Discovery of potent and selective inhibitors of 11β-HSD1 for the treatment of metabolic syndrome

Article abstract of DOI:10.1016/j.bmcl.2006.09.035

High throughput screening efforts have identified a novel class of dichloroaniline amide 11β-HSD1 inhibitors. SAR studies initiated from dichloroaniline 4 focused on retaining the potency and selectivity profile of the lead.

Full text of DOI:10.1016/j.bmcl.2006.09.035

Bioorganic & Medicinal Chemistry Letters 16 (2006) 6241–6245
Discovery of potent and selective inhibitors of 11b-HSD1 for the
treatment of metabolic syndrome
Steven Richards,* Bryan Sorensen, Hwan-soo Jae, Marty Winn, Yixian Chen,
Jiahong Wang, Steven Fung, Katina Monzon, Ernst U. Frevert, Peer Jacobson,
Hing Sham and J. T. Link
Metabolic Disease Research, GPRD, Abbott Laboratories, 100 Abbott Park Rd., Abbott Park, IL 60064, USA
Received 8 August 2006; revised 4 September 2006; accepted 7 September 2006
Available online 26 September 2006
Abstract—High throughput screening efforts have identified a novel class of dichloroaniline amide 11b-HSD1 inhibitors. SAR
studies initiated from dichloroaniline 4 focused on retaining the potency and selectivity profile of the lead.
Ó 2006 Elsevier Ltd. All rights reserved.
The glucocorticoid hormones, cortisol in human and
corticosterone in rodents, play key roles in lipid and glu-
cose metabolism. Patients with glucocorticoid excess
(Cushing’s disease) are typically obese, insulin resistant,
and have high fasting blood glucose levels. However, in
diabetic patients, circulating cortisol levels are not ele-
vated.1 To determine if there is a link between glucocor-
ticoid concentrations and human diabetes, researchers
focused on the reductase 11b-hydroxysteroid dehydro-
genase (11b-HSD1). This enzyme converts inactive
cortisone to cortisol in tissues such as liver and fat,
increasing the local concentration of active glucocorti-
coid. Evidence of the importance of inhibition of 11b-
HSD1 came from homozygous 11b-HSD1 knockout
mice as well as animals engineered to overexpress 11b-
HSD1.2 The knockout animals showed decreased fast-
ing blood glucose levels and decreased glycogenolysis.
Selective over-expression of 11b-HSD1 in mouse adi-
pose tissue increased visceral fat mass, when fed a high
fat diet, and the animals were hyperglycemic and insulin
resistant.3
of 11b-HSD1 (2 and 3) that showed efficacy in rodents.
The Biovitrum inhibitor, BVT.2733, when dosed via
mini-pump, lowered glucose and insulin levels. Addi-
tionally, continuous infusion of 2 reduced expression
levels of mRNA encoding for PEPCK and G6Pase.5
The short-acting Merck inhibitor 3 was reported to low-
er glucose levels, decrease food intake, improve insulin
sensitivity, and decrease the rate of body weight gain
in STZ treated DIO mice. In ApoE knockout mice, tri-
azole 3 was shown to decrease the rate of atherosclerotic
plaque formation.6 In light of the preliminary validation
of 11b-HSD1 as a treatment for the metabolic syn-
drome, we present a novel series of potent, selective
dichloroaniline 11b-HSD1 inhibitors.
Carbenoxolone (1), a non-selective 11b-HSD1 inhibitor,
has been shown to decrease glycogenolysis in diabetic
patients (Fig. 1).4 However, carbenoxolone also caused
a rise in blood pressure, likely due to inhibition of
11b-HSD2.4 Researchers at Biovitrum and Merck have
reported the preparation of small molecule inhibitors
Keywords: 11b-HSD1 inhibitor; 11b-hydroxysteroid dehydrogenase.
*
Figure 1. Representative 11b-HSD1 inhibitors.
Corresponding author. E-mail: Steven.Richards@abbott.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.09.035

6242
S. Richards et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6241–6245
Scheme 1. Reagents and conditions: (a) amines, TBTU, DMF, 100 °C,
12 h, 70%.
Dichloroaniline amides (6) were initially prepared as
shown in Scheme 1 where 4-amino-3,5-dichlorobenzoic
acid was coupled with a variety of secondary amines.
Scheme 4. Reagents and conditions: (a) benzyl chloroformate, MeM-
gBr, THF, À25 °C, 2 h, 1 M HCl, rt, 1 h, 50%; (b) Zn, AcOH, 80 °C,
12 h, 80% or MeMgBr, THF, CuBr DMS, BF3ÆOEt2, À78 °C, 2.5 h,
60%; (c) H2, Pd/C, MeOH, 1 atm, 12 h; (d) TBTU, DMF, 4-amino-3,5-
dichlorobenzoic acid, 100 °C, 12 h, 40%; (e) anilines, Na(OAc)3BH,
AcOH, DCE, rt, 12 h, 20–50%; (f) triphosgene, Et3N, CH2Cl2, 0 °C,
2 h, 10–30%.
Elaboration of amide 6 began with hydrolysis, which
was accomplished under standard conditions, to provide
the carboxylic acid (Scheme 2). Coupling of the acid
with amines provided a second group of dichloroaniline
11b-HSD1 inhibitors.
Another series of compounds was prepared by the route
outlined in Scheme 3. The least hindered amine in cis-
2,6-dimethyl piperazine was selectively masked as a
tert-butoxy carbamate.7 Coupling with 4-amino-3,5-
dichlorobenzoic acid furnished the desired amide
product. Cleavage of the protecting group under acidic
conditions provided a crystalline amine hydrochloride
core. Analogs were prepared by standard nucleophilic
substitution reactions. Alternatively, palladium cata-
lyzed cross-coupling of 2,6-dimethyl piperazine with aryl
halides provided substituted piperazines that were, in
turn, coupled with 4-amino-3,5-dichlorobenzoic acid.8
Such compounds were prepared as shown in Scheme 4
using chemistry pioneered by the Comins group.9 By
sequential addition of benzyl chloroformate, MeMgBr
and aqueous acid to 4-methoxypyridine 10, enone 11
was efficiently produced. Enone 11 was then processed
in two different ways. Reduction with zinc was followed
by hydrogenolysis of the amine protecting group.10 The
resulting amine was coupled with 4-amino-3,5-dichloro
benzoic acid. Reductive amination with 2-aminophenol
or 1,2-diaminobenzene provided monomethyl piperi-
dines 12 (R = H). Imine reduction favored the syn iso-
mer with a 10:1 selectivity ratio. Subsequent reaction
with triphosgene afforded either the cyclic carbamate
or urea (13, R = H). Alternatively, enone 11 was treated
with methyl cuprate. The 2,6-dimethyl ketone was
formed in a 3:1 trans/cis ratio. Subsequent transforma-
tions to analogs of 13 (R = Me) proceeded as described
above with an additional reductive amination carried
out using 1,2-diamino benzene.
To explore the steric environment of the 11b-HSD1 en-
zyme, we investigated 2,6-dimethyl-4-amino piperidines.
The dichloroaniline amides were evaluated in 11b-HSD1
binding assays using truncated human and rat enzymes.
The cellular assay for 11b-HSD1 activity was performed
in HEK293 cells transfected with full length h-HSD1
cDNA.11
Scheme 2. Reagents and conditions: (a) LiOH, THF/H2O, 80 °C, 12 h,
90%; (b) R1R2NH, TBTU, DMF, 100 °C, 12 h, 80%.
High throughput screening efforts led to the identifica-
tion of dichloroaniline amide 4 that was a potent inhib-
itor of both the human and rat isoforms of 11b-HSD1
and was highly selective against 11b-HSD2 (Table 1).12
In rat, dichloroaniline 4 was shown to be a short acting
inhibitor of 11b-HSD1. When compound 4 was tested
for metabolic stability in rat microsomes, it was rapidly
metabolized. Subsequent metabolite identification
experiments demonstrated that the major mode of
metabolism was hydroxylation of the cyclopentyl ring.
No aniline oxidation was observed. Lead 4 also had
poor water solubility. Thus, we began a research effort
to identify additional potent, selective dichloroaniline
inhibitors of 11b-HSD1.
Scheme 3. Reagents and conditions: (a) (Boc)2O, THF, rt, 12 h, 80%;
(b) TBTU, DMF, 4-amino-3,5-dichlorobenzoic acid, 100 °C, 12 h,
40%; (c) 4 M HCl in dioxane, THF, rt, 3 h, 90%; (d) RX, DMF,
Cs2CO3, 180 °C, 60 min, 10–50%; (e) Pd2(dba)3, ArBr, NaOtBu,
BINAP, toluene, 150 °C, 15 min, 30–50%.

S. Richards et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6241–6245
Table 1. First generation dichloroaniline 11bHSD1 inhibitors (6)
6243
In an effort to lower the lipophilicity of this series, we
attempted to introduce polar functionality. Compounds
6d and 6e indicated human 11b-HSD1 could tolerate a
basic nitrogen. Unfortunately, the amine had a deleteri-
ous effect on cellular activity. The cyclopropyl cyclohex-
yl amide 6f was as potent as lead compound 4.
Functionalization of the cyclohexyl ring at the 4 posi-
tion with a carboethoxy group led to compound 6g. This
ester retained potency against rat 11b-HSD1 while suf-
fering a drop in cellular potency. Analogs of 6g were
pursued in hopes of recovering cellular activity.
Compound R
h-HSD1 r-HSD1 HEK HSD1
a
Kia (nM) Kia (nM) IC50 (nM)
N
4
8
4
16
140
In general, lipophilic secondary amides were preferred
by rat 11b-HSD1 (Table 2). Compounds 7a and 7b were
representative examples of a group of compounds that
contained small, non-polar groups. These potent com-
N
6a
8
17
Table 2. Dichloroaniline amides 11bHSD1 inhibitor 7
N
6b
6c
10
6
100
41
N
4
4
Compound R1
R2
H
h-HSD1 r-HSD1 HEK HSD1
a
Kia (nM) Kia (nM) IC50 (nM)
N
N
6d
6e
16
53
2700
280
960
N
7a
17
39
640
N
7b
7c
7d
H
20
60
68
14
20
330
210
O
O
Me
H
5700
O
N
6f
14
5.3
130
700
>30,000
OH
N
6g
28
4
900
7e
7f
=R1
=R1
130
560
5.2
6400
CO2Et
a Values are geometric means of two experiments.
9.0
>30,000
N
Polycyclic amides 6a–c were potent against human and
rat 11b-HSD1. They are also selective against both hu-
man and rat 11b-HSD2. Cellular potency of compounds
6a and 6c was significantly improved over the lead
compound.13
7g
H
1100
17
>30,000
N
a Values are geometric means of two experiments.

6244
S. Richards et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6241–6245
Table 3. 2,6-Dimethyl piperazine dichloroaniline analogs with general
structure 9
Compound
R
h-HSD1
r-HSD1
HEK HSD1
a
Kia (nM) Kia (nM) IC50 (nM)
N
9a
100
13
6
>10,000
Scheme 5. Steric bulk around amide increases h-HSD1 potency.
9b
9c
14
490
O
O
O
pounds did not show improved cellular potency. The
move to small, polar substituents, such as ester 7c, also
did not improve cellular potency. Similarly, tertiary
amide 7e was also significantly less active in the HEK as-
say. The incorporation of acidic (7d) or basic (7f and 7g)
substituents at this position also decreased the human
11b-HSD1 potency and led to a complete loss of cellular
activity. However, these compounds were potent against
r-HSD1, indicating that charged substituents were pre-
ferred at this position in the rat isoform.
10
18
190
O
O
9d
9e
19
15
23
360
370
N
The decision to pursue substituted piperazine and amino
piperidine dichloroaniline analogs was made based on
the compounds shown in Scheme 5. Cyclohexyl amide
14 was found to be a weak inhibitor of human 11b-
HSD1. However, when methyl groups were added as
steric bulk around the amide carbonyl, the resulting
amide 15 was greater than 20-fold more potent. Similar-
ly, 2,6-dimethyl aryl piperazine 17 was marginally active
while the non-methylated analog 16 was completely
inactive. Thus, a goal became to exploit these observa-
tions to make polar analogs that maintained cellular
potency.
NH2
5
N
a Values are geometric means of two experiments.
r-HSD1, was inactive in cells. Alkyl and aryl piperazines
9b and 9c provided the first indication that this scaffold
could possess cellular potency. Introduction of both a
nitrogen and a second substituent at the ortho posi-
tions of the aryl piperazine core also furnished active
analogs. Meta- and para-substituted aryl piperazines
were inactive or less active than their ortho-substituted
comparators. Compounds 9d and 9e were examples that
displayed polar functional groups and still had cellular
Incorporation of the cis-2,6-dimethyl piperazine moiety
to the dichloroaniline scaffold was generally tolerated
by rat 11b-HSD1. The potency depended on the substit-
uents appended to the 4 position of the piperazine
(Table 3). Pyridyl piperazine 9a, while potent against
Table 4. 4-Aminopiperidine dichloroaniline SAR
a
HEK HSD1 IC50 (nM)
Compound
R
X
h-HSD1 Kia (nM)
r-HSD1 Kia (nM)
13a
13b
13c
H
O
O
N
13
12
22
140
160
39
380
72
Me
Me
76
a Values are geometric means of two experiments.

S. Richards et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6241–6245
6245
B.; Li, Z.; Metzger, J. M.; Mundt, S. S.; Noonan, H.;
Nunes, C. N.; Olson, S. H.; Pikounis, B.; Ren, N.;
Robertson, N.; Schaeffer, J. M.; Shah, K.; Springer, M.
S.; Strack, A. M.; Strowski, M.; Wu, K.; Wu, T.; Xiao,
J.; Zhang, B. B.; Wright, S. D.; Thieringer, R. J. Exp.
Med. 2005, 202, 517.
activity. Additionally, 9e was found to be more than 10-
fold more soluble than lead compound 4.
The highly substituted trans piperidine amides were also
well tolerated by h-HSD1 (Table 4). Dimethyl urethane
13b was significantly more active against 11b-HSD1 in
cells than the monomethyl variant (13a). Urea 13c
proved to be a dual potent human/rat 11b-HSD1 inhib-
itor that maintained cellular potency.
7. Jacobsen, E. J.; Stelzer, L. S.; TenBrink, R. E.; Belonga,
K. L.; Carter, D. B.; Im, H. K.; Im, W. B.; Sethy, V.
H.; Tang, A. H.; VonVoightlander, P. F.; Petke, J. D.;
Zhong, W.; Mickelson, J. W. J. Med. Chem. 1999, 42,
1123.
8. Compound 9e was prepared via a Curtius rearrangement
of the acid derived from ester 9d.
9. Brown, J. D.; Foley, M. A.; Comins, D. L. J. Am. Chem.
Soc. 1988, 110, 7445.
10. Comins, D. L.; Brooks, C. A.; Ingalls, C. L. J. Org. Chem.
2001, 66, 2181.
In summary, we have described SAR for a novel series
of dichloroaniline amide 11b-HSD1 inhibitors and iden-
tified several dual human/rat potent compounds. The rat
isoform tended to accommodate polar analogs better
than the human enzyme. While the SAR for cellular
potency was less well defined, we were able to discover
several dual potent inhibitors, 6a, 6c, and 13c, which
were also potent against 11b-HSD1 in cells. All of the
compounds identified were selective against 11b-
HSD2. The discovery of compounds with this desirable
biochemical profile is a key step forward in the identifi-
cation of new small molecule inhibitors of 11b-HSD1
with therapeutic potential.
11. The human and mouse HSD1 assays were performed in a
SPA format using truncated enzyme, expressed in Esch-
erichia coli. The HSD2 SPA assay used full-length human
and mouse enzyme expressed in baculovirus. The test
compounds were incubated with enzyme and 3H-cortisone
substrate for 30 min at room temperature. The radioactive
cortisol produced was captured and quantified. The
cellular assay for 11b-HSD1 activity was performed in
HEK293 cells transfected with full length h-HSD1 cDNA.
Test compounds were pre-incubated for 30 min with the
cells before introduction of the cortisone substrate. After
2 h incubation, the cortisol concentration was determined
by fluorescence polarization immuno-assay (FPIA). For
more details, see Rohde, J. J.; Pliushchev, M. A.;
Sorensen, B. K.; Wodka, D.; Shuai,Q.; Wang, J.; Fung,
S.; Monzon, K. M.; Chiou, W. J.; Pan, L.; Deng, X.;
Chovan, L. E.; Ramaiya, A.; Mullally, M.; Henry, R. F.;
Stolarik, D. F.; Imade, H. M.; Marsh, K. C.; Beno, D.;
Fey, T. A.; Droz, B. A.; McDowell, C.; Brune, M. E.;
Camp, H. S.; Sham, H. L.; Frevert, E. U.; Jacobson, P. B.;
Link, J. T. J. Med. Chem., submitted for publication.
12. All of the compounds described in this paper were assayed
against both human and rat 11b-HSD2 and found to be
inactive (IC50 > 10 lM).
References and notes
1. For a review, see: Seckl, J.; Walker, B. R. Trends
Endocrinol. Metab. 2004, 15, 418.
2. Kotelevtsev, Y.; Holms, M. C.; Burchell, A.; Houston, P.
M.; Schmoll, D.; Jamieson, P.; Best, R.; Brown, R.;
Edwards, C. R. W.; Seckl, J. R. Proc. Natl. Acad. Sci.
U.S.A. 1997, 94, 14924.
3. Masuzaki, H.; Peterson, J.; Shinyama, H.; Morton, N. M.;
Mullins, J. J.; Seckl, J. R.; Flier, J. S. Science 2001, 294,
2166.
4. Andrews, R. C.; Rooyackers, O.; Walker, B. R. J. Clin.
Endocrinol. Metab. 2003, 88, 285.
13. A strategic decision was made to move away from amides
such as 6a due to the fact compounds containing the same
lipophilic amine substituent began appearing in the patent
literature. See Kampen, Gita, Camilla, Tejlgaard, Ander-
sen, Henrik, Sune, WO 2004/089415, 2004.
5. Alberts, P.; Engblom, L.; Edling, N.; Forsgren, M.;
Klingstro¨m, G.; Larsson, C.; Ro¨nquist-Nii, Y. B.; Abra-
´
hmsen, L. Diabetologia 2002, 45, 1528.
6. Hermanowski-Vosatka, A.; Balkovec, J. M.; Cheng, K.;
Chen, H. Y.; Hernandez, M.; Koo, G. C.; Le Grand, C.