Synthesis and SAR of α-sulfonylcarboxylic acids as potent matrix metalloproteinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.03.065

A series of novel carboxylic acid-based α-sulfone MMP inhibitors have been synthesized and the in vitro enzyme SAR is discussed. A potential binding mode in the active site of the MMP-9 homology model was highlighted. These compounds are potent MMP-9 inhi

Full text of DOI:10.1016/j.bmcl.2006.03.065

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3096–3100
Synthesis and SAR of a-sulfonylcarboxylic acids as potent matrix
metalloproteinase inhibitors
Yue-Mei Zhang,* Xiaodong Fan, Bangping Xiang, Devraj Chakravarty,
Robert Scannevin, Sharon Burke, Prabha Karnachi, Kenneth Rhodes and Paul Jackson
Drug Discovery, Johnson & Johnson Pharmaceutical Research and Development, 1000 Route 202, Raritan, NJ 08869, USA
Received 16 February 2006; revised 20 March 2006; accepted 21 March 2006
Abstract—A series of novel carboxylic acid-based a-sulfone MMP inhibitors have been synthesized and the in vitro enzyme SAR is
discussed. A potential binding mode in the active site of the MMP-9 homology model was highlighted. These compounds are potent
MMP-9 inhibitors and are selective over MMP-1.
Ó 2006 Elsevier Ltd. All rights reserved.
O
The wide variety of matrix metalloproteinases (MMPs)
normally participate in many different homeostatic tis-
sue remodeling events. The over-expression of MMP
activities has been shown to play a negative role in a
variety of disease states, such as oncology,1a rheumatoid
arthritis,1b cardiovascular diseases,1c and neurological
disorders.1d Therefore, design and synthesis of potent,
small molecule MMP inhibitors (MMPIs) has been an
area of intense focus in medicinal chemistry.
O
O
X
O
O
HO
S
HO
N
Cl
N
S
O
H
O
O
O
2 (X = O or N-R)
1(RS-130,830)
Cl
O
HO
O
S
IC50(nM):
MMP-2: 11; MMP-3: 143;
To date the hydroxamate-based compounds have
attracted the most interest due to their high in vitro
potency. With the disclosure of the spiro-b-sulfone
hydroxamate 1 as a clinical candidate, reports of a series
of hydroxamate-based a-sulfone MMPIs have fol-
lowed,2,3 exemplified by structure 2 (Fig. 1).3 The en-
zyme profile and ADME properties of the a-sulfone
hydroxamates were reported to be superior to the b-sul-
fone series.3 Nevertheless, it has been suggested that the
poor pharmacokinetic (PK) profiles of the hydroxamate
inhibitors may be associated with in vivo toxicity.4 The
carboxylic acids, the precursors of the hydroxamates,
have been shown to demonstrate more favorable PK
properties.4 However, the simple replacement of
hydroxamate by carboxylic acid as zinc-binding group
(ZBG) causes 100- to 2000-fold loss in potency.5 More-
3 Tanomastat (Bay 12-9566)
MMP-9: 301; MMP-13: 1470
Figure 1. Selected MMPIs.
over, high serum protein binding is typically suffered by
carboxylic acid MMPIs.6 A representative carboxylic
acid MMPI is tanomastat (3) with moderate MMP
activities (Fig. 1).
Our program has focused on the discovery of potent and
selective MMPIs. Here, we disclose a novel family of
carboxylic acid-based a-sulfone inhibitors 4 with excel-
lent enzyme potency and good selectivity over MMP-
1. We will describe the enzyme SAR and highlight a
potential binding mode with S01 interaction in the MMP-
9 homology model. The impact of the a-substitution of
4 on the MMP inhibitory activities will also be discussed.
O
O2
S
X
R3
Keywords: MMP inhibitors; MMP-9; MMP-2; a-sulfonylcarboxylic
acids.
HO
R2
R1
X = bond, O,
N
4
*
Corresponding author. Tel.: +1 908 281 6788; e-mail:
xufangymz@yahoo.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.03.065

Y.-M. Zhang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3096–3100
3097
O
O
O
R2
b,c
O
O
a
R2
R2
S
Br
S
Br(F)
ButO
b, c
S
ButO
Br(F)
Br
O
a
O
Br
O
O
R1
Br
R1
R3
HO
O
10
9
5
6
8
O
2O
S2
Z
O2
S
O
(O)n
S
R2
R
HN
f
O
(O)n
S
R2
d,e
Z
X
HO
d,e
HO
Br
ButO
R1
g, h
Cl
R1
Z
N
4
R2 =CH3
n = 1,2
OTs
7
11
12a Z = CO
12b Z = CH2
Z
O
R2 = H
O
O
O
f, g
S
Cl
f
O
HO
O2
S
O2
S
when R2=H, R3=Cl
i, h
HO
O
O
R1 =
N
N
Cl
4f
14
Cl
SPh
13
O
O
Scheme 2. Reagents and conditions: (a) HSPh(4-Br), TEA, CH2Cl2;
(b) NaOH (1.0 equiv), DMF/H2O (4:1); (c) MeI (1.5 equiv), NaHCO3
(1.0 equiv), 18-crown-6 (0.1 equiv); (d) TsCl, TEA, CH2Cl2; (e)
m-CPBA, CH2Cl2; (f) K2CO3, DMF; (g) (4-Cl)PhB(OH)2, 5 mol%
Pd(PPh3)4, Na2CO3 (2 M), toluene, 90 °C; (h) 12a: 0.5 N HCl/AcOH
(1:1), 100 °C; 12b: aq LiOH, THF/MeOH; i—PhSH, NaOMe, toluene,
90 °C.
Scheme 1. Reagents and conditions: (a) HSPh(4-Br or -F), TEA,
CH2Cl2; (b) OXONEÒ, MeOH/THF (1:1); (c) R2Br, K2CO3, 18-
crown-6(cat.), acetone, reflux; (d) i—when X = bond, (4-
R3)PhB(OH)2, 5 mol% Pd(PPh3)4, Na2CO3 (2 M), toluene, 90 °C;
ii—when X = acetylenyl, H– „ –Ph(4-R3), 5 mol% CuI, 5 mol%
Pd(PPh3)2Cl2, TEA, 60 °C; iii—when X = O, (4-R3)PhOH, K2CO3,
DMA, 90 °C; iv—when X = –NR0R00–, HNR0R00-Ph(4-R3), K2CO3,
DMF, 90 °C; (e) TFA/CH2Cl2 (1:2); (f) Zn/HgCl2, concd HCl, reflux.
ester 10 failed under various conditions. Finally, hydro-
lysis and alkylation with methyl iodide in one-pot reac-
tion furnished 10. Tosylation of 10 and subsequent
oxidation gave 11. The reaction of 11 with various
nucleophiles, followed by Suzuki coupling and hydroly-
sis, yielded 12 (exemplified by 12a and 12b). a-Spiro sul-
fone 13 was obtained from 11 by base-promoted
intramolecular cyclization. Further manipulation of
electron-deficient cyclopropane 13 via Michael-type
addition followed by hydrolysis gave the ring-opened
product 14.
The carboxylic acid 4 was synthesized as shown in
Scheme 1. Reaction of a-bromo tert-butyl ester 57 with
thiophenols gave sulfide 6. Oxidation of 6 with OX-
ONEÒ followed by alkylation at the acidic a-carbon
with halides gave sulfone 7 (n = 2). a-Sulfinyl analogs
7 (n = 1) may be obtained by controlling the reaction
time and equivalents of OXONEÒ used. Manipulation
of aryl bromide or fluoride of 7 via various chemical
transformations such as Suzuki coupling, Sonagishiro
reaction, and simple displacement of fluorine with phe-
nols or amines, followed by TFA deprotection of Boc,
yielded a variety of carboxylic acids 4. Further reduction
of phthalimide 4 according to Hall’s procedure8 gave
phthalimidine 4f.
Initially, as we fixed the P01 moiety as a linear 4-chloro
biaryl group, a series of a-sulfonylcarboxylic acids with
diverse a-side chains were synthesized and their inhibito-
ry activities were examined against MMP-2 and MMP-9
(Table 1). All compounds in Table 1 demonstrated mod-
erate to potent activity against MMP-2. However, an
amide or phthalimide side chain seems critical for
MMP-9 inhibitory activity. Simple hydrophobic side
chains (4a and 14) gave no MMP-9 activity. Compounds
12a and 12b were compared, the absence of a carbonyl
To rapidly explore the SAR of a-substitutions of 4, an
alternative synthesis from a-bromo lactone 8 was pur-
sued which is outlined in Scheme 2. Displacement of
the bromide 8 with 4-bromothiophenol gave 9. The
attempts of trans-esterification of lactone 9 into methyl
Table 1. In vitro IC50 and SAR of a-substitution of compounds 4
O
O2
S
O
Phth:
O
O2
HO
Ar
N
S
Cl
N
R2
R1
4d
HO
H
R2
R1
O
Compound
R1
R2
MMP-2a
MMP-9a
4a
14
–(CH2)2CH3
–(CH2)2SPh
–CH2Phth
H
H
H
H
H
H
H
792
295
556
1.8
10,005
>10,000
1049
14
4b
4c
–(CH2)2Phth
–(CH2)2Phth
–(CH2)3Phth
See Scheme 1
–(CH2)2Phth
See Scheme 2
4d (HOHNOC–)
4e
0.62
145
8.6
1.4
2124
96.5
4f
12a
12b
–CH3
–CH3
52
3060
114
>10,000
a IC50 (nM) the average value of at least two experiments.

3098
Y.-M. Zhang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3096–3100
functionality on 12b resulted in no inhibition of MMP-
9. Phthalimidine 4f gave an IC50 of 8.6 nM on MMP-2
and about 7-fold loss of MMP-9 inhibition compared
to 4c. This observation was consistent with the reported
work on b-sulfone hydroxamate inhibitors.9 The length
of the side chain was also important. The optimal ethyl
phthalimido derivative 4c was over 150-fold more po-
tent than the propyl homolog 4e and 75-fold more than
methyl analog 4b. Fully a-substituted 12a gave 8-fold
less potency against MMP-9 and about 30-fold less
against MMP-2 than 4c. This is in contrast to SAR ob-
served on the reported a-sulfone hydroxamate MMPI.2
Interestingly, different from previously reported a- and
b-sulfone hydroxamate series,2,3,9,10 there was only a
10-fold potency loss on MMP-9 and a 3-fold loss on
MMP-2 by replacing the hydroxamate ZBG (4d) with
the carboxylic acid (4c). This suggested that the novel
a-sulfonylcarboxylic acid scafold provides optimal inter-
actions with enzyme subsites and therefore compensates
for the loss of potent bidentate bindings with zinc cation
and hydrogen bond interactions with the enzyme back-
bones offered by hydroxamates.
the enzyme backbone. The IC50 on MMP-9 for 4r0 (dia-
stereomer mixture) was improved to 1.4 nM, 22-fold
more potent than 4r00, however, an additional oxygen
of 4r gave only slight improvement. The linear, rigid
S1
S2
NH
NH
A191
A189
Zn
S1’
Since an ethyl phthalimido group proved to be the most
favorable P1/P2 substituent and thus was chosen as the
benchmark, we turned our attention to SAR of the P10
substituent. A series of P10 modifications were performed
and the results of in vitro MMP-2/-9 inhibitions are
shown in Table 2. It appeared that at least one oxygen
on the sulfur of 4 is required for better interaction with
Figure 2. Compound 4r docked in homology model of MMP-9.
Table 2. MMP IC50 and SAR of P01 group of compounds 4
Compounda
X
R
MMP-2b
MMP-9b
4c
4g
4h
4i
(4-Cl)Ph–
1.8
14.0
4.4
15
(4-OMe)Ph–
(4-CN)Ph–
(4-Me)Ph–
(4-SMe)Ph–
(4-iPr)Ph–
(4-F)Ph–
0.84
3.0
1.2
4.9
4j
0.74
0.98
6.5
0.96
1.6
53
20
1.8
4k
41
4m
4n11
(4-CF3)Ph–
N
4.4
F
0.7
HN
4o
4p
O
(4-OCF3)Ph–
(4-Cl)Ph–
19
12
23
258
N
4q
(4-CN)Ph–
6.8
120
4r0
4r00 (n = 0)
(4-Me)Ph–
(4-Me)Ph–
(4-Me)Ph–
(4-Et)Ph–
(4-nPr)Ph–
(4-nBu)Ph–
(4-iPr)Ph–
(4-OCF3)Ph–
(4-Cl)Ph–
0.7
1.5
33
0.5
1.4
31
4r (n = l)
4s
1.3
1.1
2.3
1.7
1.3
1.9
4.0
54
0.8
9.7
4t
4u
4v
22
5.1
0.6
1.2
4w
4x
4y
4z12
(3-OMe)Ph–
(4-OMe)Ph–
25
101
O
a n = 2 for all compounds except indicated otherwise. 4r0 is a diastereomer mixture.
b IC50 (nM) the average value of at least two determinations.

Y.-M. Zhang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3096–3100
Table 3. IC50 values against selected MMPs of 4n and 4r
3099
a
Compound
MMP-1
MMP-2
MMP-3
MMP-9
MMP-13
4n
4r
>1000
>1000
0.7
0.7
34
28
1.8
0.5
0.8
1.4
a IC50 (nM) value of single determination.
Table 4. Pharmacokinetic parametersa of 4c and 4d in rats and in vitro metabolism
Compound
AUCb (lg-h/mL)
CLb (mL/min/kg)
HLMc T1/2 (min)
b
C0 (lM)
b
T1/2 (h)
b
Vss (L/kg)
4c
4d
41.4
13.9
1.4
0.98
18.23
2.94
0.4
0.1
0.6
2.9
230
12
a Mean value of four animals (rat) at 0.5 mg/kg iv dose.
b C0: concentration at T = 0; T1/2, apparent elimination half-life; AUC, area under the concentration-time curve; Vss, volume of distribution at steady
state; CL, systemic clearance.
c Human liver microsome.
P01 groups such as biaryl or biaryl acetylenyl groups were
preferred. More flexible linkers between two phenyl
groups of P01 resulted in loss of activity against MMP-
9 (4o, 4p, 4q, and 4z). The para-substitution on the ter-
minal phenyl was preferred over substitution at the
meta-position, 30-OMe compound 4y gave an IC50 value
of 25 nM for MMP-9. Generally, hydrophobic, electron-
donating substitutions on the terminal phenyl increased
MMP potencies, for example, 40-SMe on 4j gave an IC50
value of 0.96 nM for MMP-9 and 0.74 nM for MMP-2.
Polar, electron-withdrawing groups (EWG) such as –
CN (4h), –F (4l), and –CF3 (4m) shifted IC50 values to
double-digit nanomolar range. However, compound
4n, even with an EWG, fluorine, still gave excellent
potency, illustrating benzoimidazole as an effective P10
group. Longer biaryl acetylenyl groups as P01 moiety
were shown to be superior to biaryl groups for MMP-
9 inhibition, although MMP-2 inhibitory activity was
unchanged. Further variations on the terminal acetyle-
nyl phenyl substitution were focused on hydrophobic,
electron-rich groups based on the SAR of the biaryl P10
series. The MMP-2 inhibitory activities of compounds
4r–4x were insensitive to the length of the para-substit-
uents. However, the methyl (4r) or ethyl (4s) derivatives
have reached the maximal activities against MMP-9.
The further elongation of the para-substitution leads
to significant increase of the IC50 values of MMP-9.
These results agreed with the same SAR trend of P01 as
the previously reported sulfonamide MMP inhibitors13
and suggest a rigid, hydrophobic S01 subsite with a floor-
board at the bottom for MMP-9 and a tunnel-like S10
subsite with an open bottom for MMP-2.
the S01 pocket. This may contribute to the loss in activity
against MMP-9 for 12b.
The selectivity across several MMPs of the represen-
tative compounds is summarized in Table 3. As
expected, 4n and 4r showed high selectivity over
MMP-1, which has a shallow S01 subsite, and moder-
ate selectivity for MMP-9 vs. MMP-3 (20- to 50-
fold). The compounds 4n and 4r are also potent
MMP-13 inhibitors.
The pharmacokinetics of carboxylic acid 4c and its
hydroxamate analog 4d in rats are evaluated in Table
4. Human liver microsomes (HLM) were also used to
predict metabolic stability. Both the AUC and C0 were
improved for carboxylic acid 4c.
In summary, we have prepared a series of a-sulfone car-
boxylic acid inhibitors that are potent against MMP-2/-
9 and selective over MMP-1. The SAR of both P01 and
P1/P2 has shown that a straight, hydrophobic P01 group
with the optimal length is preferred and an amido- or
imido-type group is critical for MMP-9 inhibitory activ-
ity. The carboxylic acid series exhibits enhanced PK
properties in rats compared to its hydroxamate analog.
However, the protein binding data with highly lipophilic
compounds such as 4r, 4s, and 4x are greater than 99%,
indicating that these carboxylic acids are highly protein
bound in the blood and therefore faced a hurdle for fur-
ther development. We are continuing to optimize P1/P2
group of the carboxylate inhibitors to minimize the high
protein binding issue and hope to report the results in
due course.
The binding mode of compound 4r was proposed based
on docking the molecule into a homology model of
MMP-9 (Fig. 2). The modeling indicates that a linear
biaryl acetylenyl group well fits into the S01 pocket and
the a-sulfonyl function forms a hydrogen bond interac-
tion with Ala 189. The phthalimido side chain is located
between the S1 and S2 pockets, and one of the carbonyls
may form a hydrogen bond interaction with Ala 191.
This may rationalize that an amide or imide side chain
is critical for the maintenance of the activity. To our sur-
prise, the modeling of 12b (not shown here) indicated
that the a-sulfonyl group no longer had any interactions
with Ala 189 and the biaryl group does not fit well into
Acknowledgment
The authors thank the Chem/Bio Support group for
ADME tests.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2006.03.065.

3100
Y.-M. Zhang et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3096–3100
6. (a) Natchus, M. G.; Bookland, R. G.; Laufersweiler,
References and notes
M. J.; Pikul, S.; Almstead, N. G.; De, B.; Janusz, M.
J.; Hsieh, L.; Gu, F.; Pokross, M.; Patel, V.; Garver,
S.; Peng, S.; Branch, T. M.; King, S. L.; Baker, T.
R.; Foltz, D.; Mieling, G. J. Med. Chem 2001, 44,
1060; (b) Leff, R. L. Ann. N.Y. Acad. Sci. 1999, 878,
201.
1. (a) Vihinen, P.; Ala-aho, R.; Kahari, V. Curr. Cancer Drug
Targets 2005, 5, 203; (b) Handsley, M. M.; Edwards, D. R.
Int. J. Cancer 2005, 115, 849; (c) Iannone, F.; Lapadula,
G. Aging Clin. Exp. Res. 2003, 15, 364; (d) Tayebjee, M.
H.; Lip, G. Y.; MacFadyen, R. J. Curr. Med. Chem 2005,
12, 917; (e) Yong, V. W. Expert Opin. Investig. Drugs
1999, 8, 255; (f) Gu, Z.; Cui, J.; Brown, S.; Fridman, R.;
Mobashery, S.; Strongin, A. Y.; Lipton, S. A. J. Neurosci.
2005, 25, 640.
7. For synthesis of 5 with R1 = –(CH2)n Phth: Kolasa, T.;
Miller, M. J. Org. Chem. 1990, 55, 4246.
8. Chapman, J.; Cocolas, G.; Hall, I. J. Med. Chem 1983, 26,
243.
9. Cholleta, A. M.; Le Diguarhera, T.; Kucharczyka, N.;
Loynela, A.; Bertrandc, M.; Tuckerb, G.; Guilbaudb, N.;
Burbridgeb, M.; Pastoureaub, P.; Fradinb, A.; Sabatinib,
`
M.; Faucherea, J.; Casara, P. Bioorg. Med. Chem. 2002,
10, 531.
10. Le Diguarher, T.; Chollet, M.; Bertrand, M.; Hennig, P.;
2. (a) Aranapakam, V.; Grosu, T.; Davis, J.; Hu, B.;
Ellingboe, J.; Baker, J.; Skotnicki, J.; Zask, A.; Sung,
A.; Sharr, M.; Killar, L.; Walter, T.; Lin, G.; Cowling, R.
J. Med. Chem 2003, 46, 2361; (b) Aranapakam, V.; Davis,
J.; Grosu, T.; Baker, J.; Ellingboe, J.; Zask, A.; Levin, J.;
Du, M.; Skotnicki, J.; Sung, A.; Sharr, M.; Killar, L.;
Walter, T.; Lin, G.; Tillett, J.; Zhao, W.; Xu, Z. J. Med.
Chem 2003, 46, 2376.
3. Becker, D. P.; Villamil, C. I.; Barta, T. E.; Bedell, L. J.;
Boehm, T. L.; DeCrescenzo, G. A.; Freskos, J. N.;
Getman, D. P.; Heintz, R.; Howard, S. C.; Li, M. H.;
McDonold, J.; Carron, C.; Mehta, P.; Swearingen, C. A.
J. Med. Chem 2005, 48, 6713.
´
Raimbaud, E.; Sabatini, M.; Guilbaud, N.; Pierre, A.;
Tucker, G.; Casara, P. J. Med. Chem. 2003, 46, 3840.
11. For the synthesis of 4n:
O
i. Swern ox.
TFA/CH2Cl2
O2
S
4n
ButO
H N
2
F
ii.
OH
R1
H2N
NaHSO3
4. Tamura, Y.; Watanabe, F.; Nakatani, T.; Yasui, K.; Fuji,
M.; Komurasaki, T.; Tsuzuki, H.; Maekawa, R.; Yoshio-
ka, T.; Kawada, K.; Sugita, K.; Ohtani, M. J. Med. Chem
1998, 41, 640.
5. Skiles, J. W.; Gonnella, N. C.; Jeng, A. Y. Curr. Med.
Chem. 2001, 8, 425.
12. The linker methyl ketone of 4z was obtained through
hydration of acetylene under TFA treatment at the last
Boc cleavage step.
13. Kiyama, R.; Tamura, Y.; Watanabe, F.; Tsuzuki, H.;
Ohtani, M.; Yodo, M. J. Med. Chem 1999, 42, 1723.