Synthesis and structure-activity relationship of RXR antagonists based on the diazepinylbenzoic acid structure

Article abstract of DOI:10.1016/j.bmcl.2007.06.079

Synthesis and structure-activity relationship of RXR antagonists employing a diazepinylbenzoic acid scaffold are described. Of those antagonists, sulfonamide derivatives (6v and 6w) reveal a high antagonistic activity and good pharmacokinetic properties.

Full text of DOI:10.1016/j.bmcl.2007.06.079

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4808–4811
Synthesis and structure–activity relationship of RXR
antagonists based on the diazepinylbenzoic acid structure
Junichi Sakaki, Kazuhide Konishi, Masashi Kishida,* Hiroki Gunji, Takanori Kanazawa,
Hidefumi Uchiyama, Hiroaki Fukaya, Hironobu Mitani and Masaaki Kimura
Novartis Institute for Biomedical Research, Tsukuba Research Institute, Ohkubo 8, Tsukuba-shi, Ibaraki 300-2611, Japan
Received 4 April 2007; revised 15 June 2007; accepted 18 June 2007
Available online 30 June 2007
Abstract—Synthesis and structure–activity relationship of RXR antagonists employing a diazepinylbenzoic acid scaffold are
described. Of those antagonists, sulfonamide derivatives (6v and 6w) reveal a high antagonistic activity and good pharmacokinetic
properties.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Retinoid X receptors (RXRs) are members of the nucle-
N
ar receptor superfamily and act as ligand-inducible tran-
scriptional factors.1,2 RXRs regulate a wide variety of
biological functions such as cell differentiation, prolifer-
ation, and embryonic development in vertebrates.3
RXRs are divided into three sub-types a, b, c and form
a heterodimer with various other nuclear receptors (e.g.,
peroxisome proliferator-activated receptors: PPARs, li-
ver X receptors: LXRs, farnesoid X receptor: FXR
and retinoic acid receptors: RARs).4
NO2
N
HO
Figure 1. Structure of HX531.
O
HX531 (Fig. 1) was identified as an RXR antagonist on
the basis of inhibitory activity on retinoid-induced cell
differentiation of human promyelocytic cells HL-60
and transactivation assays using RARs and RXRs in
COS-1 cells by Kagechika et al.5 HX531 inhibited HL-
60 cell differentiation induced by the combination of
the retinoid agonist Am80 with an RXR agonist (a ret-
inoid synergist, HX600), indicating that HX531 inhib-
ited the activation of RAR–RXR heterodimers as well
as homodimers. Kadowaki et al. reported that treatment
of KKAy mice fed a high-fat diet with HX531 reduced
plasma glucose and insulin levels, improved insulin sen-
sitivity, and increased energy expenditure.
It also decreased body weight gain though no effect
on food consumption was observed.6 It suggested
Keywords: RXR; Antagonist; Diazepinylbenzoic acid.
*
Figure 2. Docking of HX531 in a homology model of the antagonist-
bound form of RXRa.
Corresponding author. Tel.: +81 29 865 2306; fax: +81 29 865
2308; e-mail: masashi.kishida@novartis.com
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.06.079

J. Sakaki et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4808–4811
4809
N
R 2
N
H
N
O
O
N
OH
3a-c
R3
N
H
N
b,c
d,c
O
R1
N
R1
OH
4a-m
N
NO2
NH2
a
N
N
O
R4
N
N
e,c
N
H
H
O
O
N
O
O
1
2
O
f,c
OH
5a-e
R1
N
O
S
O
R5
N
H
N
O
OH
6a-z
Scheme 1. Reagents and conditions: (a) SnCl2–2H2O, DMF, 60 ꢁC; (b) i—(Boc)2O, Et3N, THF; ii—R2X, NaH, DMF; iii—TFA, CH2Cl2; (c) 2 M
NaOH, DMF; (d) R3COX, Et3N, CH2Cl2; (e) R4NCO, Et3N, CH2Cl2; (f) R5SO2Cl, Et3N, CH2Cl2.
that appropriate functional antagonism of PPARc/
RXR may be logical approach to protection
against obesity and related diseases such as type 2
diabetes.
was synthesized by the reported procedures.5,7 The
Boc protection of 2 followed by alkylation and depro-
tection afforded N-alkylamine derivatives 3a–c. Amide
4a–m, ureido 5a–e, and sulfonamide 6a–z analogues
were synthesized by direct acylation of 2 with acid chlo-
ride, isocyanate, and sulfonyl chloride, respectively
(Scheme 1).8,9
a
Thus, we initiated to seek more potent RXR antagonists
with good pharmacokinetic properties based on the
structure of HX531.
The diazepinylbenzoic acid derivatives possessing a vari-
ety of functional groups replacing the nitro group of
HX531 were evaluated in the inhibition assay of the
transactivation activity for RXRa homodimer using
LG100268 as an RXR selective ligand and 9-cis-RA as
a natural ligand.10
Figure 2 shows a docking of HX531 in a homology
model of the antagonist-bound form of RXRa based
on the X-ray structure of agonist-bound RXRa and
antagonist-bound PPARa. On the basis of the structural
information from the modeling study, we initially fo-
cused our attention of structural modification on two
functionalized sites, the nitro group and the N-methyl
group, shown in Figure 1 because an open molecular
space around these two sites is available to accommo-
date structural modifications. More importantly, these
two sites were found to be key determinants for antago-
nism and play an important role as agonist/antagonist
switch.5
The amine derivatives 3 turned out to be inactive in the
reporter gene assay although other analogues, the
amides 4, the ureidos 5, and the sulfonamides 6, were
comparably active to HX531. It may suggest that an
electron-withdrawing group seems to be essential as a
functional group for the antagonistic activity. The
amide derivatives 4 inhibited the transactivation with
9-cis-RA more strongly than with LG100268 whereas
the ureidos 5 inhibited it with 9-cis-RA and LG100268
equally (Table 1).
The common intermediate amine 2 was prepared by
SnCl2 reduction of the corresponding nitro 1 which

4810
J. Sakaki et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4808–4811
Table 1. Inhibition of the transactivation of RXRa with amine
derivatives (3–5)
Table 2. Inhibition of the transactivation of RXRa with amine
derivatives (6)
Compound
3: R2
4: R3
5: R4
LGa
9-RAb
Compound R1
R5
LGa
9-RAb
IC50 (lM)
IC50 (lM)
IC50 (lM) IC50 (lM)
6a
6b
6c
6d
6e
6f
Me
Me
Me
Me
Et
Me
Et
4.4
4.9
2.4
HX531
1.2
0.9
0.84
0.25
0.37
0.34
0.55
0.2
n-Pr
n-Bu
n-Pr
2.1
3a
3b
3c
4a
4b
4c
4d
4e
4f
Me
n-Bu
>100
>100
>100
7.1
>10
>10
>10
2.2
0.91
0.57
0.84
0.32
0.85
5.2
Benzyl
Et
n-Pr n-Pr
Et n-Bu
n-Pr n-Bu
6g
6h
6i
n-Pr
i-Pr
3.9
5.6
1.8
1.9
0.14
1.0
Me
Me
Me
Me
Me
Me
Et
Benzyl
Stylyl
Phenyl
Cyclohexyl
Adamantyl
Phenyl
4.5
7.2
1.5
2.9
6j
2.9
1.4
0.4
6k
6l
0.51
2.3
2.0
1.6
0.9
1.7
2-F-phenyl
3-F-phenyl
4-F-phenyl
4-F-phenyl
4.4
2.3
4g
4h
4i
4-F-phenyl
3-CF3-phenyl
Benzyl
4-F-benzyl(R1 = Me)
4-F-benzyl(R1 = Et)
4-F-benzyl(R1 = Pr)
Stylyl
6m
6n
6o
6p
6q
6r
6s
6t
1.2
4.1
2.5
4.2
1.7
0.19
0.22
0.13
0.43
0.47
1.4
0.97
0.88
0.72
3.1
0.42
0.81
2.5
4j
1.3
1.2
n-Pr 4-F-phenyl
4k
4l
Me
Me
Me
Me
Me
Et
4-Me-phenyl
4-OMe-phenyl
2-CF3-phenyl
3-CF3-phenyl
4-CF3-phenyl
3-CF3-phenyl
3.9
5.2
2.5
2.6
4m
3.8
1.5
2.2
0.48
1.1
5a
5b
5c
5d
5e
Me
Et
14
15
6u
6v
6w
6x
6y
6z
7.4
5.2
3.4
2.5
7.3
6.2
3.3
2.8
0.16
0.19
0.095
0.076
1.2
n-Pr
n-Bu
Benzyl
n-Pr 3-CF3-phenyl
Me
Me
Me
3,5-Bis-CF3-Ph 4.1
1-Naphthyl
4-Biphenyl
3.4
6.2
0.85
1.3
Values are means of three independent experiments measured in
duplicate.
a LG100268.
b 9-cis-Retinoic acid.
Values are means of three independent experiments measured in
duplicate.
a LG100268.
b 9-cis-Retinoic acid.
Of these compounds, the sulfonamide analogues 6 were
found to be most potent RXR antagonists (Table 2). Both
simple alkyl and aryl sulfonamides were generally tolera-
ble and show an antagonistic activity with the IC50 value
below 1 lM. A propyl or butyl group is an optimum size
as an alkyl sulfonamide substitution (R5) and an ethyl
group is a most preferable substituent for N-substitution
(R1). Accordingly, the compounds 6e and 6g having
these combinations of alkyl substitution (6e: R1 = Et,
R5 = Pr, 6g: R1 = Et, R5 = Bu) exerted a high antagonistic
activity against both LG100268 and 9-cis-RA.
a
10000
i.v.
p.o.
100
1
In the case of aryl sulfonamides, the 4-F-substitution
pattern (6n) is more active than the corresponding
2-F-(6l) or 3-F-substitution (6m). Thus, its N-ethyl sub-
stitution (6o) showed an IC50 value below 0.5 lM both
for LG100268 and 9-cis-RA. The replacement of 4-F
group of phenyl sulfonamide with a methyl or methoxy
group resulted in a significant decrease in the antagonis-
tic activity. It is intriguing that the 3-substitution pattern
(6t) of the trifluoromethyl group on the phenyl group
showed a good result. The replacement of the N-methyl
group of 6t with an ethyl (6v) or a propyl group (6w)
gave best results, IC50 values of these compounds ex-
erted below 0.1 lM when 9-cis-RA was used as a ligand.
However, larger biaryl substituents (6y and 6z) and even
3,5-bis-trifluoro-methyl phenyl group (6x) were unac-
ceptable in this open space.
0
8
16
24
Time after administration (hr)
b
10000
100
1
i.v.
p.o.
0
8
16
24
Time after administration (hr)
The pharmacokinetics profiles of 6v and 6w were evalu-
ated after the intravenous (1 mg/kg bolus) and peroral
Figure 3. Plasma concentration–time profiles of 6v (a) and 6w (b) after
intravenous (iv) and peroral (po) administration in rats.

J. Sakaki et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4808–4811
4. Mangelsdorf, D. J.; Evans, R. M. Cell 1995, 83, 841.
4811
(3 mg/kg as suspension) administration in rats (Fig. 3).
Blood samples were taken at various time points after
administration and blood levels of parent compounds
were measured with an LC/MS method. After intrave-
nous administration, 6v and 6w were cleared biphasical-
ly from plasma with the CL p values of 0.5 0.2 and
0.4 0.1 L/h/kg, and the terminal elimination half-lives
were 2.1 0.3 and 1.8 0.7 h, respectively. The volumes
of distribution of 6v and 6w at the steady state were cal-
culated as 1.2 0.5 and 1.1 0.3 L/kg. After peroral
administration, 6v and 6w were absorbed well as shown
by the mean tmax of 4.7 1.2 and 4.0 3.5 h reaching
the Cmax of 468 129 and 519 270 nM, respectively.
Oral bioavailabilities of 6v and 6w at doses 1 mg/kg
for intravenous and 3 mg/kg for peroral dosing were
36 and 21%, respectively. These results suggested that
these compounds seem to be a promising candidate for
the indications requiring peroral administration.
5. Ebisawa, M.; Umemiya, H.; Ohta, K.; Fukasawa, H.;
Kawachi, E.; Christoffel, G.; Gronemeyer, H.; Tsuji, M.;
Hashimoto, Y.; Shudo, K.; Kagechika, H. Chem. Pharm.
Bull 1999, 47, 1778.
6. Yamauchi, T.; Waki, H.; Kamon, J.; Murakami, K.;
Motojima, K.; Komeda, K.; Miki, H.; Kubota, N.;
Terauchi, Y.; Tsuchida, A.; Tsuboyama-Kasaoka, N.;
Yamauchi, N.; Ide, T.; Hori, W.; Kato, S.; Fukayama, M.;
Akanuma, Y.; Ezaki, O.; Itai, A.; Nagai, R.; Kimura, S.;
Tobe, K.; Kagechika, H.; Shudo, K.; Kadowaki, T.
J. Clin. Invest 2001, 108, 1001.
7. Umemiya, H.; Fukasawa, H.; Ebisawa, M.; Eyrolles, L.;
Kawachi, E.; Eisenmann, G.; Gronemeyer, H.; Hashim-
oto, Y.; Shudo, K.; Kagechika, H. J. Med. Chem 1997, 40,
4222.
8. Synthesis of compound 6v: To a solution of 5-ethyl-
7,7,10,10-tetramethyl-7,8,9,10-tetrahydro-2-[3-(trifluoro-
methyl)phenylsulfonylamino]-5H-5,13-diazabenzo[4,5]-
cyclohepta[1,2-b]naphthalene-12-yl)benzoic acid methyl
ester (1.10 g, 0.91 mmol) in DMF (12 ml) was added 2 N
NaOH (6.0 ml, 12 mmol) at room temperature. The
reaction mixture was stirred at the same temperature
overnight. The mixture was acidified with 1 N HCl,
diluted with H2O, and extracted with ether. The organic
layer was washed twice with H2O, dried over MgSO4, and
evaporated in vacuo. The resulting solid was washed with
hexane/ether (1:1) and dried to give 6v (400 mg, 65%) as a
In summary, we found potent RXR antagonists based
on the diazepinylbenzoic acid scaffold. An antagonistic
activity of the newly synthesized derivatives is dependent
on a size of the N-alkyl group (R1) and the side chain
(R2–R5) of the functional groups on the diazepine phe-
nyl group. It is also influenced by a degree of electron-
withdrawing property of the functional group. The sul-
fonamide derivatives (6v and 6w) reveal a high antago-
nistic activity and good pharmacokinetic properties.
These compounds are promising oral agents for the
treatment of diabetes and obesity.
1
yellow solid. H NMR (400 MHz, CDCl3) d: 1.06 (3H, s),
1.14 (3H, s), 1.21 (3H, t), 1.26 (3H, s), 1.31 (3H, s), 1.62–
1.70 (4H, m), 3.53–3.72 (2H, m), 6.42 (1H, s), 6.84–6.88
(3H, m), 6.92–6.96 (2H, m), 7.59 (1H, t), 7.78 (1H, d), 7.86
(2H, d), 7.95–7.97 (3H, m), 8.13 (1H, d).
9. Compound 6w: 1H NMR (400 MHz, DMSO-d6) d: 0.80
(3H, t), 1.01 (3H, s), 1.10 (3H, s), 1.23 (3H, t), 1.27 (3H, s),
1.44–1.62 (6H, m), 3.37–3.43 (1H, m), 3.65–3.70 (1H, m),
6.84–6.99 (5H, m), 7.73 (2H, d), 7.82 (1H, t), 7.90 (1H, s),
7.99–8.07 (4H, m), 10.29 (1H, br s), 13.14 (1H, br s).
10. EK-293 cells were co-transfected with the hRXRa expres-
sion vector, pcDNA-hRXRa and the reporter plasmid,
pGL3(CRBPII)2. For transfection, 100 lL/well of the cell
suspension (1.2 · 106 cells/mL) was mixed with DNA-
LipofectAmine2000 (50 lL/well in OPTI-MEM I serum
free medium) and dispensed to the wells of a 96-well flat-
bottomed assay plate. After a 6-h incubation, 2 nM
LG100268 or 20 nM 9-cis-RA as the activating ligands
was added in the presence or absence of a test compound
as 0.001 so 10 lM (final concentration) as dilutions of
50 lL/well of the medium. After further 20 h incubation,
the cells were lysed and the luciferase activity was
measured.
Acknowledgments
The authors thank Messrs. Junichi Yamanaka and
Toshiyuki Kurihara and Ms. Toshie Kurasawa for syn-
thetic support of the antagonists. We thank Mses. Seri-
na Nakano and Akiko Kato for technical support of the
reporter gene assay.
References and notes
1. Nagpal, S.; Saunders, M.; Kastner, P. Cell 1992, 70, 1007.
2. Kastner, P.; Mark, M.; Chambon, P. Cell 1995, 83, 859.
3. Sporn, M. B.; Roberts, M. B.; Goodman, A. B. In The
Retinoids; Academic: Orlando, 1984.