Synthesis and SAR of diazepine and thiazepine TACE and MMP inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.01.053

Potent and selective TACE and MMP inhibitors utilizing the diazepine and thiazepine ring systems were synthesized and evaluated for biological activity in in vitro and in vivo models of TNF-α release. Oral activity in the mouse LPS model of TNF-α release

Full text of DOI:10.1016/j.bmcl.2005.01.053

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1641–1645
Synthesis and SAR of diazepine and thiazepine
TACE and MMP inhibitors
Arie Zask,a,* Joshua Kaplan,a XueMei Du,a Gloria MacEwan,a Vincent Sandanayaka,a
Nancy Eudy,a Jeremy Levin,a Guixian Jin,a Jun Xu,a Terri Cummons,b Dauphine Barone,c
Semiramis Ayral-Kaloustiana and Jerauld Skotnickia
aWyeth Research, 401 N. Middletown Road, Pearl River, NY 10965, USA
bWyeth Research, PO Box CN 8000, Princeton, NJ 08543, USA
cAmgen, Seattle, WA 98101, USA
Received 11 November 2004; revised 19 January 2005; accepted 21 January 2005
Abstract—Potent and selective TACE and MMP inhibitors utilizing the diazepine and thiazepine ring systems were synthesized and
evaluated for biological activity in in vitro and in vivo models of TNF-a release. Oral activity in the mouse LPS model of TNF-a
release was seen. Efficacy in the mouse collagen induced arthritis model was achieved with diazepine 20.
Ó 2005 Elsevier Ltd. All rights reserved.
Tumor necrosis factor-a (TNF-a) is a pro-inflammatory
cytokine that is believed to play a role in the etiology of
rheumatoid arthritis (RA), CrohnÕs disease and other
disease states.1 Small molecule mediators of TNF-a ac-
tion are important targets for the treatment of these
inflammatory diseases. EnbrelTM, a soluble TNF recep-
tor-Fc dimer, modulates TNF-a by interacting with
both the 26 kDa membrane-bound form of TNF-a
and 17 kDa soluble TNF-a and is a highly effective
RA therapy.2 An alternative paradigm for affecting
TNF-a levels is via the inhibition of TNF-a converting
enzyme (TACE/ADAM-17), the enzyme primarily
responsible for the shedding of membrane-bound
TNF-a to provide its soluble form.3 A number of small
molecule TACE inhibitors possessing excellent enzyme
and cellular potency have recently been disclosed having
diverse selectivity profiles versus TACE and the MMPs.4
Since a variety of MMPs have been found to be overex-
pressed in RA synovial tissue and have been implicated
in the destruction of cartilage in RA joints, the optimal
TACE/MMP selectivity profile for a drug to treat rheu-
matoid arthritis is at present unresolved.5 In addition,
several MMP inhibitors have been reported to exhibit
a musculoskeletal syndrome as a side effect, however
the cause of this toxicity (e.g. inhibition of MMP-1,
MMP-14 or sheddases) is unclear.4b
Molecular modeling studies of the TACE X-ray crystal
structure6 indicated that an acetylenic P10 group could
fit in the tunnel connecting the S10 and S30 pockets.7
The butynyloxy P10 group has been found to be effective
for increasing potency versus the TACE enzyme in a
cell-free assay, and in LPS-stimulated THP-1 cells for
anthranilic acid-based and benzodiazepine-based sulfon-
amide hydroxamic acids (Fig. 1).7,8 The butynyloxy
group has also been found to enhance the MMP-1 selec-
tivity of these compounds. More recent work on the
incorporation of polar groups onto the benzodiazepine
scaffold, to enhance aqueous solubility and in vivo activ-
ity, has also been described.9 As in the benzodiazepines,
the diazepine and thiazepine ring systems constrains the
butynyloxy P10 group relative to the hydroxamic acid.
However, removal of the phenyl ring decreases the
O
S
O
S
O
N
O
O
O
O
O
N
HOHN
HOHN
N
O
*
Corresponding author. Fax: +1 845 602 5561; e-mail:
zaska@wyeth.com
Figure 1. Butynyloxy TACE/MMP inhibitors.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.01.053

1642
A. Zask et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1641–1645
lipophilicity of the system, increases the basicity of the
ring nitrogen in the diazepines and changes the confor-
mation of the seven membered ring. Through the use
of the two possible 1,4-diazepine isomers, the location
(relative to the hydroxamic acid) of the basic nitrogens,
as well as the substituents attached to them can be
altered, providing an opportunity to explore the toler-
ances of the enzyme for interactions with these groups,
as well as their effects on cellular and in vivo activity.
Similarly, the two isomeric 1,4-thiazepines were synthe-
sized and the effects of the sulfur and sulfone moieties on
activity were evaluated.
version of the ester to a hydroxamic acid as described
above gave compounds 9–11.
The isomeric 1,4-diazepine-5-carboxylate scaffold was
synthesized in an analogous way by treating 2,4-di-
bromo-butyric acid tert-butyl ester (12) and N,N0-dibenz-
yl-1,2-ethanediamine (13) with Et3N in CH2Cl2
(Scheme 2). The benzyl protecting groups were removed
by catalytic hydrogenolysis using 10% Pd/C to give 14.
Selective protection of the more accessible nitrogen by
treatment with Boc2O followed by sulfonylation with 5
gave 15. Selective removal of the N-Boc group by treat-
ment of 15 with HCl in dioxane gave 16, whose free
amine group could be acylated or alkylated. Subsequent
removal of the t-butyl group using TFA in CH2Cl2 gave
the corresponding carboxylic acid, which was converted
to the hydroxamic acid by treatment with oxalyl chlo-
ride and hydroxylamine to provide 17 and 18. Alterna-
tively, treatment of 15 with TFA in CH2Cl2 removed
both protecting groups. The nitrogen was then Boc pro-
tected as above and the carboxylic acid was converted to
the hydroxamic acid to give 19. Removal of the Boc
group with HCl in dioxane gave 20, which could be
selectively alkylated on the ring nitrogen with CH3I
and Et3N to give 21a. Alternatively, treatment of 20
with HOAc and NaBH4 gave the N-ethyl analog 21b.
The 1,4-diazepine-2-carboxylate ring system was pre-
pared by treating N,N0-dibenzyl-1,3-propanediamine
(1) and ethyl 2,3-dibromopropionate (2) with triethyl-
amine in benzene (Scheme 1).10 The benzyl protecting
groups were removed by catalytic hydrogenolysis using
10% Pd/C to give 3. Selective protection of the more ba-
sic and sterically accessible amine using Boc2O gave 4.
Sulfonylation of 4 with 4-(2-butynyloxy)benzenesulfonyl
chloride (5) gave 6. Saponification of the ethyl ester in 6
with NaOH gave the corresponding carboxylic acid,
which was converted to the hydroxamic acid (7) by
treatment with hydroxylamine in the presence of EDC
and HOBT. Subsequent removal of the Boc protecting
group with HCl gave 8. Alternatively, removal of the
Boc group in 6 with HCl released the basic amine, which
could then be N-acylated, N-sulfonylated or N-alkylated
by treatment with acyl chlorides, sulfonyl chlorides or
alkyl halides, respectively. N-Acylation could also be ef-
fected by coupling the free nitrogen with a carboxylic
acid in the presence of EDC and HOBT. Treatment of
the basic amine with cyclohexyl isocyanate in the pres-
ence of HunigÕs base gave the corresponding urea. Con-
The isomeric 1,4-thiazepine-3-carboxylic acid scaffold
was constructed beginning with D-cysteine (Scheme 3).
Sequential treatment with 3-bromo-propan-1-ol and 5,
followed by conversion of the carboxylic acid to the t-
butyl ester gave 22. Cyclization by treatment with
PPh3 and diethyl azodicarboxylate gave 23. Deprotec-
tion of the ester by treatment with TFA and conversion
to the hydroxamic acid gave 24. Oxidation of 24
with peracetic acid gave the sulfone 25. In a similar fash-
ion, the 1,4-thiazepine-5-carboxylic acid scaffold was
NHBn
NHBn
O
H
N
O
EtO
O
H
N
Br
Ot-Bu
a, b
c
EtO
N
Boc
Br
O
H
N
4
a, b
c, d
1
12
N
H
+
+
+
t-BuO
O
NHBn
3
NH
ClO2S
O
Br
OEt
BnHN
tBuO
Br
14
13
5
2
d
O
S
O
S
O
e
O
O
O
O
S
O
S
O
O
j, f, k
or
f, c, g
O
O
O
O
N
N
e, f, g
or
O
O
HOHN
N
N
g, h, e, f
N
N
HOHN
EtO
or
R
R
g, i, e, f
N
Boc
N
R
15 R = Boc
16 R = H
17 R = COR
18 R = Bn
19 R = Boc
20 R = H
6
7 R = Boc
8 R = H
h
i
l
21a R = CH3
21b R = Et
9 R = COR', SO2R'
10 R = CH2R'
11 R = CONHC6H13
Scheme 2. Reagents and conditions (a) Et3N, CH2Cl2; (b) H2, 10% Pd/
C; (c) Boc2O, dioxane, NaOH (aq.); (d) 5, Et3N, DMAP; (e) 4 N HCl/
dioxane; (f) TFA, CH2Cl2; (g) EDC, HOBT, NH2OH, DMF; (h) 2 N
HCl/dioxane; (i) CH3I, Et3N, MeOH; (j) RCOCl or BnBr, Et3N,
CH2Cl2; (k) (COCl)2, DMF, NH2OH; (l) HOAc, NaBH4.
Scheme 1. Reagents and conditions (a) Et3N, PhH; (b) H2, 10% Pd/C;
(c) Boc2O, dioxane, NaOH (aq.); (d) pyridine; (e) NaOH; (f) EDC/
HOBT/NH2OH; (g) HCl; (h) RCOCl or RSO2Cl or RX, i-Pr2NEt or
RCO2H/EDC/HOBT; (i) C6H13NCO, i-Pr2NEt.

A. Zask et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1641–1645
1643
were the analogs with an N-COCH3 (9c) group, an N-
COCH2NHCO2tBu (9d) group or a N-SO2CH3 (9g)
group. Activity was also seen in the LPS stimulated
THP-1 cellular assay.12 Compounds 9c, 9d and 9g with
the most potent TACE enzyme activity were also among
the best inhibitors in the cellular assay for this scaffold.
Side chains with basic amine substituents (9e, f), or a ba-
sic ring nitrogen (8, 10a, 10b), were not effective in
improving cellular activity in this series. Large groups
on nitrogen were tolerated indicating interaction with
a solvent exposed region in the enzyme. Several com-
pounds in this series were orally active in the mouse
LPS induced TNF-a release assay.12,13 Compound 9d
showed the greatest inhibition, with 68% inhibition of
TNF-a release 1 h after oral dosing at 25 mpk. The thi-
azepine analog 24 was a potent TACE inhibitor with 20-
fold selectivity versus MMP-1. It performed somewhat
better in the THP-1 cellular assay than the correspond-
ing diazepine analogs, and gave 80% inhibition at 3 lM.
The corresponding sulfone (25) was markedly better
than 24 in the LPS assay, and gave 77% inhibition of
TNF-a release after 1 h (50 mpk po) compared with
24% inhibition for 24.
O
S
O
O
OH
O
O
O
NH2
SH
a, b, c
NH
HO
tBuO
S
22
O
S
O
S
O
O
O
O
O
N
N
e, f
d
tBuO
HOHN
S
X
23
24 X = S
g
25 X = SO2
Scheme 3. Reagents and conditions (a) Br(CH2)3OH, NaOH, EtOH;
(b) 5, Et3N; (c) t-BuBr, DMA; (d) PPh3, DEAD, THF; (e) TFA,
CH2Cl2; (f) EDC, HOBT, NH2OH, DMF; (g) CH3CO3H.
O
S
O
O
OH
O
O
O
NH2
SH
a, b, c
NH
HO
MeO
S
26
27
The isomeric 1,4-diazepine-5-hydroxamic acid scaffold
provided TACE inhibitors that were comparable in po-
tency to analogs utilizing the 1,4-diazepine-2-hydroxa-
mic acid system (Table 2). Potent TACE activity was
achieved independent of the nature of the substituent
on nitrogen. Hydrogen (20), alkyl (18, 21), carbonyl
(17a–d) or Boc (19) substitution gave potent inhibitors.
Moreover, this scaffold gave analogs that were more po-
tent inhibitors in the cellular THP-1 assay relative to the
isomeric diazepine scaffold. More potent MMP-13 inhi-
bition was seen for this scaffold as well. Selectivity versus
MMP-1 was approximately 10-fold in most cases. The
potent cellular activity of this series led to potent activity
in the mouse LPS induced TNF-a release assay. Several
analogs had good activity 6 h after a single 50 mpk po
dose. Although basicity had no effect on potency in
the in vitro assays, in vivo activity was best for the basic
NH compound, 20, that displayed excellent activity both
at 1 h (97% inhibition) and 6 h (62% inhibition) after
dosing at 50 mpk po in the LPS assay. The Boc analog
(19) similarly had good in vivo activity, although loss
of the Boc group may occur in vivo. Due to its superior
activity profile, compound 20 was tested in an efficacy
model of arthritis, the mouse collagen induced arthritis
(CIA) model of arthritis.14 Upon dosing at 20 mpk bid
po, 20 gave a 56% reduction of the mean clinical score
relative to control (HulgG) after 35 days. The thiazepine
analog 29, and the corresponding sulfone 30, showed
potent TACE enzyme activity and excellent inhibition
of TNF-a release in the THP-1 cellular assay. Both ana-
logs showed potent oral activity in the mouse LPS assay
of TNF-a release.
O
S
O
S
O
O
O
O
e, f
or
e, g, f
O
N
N
d
MeO
HOHN
S
X
28
29 X = S
30 X = SO2
Scheme 4. Reagents and conditions (a) Br(CH2)2OH, NaOH, EtOH;
(b) 5, Et3N; (c) TMSCHN2; (d) PPh3, DEAD, THF; (e) 5 N NaOH,
MeOH; (f) EDC, HOBT, NH2OH, DMF; (g) oxone.
synthesized by S-alkylation of 2-amino-4-mercapto-
butyric acid (26) with 2-bromoethanol (Scheme 4). Sub-
sequent N-sulfonation with 5 and esterification using
TMSCHN2 gave 27. Cyclization of 27 by treatment with
PPh3 and diethyl azodicarboxylate gave 28, which was
hydrolyzed and converted to the hydroxamic acid 29
as described above. Alternatively, saponification of 28
followed by oxidation to the sufone and conversion to
the hydroxamic acid gave 30.
Analogs were tested for their ability to inhibit TACE,
MMP-1 and MMP-13.11 The two collagenases, MMP-
1 and MMP-13, are representative of enzymes with deep
and shallow S10 pockets, respectively, and thus may be
predictive of selectivity towards other MMPs. Further-
more, MMP-13 is a key enzyme involved in type-II cartil-
age degradation and its inhibition may be particularly
beneficial in arthritis. The 1,4-diazepine-2-hydroxamic
acid scaffold provided potent inhibitors of cell-free
TACE enzyme, with some analogs having IC50 values
of less than 10 nM (9c, d, g) and greater than 100-fold
selectivity for TACE over MMP-1 (9c–e, g) (Table 1).
Inhibition of MMP-13 activity was seen as well. The
most potent compounds in the TACE enzyme assay
Due to the excellent in vivo activity of 20, the pharmaco-
kinetics were examined via three different modes of
administration, intravenous (iv), intraperitoneal (ip)
and oral (po), to female Balbc mice. Animals received
a single iv bolus of 5 mpk, an ip dose of 10 mpk, and
a po dose of 50 mpk. The compound exhibited moderate

1644
A. Zask et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1641–1645
Table 1. 1,4-Diazepine-2-hydroxamic acids and 1,4-thiazepine-3-hydroxamic acids
O
O
S
O
O
N
HOHN
X
Compound
X
MMP-1a
MMP-13a
TACEa
THPb
LPSc
t
7
NCO2 Bu
NH
NCOPh
26%
576
363
225
65
35
11
17
10
4
13
43
32
41
74
71
62
33
71
74
62
42
41
36
80
83
—
4
8
9a
9b
9c
9d
9e
9f
362
645
—
—
10
68
—
—
35
0
NCO-2-thienyl
NCOCH3
142
305
446
1063
395
128
140
88
2117
4584
3031
19%
40%
1136
39%
36%
212
t
NCOCH2NHCO2 Bu
NCOCH2NH2
NCO(CH2)3N(CH3)2
NSO2CH3
7
12
75
5
9g
9h
9i
NSO2(CH2)2CH3
NSO2Ph
NCH3
30
68
24
61
14
20
28
46
—
—
—
24e
77e
10a
10b
11
24d
25d
180
52
NCH2Ph
NCONHC6H3
S
SO2
34%
408
307
36
334
311
a IC50 (nM) or % inhibition at 1 lM.
b % inhibition at 3 lM.
c % inhibition after 1 h (25 mpk po).
d S configuration.
e % inhibition after 1 h (50 mpk po).
Table 2. 1,4-Diazepine-5-hydroxamic acids and 1,4-thiazepine-5-hydroxamic acids
O
O
S
O
O
N
HOHN
X
Compound
X
MMP-1a
MMP-13a
TACEa
THPb
LPS (1 h)c
LPS (6 h)c
17a
17b
17c
17d
18
NCOPh
113
207
142
277
287
92
4
20
9
10
19
23
15
83
74
24
23
37
13
19
93
63d
95
79
45
75
83
82
57d
87
92
35
—
—
—
17
88
97
32
—
70
94
0
—
—
—
0
NCOCH2CH3
NCOCH(CH3)2
NCOCH2OCH3
NCH2Ph
31
7
t
19
20
NCO2 Bu
5
43
62
39
—
41
—
NH
NCH3
NEt
S
274
238
285
171
114
32
21
17
11
8
21a
21b
29
30
SO2
a IC50 (nM).
b % inhibition at 3 lM.
c % inhibition (50 mpk po).
d % inhibition at 1 lM.
clearance (3.86 L/h/kg) and volume of distribution
(2.49 L/kg). The oral bioavailability was 18% and the
half-life was similar and moderate in length by each
mode of administration (iv = 1.43 h; ip = 2.64 h;
po = 0.86 h). These results are consistent with the supe-
rior in vivo efficacy seen for this compound.
and MMP inhibitors containing the butynyloxy P10
group. Potent TACE and MMP-13 inhibitors that were
selective versus MMP-1 were prepared. TACE enzyme
activity translated into cellular activity in the LPS stimu-
lated THP-1 assay. The 1,4-diazepine-5-hydroxamic
acid and 1,4-thiazepine-5-hydroxamic acid scaffolds pro-
vided inhibitors with superior cellular and in vivo activ-
ity. Diazepine 20, with a basic ring nitrogen, gave
excellent in vivo activity in the mouse LPS induced
In summary, the 1,4-diazepine and 1,4-thiazepine ring
systems are a useful scaffold for the synthesis of TACE

A. Zask et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1641–1645
1645
Pat.; March, C. J.; Paxton, R. J.; Black, R. A.; Bode, W.
Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 3408.
TNF-a release assay as well as oral efficacy in the CIA
model of arthritis.
7. (a) Chen, J. M.; Jin, G.; Sung, A.; Levin, J. I. Bioorg. Med.
Chem. Lett. 2002, 12, 1195; (b) Levin, J. I.; Chen, J. M.;
Du, M. T.; Nelson, F. C.; Killar, L. M.; Skala, S.; Sung,
A.; Jin, G.; Cowling, R.; Barone, D.; March, C. J.;
Mohler, K. M.; Black, R. A.; Skotnicki, J. S. Bioorg. Med.
Chem. Lett. 2002, 12, 1199.
8. Nelson, F. C.; Delos Santos, E.; Levin, J. I.; Chen, J. M.;
Skotnicki, J. S.; DiJoseph, J. F.; Sharr, M. A.; Sung, A.;
Killar, L. M.; Cowling, R.; Jin, G.; Roth, C. E.; Albright,
J. D. Bioorg. Med. Chem. Lett. 2002, 12, 2867.
Acknowledgements
We thank R. Cowling and A. Sung and Wyeth DSM for
technical support, and R. A. Black and K. M. Mohler
for their insight and guidance during the course of this
work.
9. Levin, J. I.; Nelson, F. C.; Delos Santos, E.; Du, M. T.;
MacEwan, G.; Chen, J. M.; Semiramis Ayral-Kaloustian,
S.; Xu, J.; Jin, G.; Cummons, T.; Barone, D. Bioorg. Med.
Chem. Lett. 2004, 14, 4147.
10. Barraclough, P.; Farrant, R. D.; Kettle, D.; Smith, S. J.
Chem. Res., Synop. 1991, 11, 306.
References and notes
1. Newton, R. C.; Decicco, C. P. J. Med. Chem. 1999, 42, 2295.
2. Nanda, S.; Bathon, J. M. Exp. Opin. Pharmacother. 2004,
5, 1175.
11. (a) Weingarten, H.; Feder, J. Anal. Biochem. 1985, 147,
437; (b) Jin, G.; Huang, X.; Black, R.; Wolfson, M.;
Rauch, C.; McGregor, H.; Ellestad, G.; Cowling, R. Anal.
Biochem. 2002, 302, 269–275.
12. Beck, G.; Bottomley, G.; Bradshaw, D.; Brewster, M.;
Broadhurst, M.; Devos, R.; Hill, C.; Johnson, W.; Kim,
H.-J.; Kirtland, S.; Kneer, J.; Lad, N.; Mackenzie, R.;
Martin, R.; Nixon, J.; Price, G.; Rodwell, A.; Rose, F.;
Tang, J.-P.; Walter, D. S.; Wilson, K.; Worth, E. J.
Pharmacol. Exp. Ther. 2002, 302, 390.
13. Mohler, K. M.; Sleath, P. R.; Fitzner, J. N.; Cerretti, D.
Pat.; Alderson, M.; Kerwar, S. S.; Dauphine, S. T.; Otten-
Evans, C.; Greenstreet, T.; Weerawarna, K.; Kronheim, S.
R.; Petersen, M.; Gerhart, M.; Kozlosky, C. J.; March, C.
J.; Black, R. A. Nature 1994, 317, 218.
3. Killar, L.; White, J.; Black, R.; Peschon, J. Ann. N. Y.
Acad. Sci. 1999, 878, 442.
4. (a) Skotnicki, J. S.; Levin, J. I. In Annual Reports in
Medicinal Chemistry; Bristol, J. A., Ed.; Academic: San
Diego, 2003; Vol. 38, pp 153–162; (b) Skotnicki, J. S.;
DiGrandi, M. J.; Levin, J. I. Curr. Opin. Drug Discovery
Dev. 2003, 6, 742.
5. (a) Yoshihara, Y.; Nakamura, H.; Obata, K.; Yamada,
H.; Hayakawa, T.; Fujikawa, K.; Okada, Y. Ann. Rheum.
Dis. 2000, 59, 455; (b) Konttinen, Y. T.; Ainola, M.;
Valleala, H.; Ma, J.; Ida, H.; Mandelin, J.; Kinne, R. W.;
Santavirta, S.; Sorsa, T.; Lopez-Otin, C.; Takagi, M. Ann.
Rheum. Dis. 1999, 58, 691.
6. Maskos, K.; Fernandez-Catalan, C.; Huber, R.; Bouren-
kov, G. P.; Bartunik, H.; Ellestad, G. A.; Reddy, P.;
Wolfson, M. F.; Rauch, C. T.; Castner, B. J.; Davis, R.;
Clarke, H. R. G.; Petersen, M.; Fitzner, J. N.; Cerretti, D.
14. Trentham, D. E.; Townes, A. S.; Kang, A. H. J. Exp.
Med. 1977, 146, 857.