Array synthesis of progesterone receptor antagonists: 3-Aryl-1,2-diazepines

Article abstract of DOI:10.1016/j.bmcl.2006.04.055

New non-steroidal chemotypes are required for the development of drugs targeting the steroid hormone receptors. The parallel array synthesis of 3-aryl-1,2-diazepines employing solid-supported reagents is described. The resulting compounds demonstrated high affinity binding to the progesterone receptor.

Full text of DOI:10.1016/j.bmcl.2006.04.055

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3777–3779
Array synthesis of progesterone receptor antagonists:
3-Aryl-1,2-diazepines
Robert W. Wiethe, Eugene L. Stewart, David H. Drewry, David W. Gray,
Abdul Mehbob and William J. Hoekstra*
Discovery Research, GlaxoSmithKline, Research Triangle Park, NC 27709-3398, USA
Received 21 March 2006; revised 17 April 2006; accepted 17 April 2006
Available online 5 May 2006
Abstract—New non-steroidal chemotypes are required for the development of drugs targeting the steroid hormone receptors. The
parallel array synthesis of 3-aryl-1,2-diazepines employing solid-supported reagents is described. The resulting compounds demon-
strated high affinity binding to the progesterone receptor.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Progesterone receptor (PR) ligands have utility as treat-
ments for a range of reproductive diseases and have
been the target of several medicinal chemistry pro-
grams.1 The PR antagonist mifepristone (RU-486) is
prescribed for termination of pregnancy and also offers
promise for the treatment of endometriosis, uterine fib-
roids, and breast cancer.2 Recently, non-steroidal PR
antagonists and partial agonists have been identified
with improved reproductive tissue safety profiles.3,4
Nevertheless, new chemotypes that are amenable to
high-throughput synthesis are useful to further structur-
al understanding of PR modulation.
this disclosure, 1-N-alkyl-3-aryl-1,2-diazepines had been
isolated as pyrolysis side products.8 The template has
also been employed to prepare a diazepine thiosemicar-
bazide.9 The key annulation step in the synthesis of
diazepine targets involved treatment of phenyl-d-chlo-
robutyl ketones with hydrazine (Scheme 1). Unfortu-
nately, decomposition of the product occurred upon
standard aqueous workup. Modifications to the
protocol were required to allow parallel synthesis of
the diazepine targets. Details of the improved synthesis
are presented below.
Compounds were prepared starting from substituted
aryl-d-chlorobutyl ketones 4 (Scheme 1, for monomers
4, see Ref. 14).10 Initial experiments showed that ring
closure using 4 equiv of hydrazine hydrate in ethanol
at reflux overnight gave good yields of the diazepine
intermediates 5. Solvation with 2-propanol gave
improved annulation yields when o-substituted aryl
ketones (R1) were employed. The workup consisted of
a rapid filtration of the reaction to remove the majority
of the insoluble hydrazine hydrochloride byproduct,
followed by concentration of the filtrate to yield the
crude product. The remaining traces of hydrazine
hydrochloride were removed by suspending the residue
in dry dichloromethane, followed by filtration and
concentration of the solution. Attempts to chromato-
graph or extract the diazepine intermediate resulted in
typically >50% decomposition of product. In contrast,
the intermediates were stable at room temperature for
weeks if kept dry. The diazepine intermediate 5 could
also be precipitated from a dry diethyl ether solution
3-Aryl-pyridazine PR agonist 15 and 3-aryl-pyrazoline
PR antagonist 26 have been previously disclosed. The
binding mode of the pyrazolines has been reported,
wherein the 3-aryl moiety occupies the steroid A-ring
binding pocket of PR.6 In our medicinal chemistry pro-
gram, we aimed to introduce a variety of pharmaco-
phores to a structurally related 1,3-disubstituted-1,2-
diazepine core via parallel synthesis. The strategy
furnished N-arylsulfonyl-3-aryl-1,2-diazepine based PR
ligands. The synthesis of high affinity 3-aryl-1,2-diaze-
pines 3 is disclosed herein.
The targeted synthesis of the 3-aryl-1,2-diazepine core
was first reported by Koenig and Wermuth.7 Prior to
Keywords: Progesterone receptor; Diazepine.
*
Corresponding author. Tel.: +1 919 483 8212; fax: +1 919 315
0430; e-mail: william.j.hoekstra@gsk.com
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.04.055

3778
R. W. Wiethe et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3777–3779
Me2N
Me
Me2N
OH
H
O
S
N
H
N
N
O
S
O
R1
N
N
N
Cl
O
S
I
O
CF3
O
O
R2
Cl
Cl
Cl
Mifepristone
1
2
3
1) R2SO2Cl, PS-DMAP,
DCM, RT, 16 h
O
NH2-NH2(4 eq),
iPrOH, 75oC, ~16 h
N
O
Cl
N
H
O
N
N
2) Si-TrisAmine
S
R2
R1
R1
5
R1
4
3
Scheme 1. Parallel synthesis of diazepines.
by addition of HCl in ether to give a stable, hygroscopic
HCl salt. Cyclization could also be effected using micro-
wave heating at temperatures up to 170 ꢁC for 5–30 min;
the forcing conditions were required to form some of the
diazepines with o-substituents (e.g., R1 = 2-bromo, 2-tri-
fluoromethyl). In some of these sterically hindered cases,
additional hydrazine hydrate was required, but 4 equiv
was adequate for most reactions. Use of fewer than
4 equiv of hydrazine led to side products derived from
incomplete ring closure and polymerization. Using this
methodology, solution-phase parallel synthesis in a 96-
well format allowed the isolation of >1300 sulfonamide
products 3 without chromatography on either step.
(<20%). Presumably, formation of HCl promoted
decomposition of the diazepine intermediates 5. An
alternate route employed the ‘catch and release’ method
wherein 4 equiv of arylsulfonyl chloride was reacted
with 4 equiv of polystyrene-supported 4-dimethylamino-
pyridine (PS-DMAP) for 2 h.10 The resin was washed
with dichloromethane to remove the excess arylsulfonyl
chloride followed by addition of diazepine in dichloro-
methane to give products of >90% purity after stirring
overnight. The formation of traces of decomposition
products could be avoided through addition of merely
a catalytic amount of diisopropylethylamine (DIPEA)
or triethylamine to the resin activation medium to buffer
the reaction mixture. Finally, Silicycle Si-TrisAmineꢂ
scavenger was added and agitation continued for 1–2
more hours to remove any remaining arylsulfonyl chlo-
ride. Filtration and evaporation of solvent furnished
diazepines 3 with purities acceptable for biological test-
ing (85–100%, Table 1).
Standard conditions for the preparation of sulfonamides
3 from intermediates 5 (1 equiv arylsulfonyl chloride,
pyridine or 4-dimethylaminopyridine in dichlorometh-
ane) gave variable results (12–47% isolated yields). Typ-
ically, poorer yields were realized as scale was increased
Table 1. Isolated chemical yields and PR activities of diazepines
R1
N
N
O
S
O
R2
3
Compound
R1
R2
Isolated yield (%)
LC purity (%)
PR Bdg.a pKi
CV-1 cellb pIC50 (% max)
3a
3b
3,4-Cl2
3,4-Cl2
3,4-Cl2
4-CN
4-CN
4-Cl
5-Br-2-thienyl
5-Cl-2-thienyl
2,4-Cl2-Ph
64
93
85
100
94
7.6
7.4
7.3
6.0
7.2
6.2
6.7
6.1
6.5
7.2
7.3
8.0
6.5 (60)
NTc
3c
3d
32
86
6.7 (82)
6.4 (89)
NT
2-F–Ph
5-Br-2-thienyl
3-Cl–Ph
100
100
100
96
3e
85
3f
3g
100
72
5.9 (89)
6.2 (68)
6.3 (91)
5.7 (86)
6.6 (71)
6.8 (65)
9.6 (100)
2-Cl
3-F
5-Cl-2-thienyl
4-Cl–Ph
3h
3i
82
95
96
3-Cl
3-Br
3-Cl–2-Me–Ph
5-Br-2-thienyl
4-CF3O–Ph
100
87
3j
93
92
3k
RU-486
4-NO2
74
a Assay measures compound interaction with the ligand binding domain of PR by displacement of a fluorescent ligand (n P 3, SD = 0.25).11
b Assay measures inhibition of progesterone-stimulated (4 nM) transactivation of BacMam-expressed human PR-B in CV-1 cells using an MMTV-
Luc reporter (n P 2, SD = 0.20).12
c NT, not tested.

R. W. Wiethe et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3777–3779
3779
6. Jones, D. G.; Liang, X.; Stewart, E. L.; Noe, R. A.;
Kallander, L. S.; Madauss, K. P.; Williams, S. P.;
Thompson, S. K.; Gray, D. W.; Hoekstra, W. J. Bioorg.
Med. Chem. Lett. 2005, 15, 3203.
7. Koenig, J.-J.; Wermuth, C.-G. Tetrahedron Lett. 1973, 63.
8. Wawzonek, S.; Stephanie, J. G. J. Org. Chem. 1971, 36,
2467.
Sulfonamides 3a–k were highly active representatives of
the >1300 1,3-disubstituted diazepines that were tested
for progesterone receptor binding affinity13 and func-
tional antagonism in CV-1 cells (Table 1, array com-
pounds with highest affinity R1 groups shown).
Products wherein the R1 group was 3,4-dichloro exhib-
ited high affinity for PR (e.g., 3a–c). The 5-bromo-2-thi-
ophene-sulfonamide 3a (pKi = 7.6) bound to PR with
similar affinity to the steroidal antagonist RU-486
(pKi = 8.0). Dichloro substitution (e.g., 3a,b) on the 3-
aryl moiety provided higher affinity ligands than those
with monochloro substitution (e.g., 3f,g). Array mem-
bers wherein R1 was H, 3-methyl, 4-methyl, or 4-meth-
oxy did not display measurable PR binding
(pKi < 5.0). Halogenated thiophenes represented one of
the optimal arene groups at R2. Generally, halo-aryl
substituents at R2 were necessary for high affinity PR
binding. Electron-rich and bicyclic aromatic R2 groups
gave poor PR binding. Consistent with the related pyr-
azoline series (and unlike the pyridazine based, breast
cell stimulatory agonist series5) the diazepines generally
antagonized progesterone-stimulated reporter activity in
CV-1 cells (for 3c, pIC50 = 6.7).6 Less than 5% of array
members exhibited PR agonist activity. While a small
percentage of array members bound to the glucocorti-
coid receptor with ca. micromolar affinity (pKi = 6.0–
6.3), the diazepines 3 generally exhibited >50-fold selec-
tivity for PR over the other steroid hormone receptors
(data not shown).
9. Kane, J. M.; Stewart, K. T. J. Heterocycl. Chem. 1988, 25,
1471.
10. Diazepine product 3a was prepared as follows: 3,4-
dichlorophenyl-d-chlorobutyl
ketone14
(4a,
2.8 g,
10.5 mmol) was dissolved in 2-propanol (90 mL) in a
120 mL bottle and then hydrazine hydrate (2.0 mL,
41.9 mmol) was added. The bottle was heated in a stirring
heat block at 75 ꢁC overnight (ꢀ16 h) and allowed to
slowly cool to rt. The colorless precipitate was removed by
filtration into a large test tube and all of the solvent and
excess hydrazine were removed in vacuo (using a Savant
Speedvac overnight). The remaining slurry was dissolved
in anhydrous dichloromethane (DCM) and filtered to
remove the hydrazine HCl salt. The filtrate was concen-
trated in vacuo to give 5a as an oil (2.5 g, 99% crude yield,
ꢀ95% purity by both LC/MS and 1H NMR). Second step:
to PS-DMAP resin (168 mg, 1.5 mmol/g, dried under high
vacuum at 50 ꢁC overnight) in one well of a 96-well
Scigene Flexchem reactor was added a solution of DIPEA
(0.02 mL) and anhydrous DCM (1 mL) at rt. This mixture
was treated with 5-bromo-2-thiophenesulfonyl chloride
(0.66 mL, 0.75 M in anhydrous DCE) via a Gilson-215
liquid handler and the plate sealed. The plate was rotated
and agitated for 2 h, and the reactor was placed into the
wash manifold under vacuum to remove the liquid from
the resin. The resin was further washed with 0.5 mL
anhydrous DCM (per well) and the bottom cover was
reinstalled on the reactor. Compound 5a (24 mg,
0.10 mmol in 1 mL anhydrous DCM) was added to the
activated resin. The plate was resealed and rotated at rt for
16 h. The top cover was removed and Silicycle Si-
Trisamineꢂ scavenger resin (135 mg) was added to each
well and the plate resealed and agitated for 2 h. The
reaction well contents were filtered into a 96-well Multi-
chem plate and washed with DCM (3· 0.5 mL). The
filtrate was concentrated in vacuo to afford 3a as a solid
(30 mg, 64% yield, 85% purity by LC/MS).
In summary, we have described a practical, two-step ar-
ray synthesis of new progesterone receptor binding 3-ar-
yl-1,2-diazepines. Solution-phase parallel synthesis using
solid-supported reagents and resin scavengers facilitated
the synthesis of >1300 products while removing the need
for purification of the intermediates and products by
column chromatography. As an application of this
chemistry, key intermediates 4a–k were carried forward
to biologically active PR ligands 3a–k.
11. The PR fluorescence polarization binding assay6,15
measures compound interaction with the ligand binding
domain of PR (590 nm) by displacement of a fluores-
cent steroidal progesterone mimetic (Panvera catalog
number P2964; Kd = 10 nM). The response is expressed
as a pKi.
Acknowledgments
We thank Tim M. Willson, R. Andy Noe, and L.
William Stuart for helpful discussions.
12. The functional assay6 measures compound-mediated
interaction of the PR-B isoform with the MMTV lucifer-
ase reporter to calculate compound potency and efficacy in
BacMam transduced, progesterone-stimulated (4 nM) CV-
1 cells. The antagonist response is expressed as a pIC50
and the percentage of maximal efficacy is measured
relative to the fully efficacious antagonist RU-486 (RU-
486 pIC50 = 9.6, 100% efficacy).
References and notes
1. Ashok, P. W.; Wagaarachi, P. T.; Templeton, A. Curr.
Med. Chem. Immun. Endocr. Metab. Agents 2002, 2, 71.
2. Hess-Stumpp, H.; Hoffmann, J.; Fuhrmann, U. Drugs
Future 2002, 27, 1113.
3. Zhi, L.; Tegley, C. M.; Pio, B.; Edwards, J. P.; Motamedi,
M.; Jones, T. K.; Marschke, K. B.; Mais, D. E.; Risek, B.;
Schrader, W. T. J. Med. Chem. 2003, 46, 4104.
4. Zhi, L.; Ringgenberg, J. D.; Edwards, J. P.; Tegley, C. M.;
West, S. J.; Pio, B.; Motamedi, M.; Jones, T. K.;
Marschke, K. B.; Mais, D. E.; Schrader, W. T. Bioorg.
Med. Chem. Lett. 2003, 13, 2075.
5. Palmer, S.; Campen, C. A.; Allan, G. F.; Rybczynski, P.;
Haynes-Johnson, D.; Hutchins, A.; Kraft, P.; Kiddoe, M.;
Lai, M. T.; Lombardi, E.; Pedersen, P.; Hodgen, G.;
Combs, D. W. J. Steroid Biochem. Mol. Biol. 2000, 75, 33.
13. Approximately 289 of the >1300 array members were
progressed from single concentration testing to the full-
curve PR binding assay. Eighteen members exhibited pKi
7–7.6, 119 members at pKi 6.0–6.9, 127 members at pKi
5.0–5.9, and 25 members at pKi <5.0.
14. d-Chlorobutyl aryl ketones were purchased from Rieke
metals. The following monomers were purchased: R1 = 3-
or 4-substituted F, Cl, Br, Me, CF3, and NO2; R1 = H, 4-
OMe, 4-CN, 3,4-F2, 3,4-Cl2, and 3-Me-4-Br.
15. Ohno, K.-I.; Suzuki, S.; Fukushima, T.; Maeda, M.;
Santa, T.; Imai, K. Analyst 2003, 128, 1091.