Synthesis and evaluation of spirobenzazepines as potent vasopressin receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2004.02.103

A novel series of spirobenzazepines was synthesized and evaluated for V_1a and V₂ receptor antagonist activity. Compounds 8b, 8i, and 8k have shown selective V_1a receptor antagonist activity. Compounds 8p and 8q were shown to be dual V_1a/V₂ receptor antagonists.

Full text of DOI:10.1016/j.bmcl.2004.02.103

Bioorganic & Medicinal Chemistry Letters 14 (2004) 2987–2989
Synthesis and evaluation of spirobenzazepines as potent vasopressin
receptor antagonists
Min Amy Xiang, Robert H. Chen,*, Keith T. Demarest, Joseph Gunnet, Richard Look,
William Hageman, William V. Murray, Donald W. Combs and Mona Patel*
Endocrine Therapeutics and Metabolic Disorders, Johnson and Johnson Pharmaceutical Research and Development,
L.L.C. 1000 Route 202, Bldg: PCC; Room: PC110, Raritan, NJ 08869-0602, USA
Received 27 January 2004; accepted 18 February 2004
Abstract—A novel series of spirobenzazepines was synthesized and evaluated for V1a and V2 receptor antagonist activity. Com-
pounds 8b, 8i, and 8k have shown selective V1a receptor antagonist activity. Compounds 8p and 8q were shown to be dual V1a/V2
receptor antagonists.
Ó 2004 Elsevier Ltd. All rights reserved.
Arginine vasopressin (AVP) is a cyclic nonapeptide
secreted from the posterior pituitary and is known to exert
a variety of biological effects. Upon release it is known
to interact with AVP receptor subtypes V1a, V1b, and V2.
Interaction at the V1a receptor results in vasoconstric-
tion, vascular smooth muscle proliferation among other
responses, whereas interaction at the V2 receptor results
predominantly in the regulation of osmotic water
channels in the kidneys.1 As a result, the development of
AVP receptor antagonists could prove to be therapeu-
tically useful in the treatment of congestive heart failure,
liver cirrhosis, hyponatremia, and nephrotic syndrome.2
tan and Lixivaptan are both selective vasopressin V2
receptor antagonists, whereas Conivaptan is a dual V1a/
V2 receptor antagonist. All three compounds are based
on the same benzazepine scaffold bearing a lower aryl
chain. Additionally, all compounds have a basic nitro-
gen either in the form of a fused heterocycle or as a
substituent on the seven-membered ring (Fig. 1).
We had previously demonstrated the effective design and
use of highly constrained quaternary systems as peptide
receptor antagonists.4 Based on this earlier work, we
sought to explore spirobenzazepines as V1a/V2 receptor
antagonists. Toward that end, commercially available
3-ethoxycyclohexene-1-one 1 with treated with benzyl
magnesium bromide followed by reduction with lithium
Mozavaptan, Conivaptan, and Lixivaptan are known
AVP antagonists undergoing clinical trials.3 Mozavap-
Figure 1. Known AVP antagonists.
* Corresponding author. Tel.: +1-908-707-3558; fax: +1-908-203-8109; e-mail: mpatel5@prdus.jnj.com
Present address: Ace Chemicals and Pharmaceutical, M. L. King, Jr. Boulevard, Newark, NJ 07102, USA.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.02.103

2988
M. A. Xiang et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2987–2989
Table 1. In vitro activity of spirobenzoxazines
aluminum hydride (LAH) to afford allyl alcohol 2 in a
two-step process in 80% yield. Alcohol 2 was treated
with sodium hydride and phenyl vinyl sulfoxide to
provide (in Michael type addition) the corresponding
allyloxy ethyl aryl sulfoxide. Upon heating in decalin,
the crude sulfoxide underwent a Claisen rearrangement
to provide aldehyde 3 in a 70% yield over two steps.5
Oxidation of aldehyde 3 with sodium chlorite to car-
boxylic acid (4) proceeded in 65% yield, and was fol-
lowed by an intramolecular Friedel Craft’s acylation
with trifluoroacetic anhydride (TFAA) and trifluoro-
acetic acid (TFA) to afford spiroketone 5 as a racemic
mixture. Schmidt rearrangement of spiroketone 5 with
sodium azide and TFA provided spirolactam 6.
Reduction of the amide to the corresponding amine (7)
with LAH was followed by acylation with an appro-
priately substituted acid chloride to afford spiro-
benzazepine 8 (Scheme 1).6
Compound
Ra
V1a Binding7
IC50 (nM)
V2 Binding7
IC50 (nM)
8a
8b
8c
8d
8e
8f
H
2-CH3
20
10
390
70
3-CH3
4-CH3
50
74
380
>1000
>1000
0.20
3,4-diCH3
2-OCH3
2,6-diOCH3
3,4-diOCH3
4-SCH3
4-SOCH3
4-NH2
140
30
8g
8h
8i
41
>1000
15
>1000
>1000
>1000
>1000
>1000
550
8j
>1000
10
10
8k
8l
Table 1 summarizes the in vitro potency of the racemic
benzazepines. Compound 8a bearing no substitutions on
the phenyl ring was found to be a potent V1a antagonist
with moderate V2 antagonist activity. The introduction
of a methyl group at positions 2, 3, or 4 of the phenyl
ring resulted in compounds 8b, 8c, and 8d, respectively.
Compound 8b bearing a methyl group at the 2 position
was clearly the most potent compound of the set with an
IC50 of 10 nM in the V1a binding assay and an IC50 of
70 nM in the V2 binding assay. The 3,4-dimethyl
substituted compound 8d showed a loss of potency in
the V1a binding assay and is a poor V2 receptor anta-
gonist. Compound 8f bearing a 2-methoxy substituent
on the phenyl ring was a potent V1a antagonist with
moderate V2 antagonist activity, but still less potent
than the 2-methyl compound 8b. The 2,6-dimethoxy
substituted compound 8g showed a moderate loss in
potency in the V1a binding assay and had negligible V2
2-F
2-Cl
8m
8n
8o
8p
8q
12
46
180
159
2,6-diCl
2,4-diCl
2-CH3, 5-F
2-Ph
16
274
67
8
26
29
a All compounds are racemic.
receptor antagonist activity, whereas compound 8h
bearing a 3,4-dimethoxy substitution pattern resulted in
a loss of potency in both V1a and V2 binding assays. The
4-thiomethyl substituted compound 8i was a potent V1a
receptor antagonist with an IC50 value of 15 nM with
very poor potency in the V2 receptor binding assay.
Compound 8i is a selective V1a receptor antagonist.
Oxidation of the thiomethyl group to the corresponding
Scheme 1. Reagents and conditions: (a) BnMgCl, THF; (b) LAH, THF, 80% over two steps; (c) phenyl vinyl sulfoxide, NaH, THF; (d) decalin,
NaH, sodium bicarbonate, 70% over two steps; (e) sodium chlorite, sodium phosphate, 65%; (f) trifluoroacetic anhydride, trifluoroacetic acid, 90%;
(g) sodium azide, trifluoroacetic acid, 65%; (h) LAH, THF, 97%; (i) ArCOCl, TEA, DCM, 65–90%.

M. A. Xiang et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2987–2989
2989
Biochem. Soc. Trans. 1979, 7, 861; (c) Jard, S. Kidney Int.,
Supp. 1988, 26, 38; (d) Thibonnier, M. Regul. Pep. 1992,
38, 1.
sulfone moiety (8j) resulted in an inactive compound in
both V1a and V2 binding assays. The 2-amino substi-
tuted compound 8k was also a V1a selective compound
with an IC50 of 10 nM and negligible potency in the V2
binding assay. Introduction of halogens on the phenyl
ring in the form of 2-fluoro (8l) and 2-chloro (8m)
substituted phenyl compounds, resulted in potent V1a
antagonists with moderate (180–550 nM) V2 antagonist
activity. The introduction of the dihalo substituted
phenyl ring compounds such as 2,6-dichloro (8n) and
2,4-dichloro (8o) did not alter the V1a/V2 potency profile
with respect to that observed for the monohalo substi-
tuted compounds (8l and 8m). Compound 8p bearing a
2-methyl,5-fluoro substituent provided a potent V1a and
V2 dual antagonist with IC50 values of 8 and 67 nM,
respectively. The introduction of a 2-phenyl group
resulted in compound 8q, bearing an excellent V1a and
V2 dual antagonist activity profile with IC50 values of 26
and 29 nM, respectively. Next, we sought to examine the
metabolic stability and solubility profiles of the com-
pounds of interest. We were disappointed to discover
that the metabolic profiles for the potent compounds of
this series were rather poor with half-lives in the human
microsomal stability assay consistently being less than
10 min. Solubilities of compounds of interest were
determined to be below detectable levels at both pH 2
and 7.4.
2. (a) Laszlo, F. A.; Laszlo, F., Jr.; De Wied, D. Pharmacol.
Rev. 1991, 43, 73; (b) Mah, S. C.; Hofbauer, K. G. Drugs
Fut. 1987, 12, 1055; (c) Paranjape, S. B.; Thibonnier, M.
Exp. Opin. Invest. Drugs 2001, 10, 825; (d) Trybulski, E. J.
In Annual Reports in Medicinal Chemistry; Doherty, A. H.,
Ed.; Academic: San Diego, 2001; Vol. 36, pp 159–168; (e)
Albright, J. D.; Chan, P. S. Curr. Pharm. Des. 1997, 3, 615;
(f) Matthews, J. W.; Hoekstra, W. J.; Andrade-Gordon, P.;
de Garavilla, L.; Demarest, K. T.; Erickson, E.; Greco, M.
N.; Gunnet, J. W.; Hageman, W.; Hecker, L. R.; Moore, J.
B.; Maryanoff, B. E. Bioorg. Med. Chem. Lett. 2003, 13,
753.
3. Lixivaptan: (a) Martinez-Castelao, A. Curr. Opin. Investig.
Drugs 2001, 2, 525; (b) Albright, J. D.; Reich, M. F.; Delos
Santos, E. G.; Dusza, J. P.; Sum, F.-W.; Venkatesan, A.
M.; Coupet, J.; Chan, P. S.; Ru, X.; Mazandarani, H.;
Bailey, T. J. Med. Chem. 1998, 41, 2442; (c) Albright, J. D.;
Delos Santos, E. G.; Dusza, J. P.; Chan, P. S.; Coupet, J.;
Ru, X.; Mazandarani, H. Bioorg. Med. Chem. Lett. 2000,
10, 695; Conivaptan: (a) Norman, P.; Leeson, P. A.;
Rabasseda, X.; Castaner, J.; Castaner, R. M. Drugs Fut.
2000, 25, 1121; (b) Matsushita, A.; Taniguchi, N.; Koshio,
H.; Yatsu, T.; Tanaka, A. Chem. Pharm. Bull. 2000, 48, 21;
Mozavaptan: (a) Prous, J.; Mealy, N.; Castaner, J. Drugs
Fut. 1993, 18, 802; (b) Ogawa, H.; Yamashita, H.; Kondo,
K.; Yamamura, Y.; Miyamoto, H.; Kan, K.; Kitano, K.;
Tanaka, M.; Nakaya, K.; Nakamura, S.; Mori, T.;
Tominaga, M.; Yabuuchi, Y. J. Med. Chem. 1996, 39,
3547.
4. Beavers, M. P.; Gunnet, J. W.; Hageman, W.; Miller, W.;
Moore, J. B.; Zhou, L.; Xiang, A.; Urbanski, M.; Combs,
D. W.; Mayo, K. H.; Demarest, K. T. Drug Des. Discov.
2001, 17, 243.
5. Mandai, T.; Ueda, M.; Hasegawa, S.; Kawada, M.; Tsuji,
J.; Saito, S. Tetrahedron Lett. 1990, 31, 4041.
6. All compounds provided satisfactory spectral data (1H
NMR, CI-MS/ESI/MS, and were homogeneous by TLC).
7. Rigas, A.; Levine, L. J. Pharmacol. Exp. Ther. 1984, 231,
230, Binding affinities were determined by measuring the
inhibition of 3[H] AVP binding to cloned human V1a and V2
receptors. Initial compounds prepared in this series were
subjected to the following functional assays and were
determined to be vasopressin antagonists: The accumula-
tion of cAMP was measured in transfected HEK-293 cells
expressing human V2 receptors and the intracellular
calcium mobilization was measured in HEK-293 cells
transfected to express human V1a receptors.
In summary, although we were able to identify novel
compounds that were both selective V1a receptor
antagonists in compounds 8b, 8g, 8i, and 8k as well as
dual V1a/V2 receptor antagonists in compounds 8p and
8q, lack of metabolic stability and a poor solubility
profile precluded further development of this series of
compounds. The spiro substituent does not appear to
regulate selectivity in this template, and could be a
probable cause of the poor metabolic stability and sol-
ubility profiles. Additional alterations in the scaffold are
underway and will be reported in due course.
References and notes
1. (a) Laszlo, F. A.; Laszlo, F., Jr. Drug News Perspect. 1993,
6, 591; (b) Michel, R. H.; Kirk, C. J.; Billah, M. M.