75%); (h) TFA/CH2Cl2 (84%); (i) trityl-ONH2/EDC/HOBT (59%); (j) TFA/Et3SiH/CH2Cl2 (78%). When R = N-methyl then N-methyl succinamide is used in step a and step b is not
performed. When R1 = F, the same sequence of reactions is used with fluorinated 5. All final compounds are racemic.

56
Z. Guo et al. / Bioorg. Med. Chem. Lett. 19 (2009) 54–57
pound 18 decreases potency by 20-fold. Modeling indicates possi-
ble steric crowding of the N-methyl group with the C@O of Gly346.
Generally speaking, carboxylate is a much weaker zinc ligand
than hydroxamate.13 In the cyclopropyl series, replacing the
hydroxamate with carboxylate rendered the compounds inactive
(>1000-fold loss in potency, data not shown). A computational
study of the binding of zinc ligands showed that the requirement
for protonation of Glu406 or the ligand carboxylate is the primary
cause of this reduced affinity.14 However, in the spirocyclopropyl
series, the Kis are reduced by only 5- to 100-fold upon substituting
carboxylate for hydroxamate. For example, compounds 19
(Ki = 40 nM) and 12 (Ki = 6.7 nM) only differ by 6-fold, and com-
pounds 21 (Ki = 143 nM) and 13 (Ki = 30 nM) by 5-fold. In addition,
2-phenyl-quinoline increases the potency by 20-fold compared
with 2-methyl-quinoline (e.g., compounds 21 vs 24). Compounds
19 and 20 have comparable potency which means that methyl sub-
stitution of the lactam is well tolerated. Finally, as expected con-
verting the carboxylate to a methyl ester rendered compound 26
inactive. Compounds 12, 13, 17, 19, 21, and 23 showed IC50 values
in a human whole blood assay for TNF production of greater than
100 micromolar.15 Due to the lack of activity in the human whole
blood assay, compounds were not resolved into their enantiomers
or tested for selectivity against other MMPs.
Figure 1 shows the crystallographic structure of compound 21
binding to a TACE mutant enzyme (V353G).16 The carboxylate
group binds to the zinc in a bidentate fashion with O–Zn dis-
tance of 1.9 and 2.5 Å, respectively. The weaker zinc-binding
oxygen also makes a hydrogen bond with the side chain of
Glu406. Here the carboxylic acid of Glu406 is expected to be
protonated.14 The carbonyl oxygen of the lactam makes a hydro-
gen bond with the backbone NH of Leu348. A crystal water is
present as a bridge for hydrogen bonding between the lactam
NH and the C@O of Gly346. Modeling indicates that this water
molecule would be displaced by the N-methyl substitution of
compound 20. The methyl group would then make hydrophobic
contacts with the side chains of Ala439 and Leu348. Therefore,
the gain in hydrophobic interactions balances the loss of hydro-
gen bonding. This is consistent with the comparable potency ob-
served in compounds 19 and 20. In contrast, modeling indicates
that the lactam NH in the hydroxamate-spirocyclopropyl series
is involved in direct hydrogen bonding interactions with the
C@O of Gly346 and that N-methyl substitution would cause
Figure 2. Overlay of the crystal structures of compounds 2 (magenta) and 21
(green). Only ligands and the active site zinc are shown for clarity.
crowding in that region. Multiple attempts to obtain a crystal
structure of a hydroxamate compound from this series were
unsuccessful. The crystal structure of compound 21 also indi-
cates a small space available to accommodate the m-fluorine
which would make favorable contacts with the side chain of
Val 402. This is consistent with the increase in potency of
compounds 19 versus 21. Figure 2 shows an overlay of the crys-
tal structure of compounds 2 and 21. The quinoline group binds
to the S10/S30 region in the same orientation as in the cyclopro-
pyl series. The 2-phenyl group makes stacking interactions with
the side chain of Glu398 and there is some change in location of
the amide moiety relative to the lactam carbonyl. It is not
obvious from the X-ray structures why compounds 22 and 25
with 4- and 3-pyridinyl would lose potency by 5- and 20-fold
relative to phenyl at this position since this group is solvent
exposed.
We attempted to explain the less than expected potency shift of
the carboxylate- versus hydroxamate-spirocyclopropyl com-
pounds. Because of the extra carbon spacer of the spirocyclopropyl
relative to the cyclopropyl scaffolds, the P10 group moved further
toward the prime side and weakened pi–pi stacking interactions
between the middle phenyl ring and His405. The phenyl group
would also have steric crowding with the side chain of Ala439.
On the other hand, the carboxylate zinc binding group being short-
er, allows the P10 group to move away slightly and therefore alle-
viate the steric conflict with Ala439.
In summary, we have reported a novel scaffold that affords po-
tent TACE inhibition for both hydroxamate and carboxylate com-
pounds. Carboxylate TACE inhibitors are particularly interesting
because of their chemical stability and potentially favorable phar-
macokinetics. Although carboxylate inhibitors binding to MMPs
have been reported, there are only limited reports on binding to
TACE.17 While further work is needed to improve potency, this
work opens up the possibility of designing potent and selective
small molecule carboxylate TACE inhibitors.
Acknowledgments
The authors thank Dr. Corey Strickland for critical reading of the
manuscript and helpful suggestions.
References and notes
1. Bemelmans, M. H.; van Tits, L. J.; Buurman, W. A. Crit. Rev. Immunol. 1996, 16, 1.
2. Aggarwal, B. B.; Natarajan, K. Eur. Cytokine Netw. 1996, 7, 93.
3. Black, R. A.; Rauch, C. T.; Kozlosky, C. J.; Peschon, J. J.; Slack, J. L.; Wolfson, M. F.;
Castner, B. J.; Stocking, K. L.; Reddy, P.; Srinivasan, S.; Nelson, N.; Boiani, N.;
Schooley, K. A.; Gerhart, M.; Davis, R.; Fitzner, J. N.; Johnson, R. S.; Paxton, R. J.;
March, C. J.; Cerretti, D. P. Nature 1997, 385, 729.
4. Maskos, K.; Fernandez-Catalan, C.; Huber, R.; Bourenkov, G. P.; Bartunik, H.;
Ellestad, G. A.; Reddy, P.; Wolfson, M. F.; Rauch, C. T.; Castner, B. J.; Davis, R.;
Figure 1. Crystal structure of compound 21 binding to TACE (PDB ID: 3EWJ). The
inhibitor (green) and the flexible loop (grey) are shown as stick model. The crystal
water bridging the lactam NH and the C@O of Gly346 is also shown. Note that the
compound was prepared and soaked as a racemic mixture. The X-ray structure
shows only the S-enantiomer bound.

Z. Guo et al. / Bioorg. Med. Chem. Lett. 19 (2009) 54–57
57
Clarke, H. R.; Petersen, M.; Fitzner, J. N.; Cerretti, D. P.; March, C. J.; Paxton, R. J.;
Black, R. A.; Bode, W. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 3408.
5. Nemunaitis, J.; Poole, C.; Primrose, J.; Rosemurgy, A.; Malfetano, J.; Brown, P.;
Berrington, A.; Cornish, A.; Lynch, K.; Rasmussen, H.; Kerr, D.; Cox, D.; Millar, A.
Clin. Cancer Res. 1998, 4, 1101.
10. Duan, J. J.-W.; Chen, L.; Wasserman, Z. R.; Lu, Z.; Liu, R.-Q.; Covington, M. B.;
Qian, M.; Hardman, K. D.; Magolda, R. L.; Newton, R. C.; Christ, D. D.; Wexler, R.
R.; Decicco, C. P. J. Med. Chem. 2002, 45, 4954.
11. Niu, X.; Umland, S.; Ingram, R.; Beyer, B. M.; Liu, Y.-H.; Sun, J.; Lundell, D.; Orth,
P. Arch. Biochem. Biophys. 2006, 451, 43.
6. Wojtowicz-Praga, S.; Torri, J.; Johnson, M.; Steen, V.; Marshall, J.; Ness, E.;
Dickson, R.; Sale, M.; Rasmussen, H. S.; Chiodo, T. A.; Hawkins, M. J. J. Clin.
Oncol. 1998, 6, 2150.
7. List, A. F.; Kurtin, S. E.; Callander, N.; Anderson, J. E.; Foran, J. M.; Rabinowitz, I.;
Glinsmann-Gibson, B. J.; Bellamy, W. T.; Bennet, J. M.; Collier, M.; Mazabel, E.
Abstract of Papers, 44th Annual Meeting of American Society of Hematology,
Philadelphia, PA, 2002; 100, Abstract 3119.
8. Zhu, Z.; Mazzola, R.; Sinning, L.; McKittrick, B.; Niu, X.; Lundell, D.; Sun, J.; Orth,
P.; Guo, Z.; Madison, V.; Ingram, R.; Beyer, B. M. J. Med. Chem. 2008, 51, 725.
9. Mazzola, R.; Zhu, Z.; Sinning, L.; McKittrick, B.; Lavey, B.; Spitler, J.; Kozlowski,
J.; Shih, N.-Y.; Zhou, G.; Guo, Z.; Orth, P.; Madison, V.; Niu, X.; Sun, J.; Lundell, D.
Bioorg. Med. Chem. Lett. 2008, 18, 5809.
12. Maldaner, A. O.; Pilli, R. A. Tetrahedron 1999, 55, 13321.
13. Skiles, J. W.; Gonnella, N. C.; Jeng, A. Y. Curr. Med. Chem. 2001, 28, 425.
14. Cross, J. B.; Duca, J. S.; Kaminski, J. J.; Madison, V. J. Am. Chem. Soc. 2002, 124,
11004.
15. Human whole blood is diluted 1:1 with serum free medium (RPMI,
glutamine, Pen-Strep, HEPES) and incubated with compounds in final
volume of 360 L for 1 h at 37 °C. Forty microliters of LPS (10 g/ml) is then
added. Supernatant is collected after 3.5 h incubation and the concentration of
TNF- is determined by ELISA (R&D Systems).
L-
a
l
l
a
16. Ingram, R. N.; Orth, P.; Strickland, C. L.; Le, H. V.; Madison, V.; Beyer, B. M.
Protein Eng. Des. Sel. 2006, 19, 155.
17. Park, K.; Levin, J. I.; Ariamala G. US2006/0160884.