Synthesis and structure-activity relationship of a novel sulfone series of TNF-α converting enzyme inhibitors

Article abstract of DOI:10.1016/j.bmcl.2004.06.049

Replacement of the amide functionality in IM491 (N-hydroxy-(5S,6S)-1- methyl-6-{[4-(2-methyl-4-quinolinylmethoxy)anilinyl]carbonyl} -5-piperidinecarboxamide) with a sulfonyl group led to a new series of α,β-cyclic and β,β-cyclic γ-sulfonyl hydroxamic acids, which were potent TNF-α converting enzyme (TACE) inhibitors. Among them, inhibitor 4b (N-hydroxy-(4S,5S)-1-methyl-5-{[4-(2-methyl-4- quinolinylmethoxy)phenyl]sulfonylmethyl}-4-pyrrolidinecarboxamide) exhibited IC₅₀ values of <1 nM and 180 nM in porcine TACE (pTACE) and cell assays, respectively, with excellent selectivity over MMP-1, -2, -9 and -13 and was orally bioavailable with an F value of 46% in mice.

Full text of DOI:10.1016/j.bmcl.2004.06.049

Bioorganic & Medicinal ChemistryLetters 14 (2004) 4453–4459
Synthesis and structure–activity relationship of a novel
sulfone series of TNF-a converting enzyme inhibitors
Chu-Biao Xue,* Xiao-Tao Chen,* Xiaohua He, John Roderick, Ronald L. Corbett,
Bahman Ghavimi, Rui-Qin Liu, Maryanne B. Covington, Mingxin Qian,
Maria D. Ribadeneira, Krishna Vaddi, James M. Trzaskos, Robert C. Newton,
James J.-W. Duan and Carl P. Decicco
Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, NJ 08543-4000, USA
Received 14 June 2004; revised 16 June 2004; accepted 16 June 2004
Abstract—Replacement of the amide functionalityin IM491 ( N-hydroxy-(5S,6S)-1-methyl-6-{[4-(2-methyl-4-quinolinylmeth-
oxy)anilinyl]carbonyl}-5-piperidinecarboxamide) with a sulfonyl group led to a new series of a,b-cyclic and b,b-cyclic c-sulfonyl
hydroxamic acids, which were potent TNF-a converting enzyme (TACE) inhibitors. Among them, inhibitor 4b (N-hydroxy-
(4S,5S)-1-methyl-5-{[4-(2-methyl-4-quinolinylmethoxy)phenyl]sulfonylmethyl}-4-pyrrolidinecarboxamide) exhibited IC50 values of
<1nM and 180nM in porcine TACE (pTACE) and cell assays, respectively, with excellent selectivity over MMPꢀ1, ꢀ2, ꢀ9 and
ꢀ13 and was orallybioavailable with an F value of 46% in mice.
ꢀ 2004 Published by Elsevier Ltd.
In our previous communication,1a we disclosed the ra-
tional design of a novel series of cyclic succinate TACE
(TNF-a converting enzyme)2 inhibitors based on a
broad spectrum MMP (matrix metalloproteinase) inhib-
itor. Extensive SAR studies from this series led to the
identification of IM4911b (Fig. 1), one of the most effec-
tive TACE inhibitors reported so far in the literature3
inhibiting TNF-a release in a LPS-stimulated human
whole blood assay(WBA, IC 50 =20nM) with excellent
selectivity(over MMPs) and pharmacokinetics profiles.
Since an aniline moiety(A, Fig. 1) was present in
IM491 through an amide linkage to the piperidine core
(B, Fig. 1), we were concerned about the break-down of
this amide bond and the potential toxicityof the
released aniline residue (A). We therefore investigated
the stabilityof the amide bond under phsyiological
conditions. In a simulated gastric fluid at pH=7.5,
about 0.5% of A was generated from IM491 over a
N
O
N
O
O
CONHOH
NH2
A
CONHOH
N
H
N
O
IM491
HO
B
N
Figure 1.
Keywords: TNF-a production; TNF-a converting enzyme inhibitors; Sulfone derivatives.
*
Corresponding authors. Tel.: +1-302-498-6706; fax: +1-302-425-2750; e-mail addresses: cxue@incyte.com; xiao-tao.chen@bms.com
0960-894X/$ - see front matter ꢀ 2004 Published byElsevier Ltd.
doi:10.1016/j.bmcl.2004.06.049

4454
C.-B. Xue et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4453–4459
period of 24h at 37ꢁC. Subsequentlycompound A was
subjected to AMES test and positive result was ob-
tained, revealing its mutagenic liability. For these rea-
sons, IM491 was discontinued for further development.
could occur under strong basic conditions to produce
a,b-unsaturated hydroxamic acid and sulfinic acid.
Structure II (cis configuration preferred based on mode-
ling) was considered to be stable to base and as such, a
series of piperidine or pyrrolidine hydroxamic acids with
a sulfonyl c to the hydroxamic acid group were prepared
as shown in Table 1. All compounds prepared contain a
basic piperidine or pyrrolidine nitrogen in order to
maintain the good water solubilityand low protein
binding of IM491.
Given the attractive profile of IM491 in cellular activity,
TACE selectivityover MMPs and in vivo property, we
set out to address the mutagenic issue associated with
the aniline residue A byexploring the possibilityof
replacing the amide bond with other hydrogen-bond
acceptors. Modeling studies suggested that the amide
functionalitycould be replaced with a sulfonyl group
(Fig. 2) to provide either b-sulfonyl hydroxamic acids
(I) or c-sulfonyl hydroxamic acids (II). Structure I was
ruled out based on the consideration that b-elimination
We first prepared compounds 1a–b, which are the clos-
est analogs to IM491. With no substitution on the pip-
eridine nitrogen, compound 1a was more potent than
IM491 in porcine TACE (pTACE) assaywith similar
IM491
CONHOH
O
S
O
N
O
A
CONHOH
O
S
A
N
O
O
γ-sulfonyl hydroxamic acid (II)
β-sulfonyl hydroxamic acid (I)
A=piperidine or pyrrolidine
Figure 2.
Table 1. In vitro and in vivo profile of a,b-cyclic c-sulfonyl hydroxamic acid inhibitorsa–d
N
N
N
N
O
O
O
O
CONHOH
CONHOH
CONHOH
CONHOH
S
O
S
O
S
O
S
O
O
O
O
O
N
R
N
N
N
R
3a-c
R
R
4a-b
1a-b
2a-g
Compd
R
pTACE WBA
IC50 (nM) IC50 (nM) Ki (nM) Ki (nM) Ki (nM) Ki (nM)
MMP-1 MMP-2 MMP-9 MMP-13 Caco-2 Papp LPS-Mousee
(·10ꢀ6 cm/s)
Inhibition
@10mg/kg (po)
IM491
1a
1b
6.2
<1
1.1
1
20
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
2.1
0.1
>50% (ED50 =3–5 mg/kg)
H
240
>3000
250
190
230
160
180
900
1070
250
510
>1000
83
754
Me
H
>5000
>5000
>5000
>5000
503
2a
2b
0.1
0.1
0.1
0.97
1.3
1.6
1.8
0.1
0.2
0.65
0.1
1.0
Me
1.9
2c
2d
Isopropyl 1.5
Allyl 1.8
Propargyl 1.0
38%
25%
24%
35%
33%
2e
377
2f
2g
n-Butyl
Isobutyl
H
<1
<1
1.5
6.5
>5000
>5000
>5000
>5000
>5000
955
3a
3b
Me
3c
4a
Isopropyl 11
H
Me
<1
<1
20%
>50% (ED50 =2.5mg/kg)
4b
180
1708
a Compounds 1a–b and 4a–b are single enantiomers. Compounds 2a–g and 3a–c are racemates.
b See Ref. 4 for pTACE, WBA and MMP assayprotocols.
c pTACE IC50 and MMP Ki values are from single determination.
d Inhibition of TNF-a release in whole blood assay(WBA) was determined with three donors.
e See Ref. 5 for LPS-mouse model studies.

C.-B. Xue et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4453–4459
4455
selectivityprofile over MMPs ( Table 1). However it
exhibited 12-fold loss of cellular activityin WBA and
much lower Caco-2 permeabilitycompared with
IM491. While the lower Caco-2 permeabilityof 1a was
disappointing, the weaker WBA activitywas not a sur-
prise since the NH analog of IM491 was six-fold less
active than IM491 in WBA.1b We expected the cellular
activityof 1a to be improved bymethylation at the pip-
eridine nitrogen. Unfortunatelythe SAR in the IM491
series did not translate into this series. N-Methylation
of 1a resulted in a dramatic loss in WBA potency. Ana-
log 1b was inactive at a concentration of 3lM in WBA
despite its potent pTACE activity.
N
O
CONHOH
S
O
A
O
α,β-cyclic γ-sulfonyl hydroxamic acid (II)
N
O
CONHOH
S
O
A
O
We next investigated the regio-effect of the piperidine
nitrogen bymoving the nitrogen one atom awayfrom
the P10 group while keeping the 4-(2-methyl-4-quinol-
inylmethoxy)phenyl P10 residue constant. Without sub-
stitution on the piperidine nitrogen, analog 2a
(racemate) was equipotent to 1a in WBA with same
low Caco-2 permeability. Methylation (2b) or isopropyl-
ation (2c) at the piperidine nitrogen maintained the
WBA activitybut did not improve Caco-2 permeability.
The low Caco-2 permeabilityof 2c is consistent with the
low in vivo anti-TNF-a activityin an LPS-mouse model
in which an oral dose of 10mg/kg achieved only38%
inhibition of TNF-a release. Lowering the basicityof
the piperidine nitrogen byreplacing the methyl in 2b
with allyl (2d) or propargyl (2e) significantlyimproved
the Caco-2 permeabilitywhile retaining the cellular
activityof 2b. In spite of the improvement in Caco-2
permeability, both analogs were inactive in the LPS-
mouse model at an oral dose of 10mg/kg, indicating that
theywere not orallybioavailable. Further improvement
in Caco-2 permeabilitywas achieved byincreasing the
size of the substituent to n-butyl (2f) and isobutyl (2g).
However, these larger alkyl analogs suffered a consider-
able loss in cellular activitydespite their better pTACE
potency, probably due to increased protein binding
caused bythe larger alkyl groups.
β,β-cyclic γ-sulfonyl hydroxamic acid (III)
Figure 3.
than IM491, indicating that analog 4b is orallybioavail-
able in mice.
In parallel to the above SAR studies, we tried to modify
the core structure of II bymoving the a,b-cyclic ring to
the b,b-position to form a new series of b,b-cyclic c-sul-
fonyl hydroxamic acids (III, Fig. 3). This modification
not onlyretains the relative orientation of the hydroxa-
mate with respect to the sulfonyl (two key components
required for binding interaction with TACE) as sug-
gested bymodeling studies but also eliminates at least
the stereogenic center a to the hydroxamate in II. We
first investigated a piperidine heterocycle (Table 2).
Without substituent on the piperidine nitrogen, com-
pound 5a exhibited potent pTACE activity(IC 50 <1nM)
with high selectivityover MMP ꢀ1, ꢀ2, ꢀ9 and ꢀ13 but
weak WBA potency(IC 50 =632nM). Substitution on the
piperidine nitrogen with an alkyl is detrimental to the
pTACE activity; the larger the alkyl group, the weaker
the pTACE activity, suggesting that a substituent at this
position mayinterfere with the orientation of the hy-
droxamate relative to the sulfonyl group.
Moving the basic piperidine nitrogen one atom further
0
awayfrom the P1 residue in 2a provided analog 3a
(racemate), which was comparable to 2a in pTACE
activity, WBA potency, selectivity over MMPs and
Caco-2 permeability. Methylation (3b) or isopropylation
(3c) on the piperidine nitrogen in 3a attenuated the
pTACE and WBA activities.
In contrast to the 2,2-disubstituted piperidine series,
substitution on the piperidine nitrogen with different size
of alkyl groups in the 3,3- and 4,4-disubstituted piper-
idine series had little effect on pTACE potency. How-
ever, larger alkyl groups such as isopropyl and
isobutyl provided analogs with much weaker WBA
potency, probably due to higher protein binding. While
the NH analog 6a in the 3,3-disubstituted series and the
N-methyl analog 7b in the 4,4-disubstituted series dis-
played desired WBA activity (IC50 =102 and 123nM,
respectively), they exhibited low Caco-2 permeability,
consistent with their low oral activityin the LPS-mouse
model.
We next turned our SAR studies to smaller nitrogen-
containing heterocycles. Shrinking the six-membered
piperidine ring in 1a to five-membered pyrrolidine ring
provided analog 4a, which was three-fold more potent
than 1a in cellular activity, with an IC50 value of
80nM in WBA and highlyselective over MMPs. Unfor-
tunatelythe Caco-2 permeabilityof 4a remained low
and it was not orallyactive in the LPS-mouse model.
Methylation at the pyrrolidine nitrogen in 4a reduced
the WBA potencybytwo-fold but improved the Caco-
2 Papp value to 1.0·10ꢀ6 cm/s. Accompanied bythe
enhancement in cell permeation was the good oral
anti-TNF-a activityin the LPS-mouse model, with an
ED50 value of 2.5mg/kg, which is slightlymore active
We next investigated a pyrrolidine heterocycle (Table 2).
All pyrrolidine analogs prepared manifested same SAR
trend as observed in the piperidine series. Increment in
the size of a substituent on the pyrrolidine nitrogen in
the 2,2-disubstituted pyrrolidine series gradually de-
creased the pTACE potency( 8a–d) whereas anysize of

4456
C.-B. Xue et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4453–4459
Table 2. In vitro and in vivo profile of b,b-cyclic c-sulfonyl hydroxamic acid inhibitorsa,b
O
S
O
S
O
S
O
S
O
S
O
S
O
O
O
O
O
O
O
O
O
X
O
X
O
X
O
X
N
N
N
N
N
N
X
X
HOHNOC
HOHNOC
HOHNOC
HOHNOC
HOHNOC
HOHNOC
10a-b
9a-f
8a-e
6a-d
7a-d
5a-c
Compd
X
p-TACE
IC50 (nM)
WBA
IC50 (nM)
MMP-1
Ki (nM)
MMP-2
Ki (nM)
MMP-9
Ki (nM)
MMP-13
Ki (nM)
Caco-2 Papp
(·10ꢀ6 cm/s)
LPS-Mouse
% inhibition
@ 5mg/kg (po)
5a
5b
5c
6a
6b
6c
6d
7a
7b
7c
7d
8a
8b
8c
8d
8e
9a
9b
9c
9d
9e
9f
N–H
<1
5.1
48
2.8
3.6
2.7
3.3
14
2
632
>3000
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>4900
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
>3300
1400
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
>2100
2100
753
1.7
5.1
50
N–Me
N-i-Bu
N–H
>5000
>5000
>5000
3869
123
208
506
2503
0.1
0.2
0.2
0.7
0.1
0.2
31
49
N–Me
N-i-Pr
N-i-Bu
N–H
>5000
>5000
>5000
2613
25
57
N–Me
O
CH2
102
258
1
>5000
1206
<1
1
867
403
22
59
19
N–H
>3300
>3300
1333
>5000
3266
0.4
1.6
0.7
0.6
5.0
0.1
0.3
0.2
0.8
N–Me
N-i-Bu
N-Ac
O
2.9
20
660
1
>3000
>5000
>5000
>5000
3266
>3300
>3300
>3300
>3300
1333
913
76
57
34
28
37
22
22
N–H
1.5
3.2
1.6
2.6
3.4
0.8
1
N–Me
N-i-Pr
N-i-Bu
109
77
>5000
653
603
161
418
420
595
223
2167
>3300
1566
>2100
>2100
1511
N-neo-Pent
CH2
1758
804
18
10a
10b
O
CH2
>3300
2566
>2100
>2100
>5000
753
<1
4.9
86
a All compounds were synthesized in racemic form if a stereogenic center was involved.
b See footnotes b–e in Table 1.
substituent on the pyrrolidine nitrogen in the 3,3-disub-
stituted pyrrolidine series was tolerated for binding (9a–
e). Of particular note in the 3,3-disubstituted pyrrolidine
series are compounds 9a–c, which were verypotent
in the cellular assay, with IC50 values at or below
100nM. Unfortunatelythese analogs exhibited low
Caco-2 permeabilityand low oral bioavailabilityin the
LPS-mouse model.
tane derivative 10b, which showed decent WBA activity
(IC50 =223nM), high selectivityover MMPs and high
Caco-2 permeability. Consistent with its Caco-2 perme-
ability, 10b is orallyactive in the LPS-mouse model,
with 86% inhibition at an oral dose of 5mg/kg.
The good in vitro and in vivo profiles of 4b and 10b
prompted us to evaluate their pharmacokinetics in mice
in a discrete manner for 4b and in rats in n-in-1 for 10b
(Table 3). After iv administration, 4b exhibited high sys-
temic clearance (4.5L/h/kg), short half-life (0.5h) and
moderate volume of distribution (2.6L/kg). Despite the
high clearance, oral absorption was impressive, with
AUC value at 4612nMh at a dose of 20mg/kg and an
oral bioavailabilityof 46%, consistent with the observed
good oral anti-TNF-a activityin the LPS-mouse model.
In contrast to 4b, 10b exhibited much longer half life
To improve the Caco-2 permeabilityof this series, we
shifted our SAR studies to replacement of the piperidine
and pyrrolidine heterocycles with less polar oxygen-con-
taining heterocycles (7c, 8e, and 10a) or carbocycles (7d,
9f, and 10b). While significant improvement in Caco-2
permeabilitywas achieved, all analogs exhibited weak
to moderate WBA potencyalbeit potent pTACE activ-
ity(IC50 ꢀs61nM). Of particular interest is the cyclobu-
Table 3. Pharmacokinetics parameters of inhibitor 4b in mice and inhibitor 10b in rats
Compd
Dose (mg/kg)
CL (L/h/kg)
4.5
T1/2 (h)
Vss (L/kg)
2.6
Tmax (h)
Cmax (nM)
AUC (nMh)
F (%)
4b
iv 6.7
po 20
0.5
1.0
3367
4612
1
3511
46
10b
iv 5
po 5
6.6
7.2
6.9
1322
217
0.25
16

C.-B. Xue et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4453–4459
4457
Ph-4-OR
S
OR
CO2-t-Bu
Ph-4-OBn
S
SPh-4-OBn
O-t-Bu
N
CO2H
O
e
a
c
Cbz
Cbz
CO2-t-Bu
Bn
N
H
N
H
O
N
Li
Cbz
CO2-t-Bu
Cbz
CO2-t-Bu
N
H
N
H
5
O
6 R=H
7 R=MeSO2
b
8 R=H
9 R=Bn
d
10
Ph-4-OBn
Ph-4-OBn
HO
O
S
O
S
CO2-t-Bu
O
O
O
COR2
i
k, l
g, h
f
Cbz
N
CO2-t-Bu
Cbz
CO2-t-Bu
S
N
H
S
H
O O
N
R
O
O
N
R1
11
13 R=H
14 R=Boc
12
j
15 R1=H, R2=OH
1a R1=H, R2=NHOH
1b R1=Me, R2=NHOH
O
m
n, o
Scheme 1. Reagents and conditions: (a) BH3, THF, 60%; (b) MeSO2Cl, TEA, CH2Cl2, 97%; (c) 4-mercaptophenol, NaH, DMF, rt, 60%; (d) BnBr,
K2CO3, DMF, 95%; (e) LDA, allyl bromide, THF, 90%; (f) Oxone, MeOH, H2O, 90%; (g) 9-BBN, THF, 67%; (h) PDC, CH2Cl2, 71%; (i) H2, Pd–C,
MeOH, 95%; (j) (Boc)2O, THF, H2O, 90%; (k) 4-chloromethyl-2-methylquinoline, K2CO3, DMF, 65%; (l) TFA, CH2Cl2, 100%; (m) HONH2, BOP,
NMM, DMF, 75%; (n) HCHO, Na(OAc)3BH, NMM, THF, 100%; (o) HONH2 ÆHCl, BOP, NMM, DMF, 75%.
(7.2h) and higher volume of distribution (6.9L/kg) but
lower oral exposure, with an F value of 16%.
O
HO2C
HO2C
HO2C
HO2C
a, b
c
O
N
N
N
Boc
Boc
Compounds 1a–b were asymmetrically synthesized using
the protocol described in Scheme 1. Reduction of N-
Cbz-D-aspartic acid (5) using borane provided the alco-
hol 6 in 60% yield. The reaction was sluggish when
2equiv of borane were used and over 4equiv of borane
were required to complete the reaction. Following con-
version of the alcohol to a mesylate, displacement of
mesylate 7 with 2equiv of 4-mercaptophenol using
NaH (4equiv) produced the thioether 8 in 60% yield.
The phenol in 8 was alkylated by treating with benzyl
bromide in the presence of potassium carbonate. The
resulting benzyl ether 9 was subjected to an LDA alkyl-
ation with allyl bromide. Formation of the Li-coordi-
nated enolate intermediate followed byattack of allyl
bromide from the less hindered face gave rise to the de-
sired diastereomer 10 as the exclusive product, which
was isolated in 90% yield. Following oxidation of the
sulfide 10 to sulfone 11 using Oxone, the olefin in 11
was converted to an alcohol using 9-BBN and the result-
ing alcohol was oxidized to an aldehyde using pyrid-
inium dichromate (PDC). Upon cyclization by
hydrogenation using palladium on carbon as a catalyst,
the resulting piperidine nitrogen in 13 was blocked with
Boc bytreating with di- tert-butyl dicarbonate. Alkyl-
ation of phenol 14 with 4-chloromethyl-2-methylquinol-
ine followed byacid treatment afforded 15 from which
opticallypure 1a was obtained byBOP coupling with
hydroxylamine. Reductive amination of 15 with formal-
dehyde using sodium triacetoxyborohydride followed by
BOP coupling with hydroxylamine completed the syn-
thesis of 1b.
18
17
O
16
O
HO
HO
d
e
HO
HO
+
N
N
Boc
Boc
19a
19b
O
O
O
+
O
3a-c
N
N
Boc
Boc
20b
O
20a
f, g, h
N
O
CO2Me
2a-g
S
N
Boc
21
Scheme 2. Reagents and conditions: (a) H2, PtO2, 1N HCl; (b)
(Boc)2O, NaOH, 74% for two steps; (c) Ac2O, THF, 100%; (d) NaBH4,
THF; (e) MeI, K2CO3, DMF, rt, 44% for 20a, 46% for 20b; (f)
4-mercaptophenol (2equiv), NaH (4equiv), DMF, 80ꢁC, 3h, 76%; (g)
MeI, K2CO3, DMF; (h) 4-chloromethyl-2-methylquinoline, K2CO3,
DMF, 80ꢁC, 59% for two steps.
treatment with acetic anhydride to form the cyclic anhy-
dride 18, which was reduced with sodium borohydride
to yield two regio-isomers of c-hydroxycarboxylic acid
19a and 19b. Without separation, the mixture was sub-
jected to treatment with iodomethane in the presence
of potassium carbonate to generate two lactones, which
were separated bycolumn chromatographyto give 20a
(racemate) in 44% yield and 20b (racemate) in 46% yield.
Both structures were characterized using 2 D NMR.
Ring opening of 20a with 4-mercaptophenol (2equiv)
The 4,5-disubstituted piperidine derivatives 2a–g were
prepared as racemates using the procedures outlined in
Scheme 2. Reduction of 3,4-pyridinedicarboxylic acid
(16) byhydrogenation using platinum oxide as a catalyst
followed bytreatment with di- tert-butyl dicarbonate
provided N-Boc-cis-3,4-piperidinedicarboxylic acid 17
in 74% yield. This dicarboxylic acid was subjected to a

4458
C.-B. Xue et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4453–4459
in the presence of NaH (4equiv), esterification of the
resulting carboxylic acid with iodomethane and alkyl-
ation of the phenol with 4-chloromethyl-2-methylquinol-
ine yielded the intermediate 21, which was converted to
the final products 2a–g following the procedures
described in Scheme 1. Starting from intermediate 20b,
compounds 3a–c were obtained in a similar manner
(Scheme 2).
EtO2C
BocN
c
a
e
O
CHR
BocN
BocN
R
32
33 R=CO2Me
34 R=CH2OH
35 R=CH2
36 R=O
b
d
EtO2C
EtO2C
BocN
g
9a-e
S
BocN
R
Syntheses of compounds 4a–b were shown in Scheme 3.
Intermediate 11 was subjected to an ozonolysis to pro-
vide the aldehyde 22 in 36% yield. The low yield of the
ozonolysis was probably caused by the interference of
the Cbz-protected amino group that, among other
things, could attack the generated aldehyde to form a
semiaminal. Ring closure to form a pyrrolidine from
aldehyde 22 was effected byhydrogenation using palla-
dium on carbon as catalyst. After protection of the
resulting pyrrolidine 23 with Boc, the intermediate 24
was converted to the final products 4a–b following the
procedures depicted in Scheme 1.
37 R=OH
OH
f
39
38 R=OMs
Scheme 5. Reagents and conditions: (a) Methyl (triphenylphosphor-
anylidene)acetate, toluene, 110ꢁC; (b) DIBAL, THF, ꢀ78ꢁC; (c)
CH3C(OEt)3, propionic acid, 180ꢁC; (d) O3, CH2Cl2, ꢀ78ꢁC, then
P(OMe)3, rt; (e) NaBH4, EtOH, rt; (f) MsCl, DIEA, CH2Cl2, 0ꢁC; (g)
NaH, 4-mercaptophenol, DMF, 0ꢁC.
allylation product 27. DIBAL reduction of 27 gave rise
to the alcohol 28, which was converted to the hemiacetal
29 upon ozonolysis. Following oxidation of the hemiac-
etal to a lactone under PDC conditions, the lactone ring
in 30 was selectivelyopened with 4-mercaptophenol and
the resulting carboxylic acid was esterified using
TMSCHN2. Conversion of ester 31 to the final product
7a–b was accomplished following the procedures
described in Scheme 1.
Most of the b,b-cyclic c-sulfonyl hydroxamic acids were
synthesized using a sequence as exemplified in Scheme 4.
The commerciallyavailable carboxylic acid 25 was con-
verted to a methyl ester 26 under K2CO3 and MeI con-
ditions. Treatment of 26 with LDA in THF at ꢀ78ꢁC,
followed byaddition of allly bromide, furnished the
A different synthetic route to compounds 9a–e was
developed due to the commercial unavailabilityof the
required carboxylic acid. As illustrated in Scheme 5,
commerciallyavailable ketone 32 was reacted with
methyl (triphenylphosphoranylidene)acetate, to provide
the unsaturated esters 33 as a mixture of cis and trans
isomers (1:1 ratio). DIBAL reduction of this mixture
led to a mixture of allylic alcohol 34, which set the stage
for the crucial Claisen rearrangement. A mixture of 34
and ethyl orthoacetate was heated up in the presence
of propionic acid in a sealed tube overnight and only
one c,d-unsaturated ester 35 was obtained. Ozonolysis
Ph-4-OBn
Ph-4-OBn
O
S
O
S
O
O
b
a
Cbz
CO2-t-Bu
Cbz
HO
CO2-t-Bu
N
H
N
H
22
11
O
CO2-t-Bu
4a-b
S
O
O
N
of 35, followed byreduction with NaBH , gave rise to
4
R
the alcohol 37, which was mesylated and displaced with
4-mercaptophenol to afford 39. Following the similar
steps as illustrated in Scheme 1, 39 was transformed to
the final products 9a–e.
23 R=H
c
24 R=Boc
Scheme 3. Reagents and conditions: (a) O3, P(OMe)3, CH2Cl2, 36%;
(b) H2, Pd–C, MeOH, 99%; (c) (Boc)2O, NaHCO3, H2O, THF, 97%.
In summary, replacement of the amide functionality in
IM491 with a sulfonyl group led to a new series of
a,b-cyclic and b,b-cyclic c-sulfonyl hydroxamic acids.
While most of the sulfone analogs prepared had higher
affinityfor pTACE than IM491, theywere less potent
in the inhibition of TNF-a production in WBA proba-
blydue to higher protein binding and/or poorer cell pen-
etration of the sulfone molecules. The best analog
identified from this series is the five-membered pyrrol-
idine analog 4b that exhibited IC50 values of <1 and
180nM in the pTACE and cell assays, respectively, high
selectivityover MMP ꢀ1, ꢀ2, ꢀ9 and ꢀ13, decent Caco-
2 permeabilityand good oral anti-TNF- a activityin the
LPS-mouse model. Pharmacokinetics studydemon-
strated that 4b was orallybioavailable, with an F value
of 46% in mice.
b
d
O
R1
R2
BocN
a
CO2R
BocN
R
BocN
29 R1=H, R2=OH
30 R1, R2=O
25 R=H
27 R=CO2Me
28 R=CH2OH
c
e
26 R=Me
S
OH
f, g
BocN
7a-b
CO2Me
31
Scheme 4. Reagents and conditions: (a) K2CO3, MeI, rt; (b) LDA,
THF, ꢀ78ꢁC, then allyl bromide; (c) DIBAL, CH2Cl2, ꢀ78ꢁC; (d) O3,
CH2Cl2, ꢀ78ꢁC, then P(OMe)3, rt; (e) PDC, CH2Cl2, rt; (f) NaH,
4-mercaptophenol, DMF, 140ꢁC; (g) TMSCHN2, MeOH, rt.

C.-B. Xue et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4453–4459
4459
Conway, M.; Cowan, D.; Gaul, M.; Glennon, K.; Lambert,
M.; Leesnitzer, M.; McDougald, D.; Moss, M.; Musso, D.;
Rizzolio, M. J. Med. Chem. 2001, 44, 4252; (c) Holms, J.;
Mast, K.; Marcotte, P.; Elmore, I.; Li, J.; Pease, L.; Glaser,
K.; Morgan, D.; Michaelides, M.; Davidsen, S. Bioorg.
Med. Chem. Lett. 2001, 11, 2907; Chen, J.; Jin, G.; Sung,
A.; Levin, J. Bioorg. Med. Chem. Lett. 2002, 12, 1195; (d)
Levin, J.; Chen, J.; Du, M.; Nelson, F.; Killar, L.; Skala, S.;
Sung, A.; Jin, G.; Cowling, R.; Barone, D.; March, C.;
Mohler, K.; Black, R.; Skotnicki, J. Bioorg. Med. Chem.
Lett. 2002, 12, 1199; (e) Letavic, M.; Axt, M.; Barberia, J.;
Carty, T.; Danley, D.; Geoghegan, K.; Halim, N.; Hoth, L.;
Kamath, A.; Laird, E.; Lopresti-Morrow, L.; McClure, K.;
Mitchell, P.; Natarajan, V.; Noe, M.; Pandit, J.; Reeves, L.;
Schulte, G.; Snow, S.; Sweeney, F.; Tan, D.; Yu, C. Bioorg.
Med. Chem. Lett. 2002, 12, 1387; (f) Duan, J.; Chen, L.;
Wasserman, Z.; Lu, Z.; Liu, R.-Q.; Covington, M.; Qian,
M.; Hardman, K.; Magolda, R.; Newton, R.; Christ, D.;
Wexler, R.; Decicco, C. J. Med. Chem. 2002, 45, 4954; (g)
Duan, J.; Lu, Z.; Xue, C.-B.; He, X.; Seng, J.; Roderick, J.;
Zelda, W.; Liu, R.-Q.; Covington, M.; Magolda, R.;
Newton, R.; Trzaskos, J.; Decicco, C. Bioorg. Med. Chem.
Lett. 2003, 13, 2035; (h) Wasserman, Z.; Duan, J.; Voss,
M.; Xue, C.-B.; Cherney, R.; Nelson, D.; Hardman, K.;
Decicco, C. Chem. Biol. 2003, 10, 215.
Acknowledgements
We thank John Giannaras, Sherrill A. Nurnberg, Dianna
L. Blessington, Persymphonie B. Miller, Theresa M.
Dimeo, Robert J. Collins, TracyL. Taylor and Paul J.
Strzemienski for performing in vitro and in vivo
assays.
References and notes
1. (a) Xue, C.-B.; He, X.; Roderick, J.; Corbett, R.; Duan, J.;
Liu, R.-Q.; Covington, M.; Newton, R.; Magolda, R.;
Wexler, R.; Decicco, C. Bioorg. Med. Chem. Lett. 2003,
13, 4293; (b) Xue, C.-B.; He, X.; Roderick, J.; Corbett, R.;
Duan, J.; Liu, R.-Q.; Covington, M.; Qian, M.; Ribadene-
ira, M.; Vaddi, K.; Christ, D.; Newton, R.; Wexler, R.;
Decicco, C. Bioorg. Med. Chem. Lett. 2003, 13, 4299.
2. (a) Black, R.; Rauch, C.; Kozlosky, C.; Peschon, J.; Slack,
J.; Wolfson, M.; Castner, B.; Stocking, K.; Reddy, P.;
Srinivasan, S.; Nelson, N.; Bolani, N.; Schooley, K.;
Gerhart, M.; Devis, R.; Fitzner, J.; Johnson, R.; Paxton,
R.; March, C.; Cerretti, D. Nature 1997, 385, 729; (b) Moss,
M.; Jin, S.-L.; Milla, M.; Burkhart, W.; Carter, H.; Chen,
W.; Clay, W.; Didsbury, J.; Hassler, D.; Hoffman, C.; Kost,
T.; Lambert, M.; Leesnitzer, M.; McCauley, P.; McGeehan,
G.; Mitchell, J.; Moyer, M.; Pahel, G.; Rocque, W.;
Overton, L.; Schoenen, F.; Seaton, T.; Su, J.; Warner, J.;
Willard, D.; Becherer, J. Nature 1997, 385, 733.
4. Xue, C.-B.; Voss, M.; Nelson, D.; Duan, J.; Cherney, R.;
Jacobson, I.; He, X.; Roderick, J.; Chen, L.; Corbett, R.;
Wang, L.; Meyer, D.; Kennedy, K.; DeGrado, W.; Hard-
man, K.; Teleha, C.; Jaffee, B.; Liu, R.-Q.; Copeland, R.;
Covington, M.; Christ, D.; Trzaskos, J.; Newton, R.;
Magolda, R.; Wexler, R.; Decicco, C. J. Med. Chem. 2001,
44, 2636.
5. Newton, R.; Solomon, K.; Covington, M.; Decicco, C.;
Haley, P.; Friedman, S.; Vaddi, K. Ann. Rheum. Dis. 2001,
60(Suppl 3), iii 25–32.
3. (a) Xue, C.-B.; He, X.; Corbett, R.; Roderick, J.; Wasser-
man, Z.; Liu, R.-Q.; Jaffee, B.; Covington, M.; Qian, M.;
Trzaskos, J.; Newton, R.; Magolda, R.; Wexler, R.;
Decicco, C. J. Med. Chem. 2001, 44, 3351; (b) Rabinowitz,
M.; Andrews, R.; Becherer, J.; Bickett, D.; Bubacz, D.;