The design, synthesis and physical chemical properties of novel human vasopressin V2-receptor antagonists optimized for parenteral delivery

Article abstract of DOI:10.1016/S0960-894X(00)00095-0

Ionizable groups were introduced onto the 10,11-dihydro-5H-pyrrolo[2,1-c][1,4]benzodiazepine scaffold of the vasopressin V₂-antagonist WAY-VPA-985 in the search for molecules optimized for parenteral formulation. The synthesis and structure-activity relationships (SAR) are presented together with solubility data in a model parenteral system. The amine, WAY-140288 (4f), was chosen for further development. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00095-0

Bioorganic & Medicinal Chemistry Letters 10 (2000) 783±786
The Design, Synthesis and Physical Chemical Properties of Novel
Human Vasopressin V2-Receptor Antagonists Optimized for
Parenteral Delivery
M. A. Ashwell,* J. F. Bagli, T. J. Caggiano, P. S. Chan, A. J. Molinari, C. Palka,
C. H. Park, J. F. Rogers, M. Sherman, E. J. Trybulski and D. K. Williams
Wyeth-Ayerst Research, Princeton, NJ 08543, USA
Received 9 December 1999; accepted 7 February 2000
AbstractÐIonizable groups were introduced onto the 10,11-dihydro-5H-pyrrolo[2,1-c][1,4]benzodiazepine sca€old of the vasopressin
V2-antagonist WAY-VPA-985 in the search for molecules optimized for parenteral formulation. The synthesis and structure±
activity relationships (SAR) are presented together with solubility data in a model parenteral system. The amine, WAY-140288 (4f),
was chosen for further development. # 2000 Elsevier Science Ltd. All rights reserved.
Arginine vasopressin (AVP, anti-diuretic hormone) cri-
tically regulates water reabsorption by interaction with
the vasopressin V2-receptors found in the collecting
ducts (aquaporin-2 water channels) of the kidney and as
a result extracellular osmolality is critically maintained.1
A vasopressin V2-antagonist would therefore be expec-
ted to normalize plasma osmolality, control excessive
water retention and dilutional hyponatremia in such
conditions as congestive heart failure, liver cirrhosis and
renal failure.
and in vivo characteristics. The synthesis, structure±
activity relationships (SAR) and solubility of these new
molecules is described.
Synthesis
The synthetic approach used in elaboration of the
10,11-dihydro-5H-pyrrolo[2,1-c][1,4]benzodiazepine tri-
cyclic sca€old and the SAR of a number of related
vasopressin V2-receptor antagonists has been published
previously.3,5,6
Several small molecule vasopressin antagonists have
been described.2 One potent human vasopressin V2-
antagonist, N-4-[3-chloro-4-(5H-pyrrolo[2,1-c][1,4]benzo-
diazepin-10(11H)-ylcarbonyl)phenyl]-5-¯uoro-2-methyl-
benzamide (WAY-VPA-985; 1, R1=Cl, R2=H,
R3=Me, R4=F)3 is undergoing phase II clinical trials.
Utilizing the previous methodology gave access to a
wide variety of 10,11-dihydro-5H-pyrrolo[2,1-c][1,4]-
benzodiazepine tricyclic sca€olds. The chemical strategy
employed in this study involved functionalization of the
pyrrole ring leading to acids, acylamines7 and Mannich
derived alkylamines8 as shown in Scheme 1. Acylamines
were prepared from the acid derivatives 2. Initial reac-
tion of 1 with trichloroacetylchloride9 was followed by
treatment with sodium hydroxide in aqueous THF to
furnish acids 2 in overall yields of 60±70%. Amine cou-
pling of the acids led directly to 3 (typically 80±90%
yield).
The goal of the work described here was to discover new
potent, selective, orally active, non-peptide antagonists
of the vasopressin V2-receptor which had the additional
feature of being optimized for parenteral administra-
tion.4 This report describes the manipulation of the
10,11-dihydro-5H-pyrrolo[2,1-c][1,4]benzodiazepine tri-
cyclic sca€old, which is common to this class of vaso-
pressin receptor agents in ways designed to increase
aqueous solubility yet maintain the desirable in vitro
Mannich reactions were performed directly on the pyrrole
sca€old.10 Generally, the corresponding secondary
amines were utilized with paraformaldehyde in good yield
(60±80%), however, for the introduction of the 3-di-
methylaminomethyl group (4a, 4f, 4g where R=NMe2)
*Corresponding author. Tel.: +1-732-274-4416; fax: +1-732-274-
4505; e-mail: ashwellm@war.wyeth.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00095-0

784
M. A. Ashwell et al. / Bioorg. Med. Chem. Lett. 10 (2000) 783±786
Scheme 1. (a) CCl3COCl, CH2Cl2; (b) NaOH (aq), THF; (c) EDC, CH2Cl2, amine; (d) MeOH, AcOH, (CH2O)n, amine.
tetramethyldiaminomethane was employed in a similar
yield. In the case of 4g demethylation of 1 (R1=OMe,
R2=H, R3=Ph, R4=H) with BBr3, preceded the Man-
nich reaction (yield 75%).
for the acylamines encouraged us to determine their
solubility.
A target solubility equal to or greater than 1 mg/mL in
a model parenteral system consisting of 30% poly-
ethylene glycol 200 (PEG-200)/water was set based on
the expected volume of delivery and pharmacological
pro®le of the drug candidate. These criteria necessitated
having crystalline material as measured by di€erential
scanning calorimetry. However, in general these acyla-
mines proved dicult to crystallize. Measurements on
amophorus material of both free bases and salts were
promising and a study was initiated to investigate alter-
native salts11 as a means to alter the physical character-
istics of the parent acylamines.
Results and Discussion
Acids
The introduction of an acid group to form 10,11-di-
hydro-5H-pyrrolo[2,1-c][1,4]benzodiazepine-3-carboxylic
acids maintains in vitro vasopressin V2-antagonism.
However, these compounds were not active orally (2a,
2b in Table 1). The lack of oral activity was attributed
to poor absorption presumably from limited dissolu-
tion. This assumption was supported by the observation
that iv dosing restored in vivo activity (2a: urine volume
ratio of 5: @ dose 10 mg/kg).
Amines
The Mannich procedure allowed the synthesis and ela-
boration of a second series of basic molecules. In gen-
eral these compounds were shown to have good in vitro
potency. However, examples containing a 2-phenyl-
benzoyl derived secondary amide such as 4a were found
to be orally superior to those containing the 2-methyl-5-
¯uorobenzoyl group present in VPA-985 (e.g., 4i in
Table 2) the former being active at doses well below 10
mg/kg. A similar e€ect is described for the related pre-
cursor pyrrolobenzodiazepine sca€old.12 The successful
combination of a 2-phenylbenzoyl derived amide and a
basic center on the pyrrole ring led to the development
of a series of amines (4a±h).
Acylamines
Utilization of the 10,11-dihydro-5H-pyrrolo[2,1-c][1,4]-
benzodiazepine-3-carboxylic acids to form amides was a
useful synthetic strategy for further modi®cation of the
sca€old. It was not until amides containing a basic cen-
ter were incorporated, however, that signi®cant increa-
ses in solubility were achieved.
A series of such acylamines was prepared (3a±c in Table
1). They were characterized by good in vitro binding
and in some cases extremely potent in vivo activity (e.g.,
3a). The incorporation of two basic centers was also
tolerated in vivo (e.g., 3b). The pharmacological results
The pharmacological pro®le of the amines having either
one or two basic centers was comparable. In general

M. A. Ashwell et al. / Bioorg. Med. Chem. Lett. 10 (2000) 783±786
Table 1. In vitroa and in vivob antagonist activity
785
Urine volume V2
No.
R
X
R1
R2
R3
R4
ratioc
IC50 nM
2a
2b
3a
OH
OH
O
O
O
Cl
Cl
Cl
H
H
H
Me
Ph
Ph
F
H
H
NEe
NE
3.6i
14
4
11.5
3b
3c
O
O
Cl
Cl
H
H
Ph
Ph
H
H
2.7h
NTf
2
8.8
4a
4b
NMe2
H2
H2
Cl
Cl
H
H
Ph
Ph
H
H
6.6
3.75
7.5
1.3
4c
4d
H2
H2
Cl
Cl
H
H
Ph
Ph
H
H
4.2
1.5
2.6
2.1
4e
H2
Cl
H
Ph
H
2.3
2
4f
4g
4h
NMe2
NMe2
H2
H2
H2
H
OH
OMe
OMe
H
H
Ph
Ph
Ph
H
H
H
5
3.8
7.8d
5.2
14.8
7
1
WAY-VPA-985
Cl
H
Me
F
7.4g
1.2
3
aCompounds were tested as the free bases (or acids) except as noted. Compounds were tested in vitro for their ability to displace H AVP in mem-
brane preparations from a murine ®broblast cell line (LV2) expressing the humna vasopressin V2-receptor.
bIn vivo studies were conducted in conscious normotensive Sprague±Dawley rats with free access to water before the experiment. The compounds
under test were dosed orally (po) at 10 mg/kg (20% dimethyl sulfoxide and 2.5% pre-boiled starch preparation.
cData reported as a ratio of urine volume collected from treatment group after 4 h versus control group.
dNo dimethyl sulfoxide in vehicle.
eNE=no e€ect.
fNT=not tested.
gConcious rats dosed (ip) with 0.4 mg/kg of AVP (in peanut oil); 20 min later given orally (by gavage) 30 mL/kg of deionized water followed by
WAY-VPA-985 20 min later.3
hHCl salt.
iCitrate salt.

786
M. A. Ashwell et al. / Bioorg. Med. Chem. Lett. 10 (2000) 783±786
Table 2. Low dose in vivoa antagonist activity
(4f±h). These molecules had the desired pharmacological
pro®le and were reproducibly crystalline. Potent oral
activity was seen in the absence of dimethyl sulfoxide in
the test vehicle (e.g. 4h) although this initial example failed
to meet the solubility cut-o€ (see entry 4h in Table 3).
The strategy proved e€ective, however, with other
examples (4f and 4g). Thus these amines met the design
needs of the program.
Summary and Conclusion
Urine volume ratiob
The utilization of the chemical reactivity of the pyrrole
nucleus present in the tricyclic 10,11-dihydro-5H-pyr-
rolo[2,1-c][1,4]benzodiazepine sca€old led to the dis-
covery of two series of amines and acylamines with
potent in vitro and in vivo vasopressin V2-antagonism.
In the case of the Mannich derived amines the target
level of solubility was only achieved when further
hydrogen bonding groups were added to the sca€old.
These materials were also reproducibly crystallized. In
the case of the acylamines the target solubility was
achieved (3a) through an appropriate salt choice
(citrate). Of these molecules WAY-140288 (N-(4-{[3-
[(dimethylamino)methyl] - 5H - pyrrolo[2,1 - c][1,4]benzo-
diazepin-10(11H)-yl]carbonyl}-2-methoxyphenyl)[1,10 -
biphenyl]-2-carboxamide) (4f) was chosen as a develop-
ment candidate based on both solubility measurements
and a promising pharmacologic pro®le.
No.
R3
R4
@ 3 mg/kg
@ 1 mg/kg
4i
4a
Me
Ph
F
H
1.4
5.6
1.4
2.3
aIn vivo studies were conducted in conscious normotensive Sprague±
Dawley rats with free access to water. The compounds under test were
dosed orally (po) at stated doses (20% dimethyl sulfoxide and 2.5%
pre-boiled starch preparation).
bData reported as a ratio of urine volume collected from treatment
group after 4 h versus control group.
Table 3. Model parenteral system solubilitya
No.
Aq solution
30% PEG mg/mL
4a
4h
3a
4f
0.72b
0.3
1.7c
2.2
References and Notes
4g
1.9b
1. Laszlo, F. A.; Laszlo, F., Jr.; De Weid, D. Pharmacol. Rev.
1991, 43, 73.
aValues reported for free bases unless noted. The samples were pre-
pared by dissolution of the material in an appropriate solvent system
on a wrist action stirrer for ca. 24±36 h. Each solution was ®ltered
through a 0.2 micro Nylon 66 ®lter disc and appropriate dilutions
made before injections were made on the HPLC system. Data reported
for crystalline materials as determined by di€erential scanning calori-
metry (DSC).
2. Albright, J. D.; Chan, P. S. Curr. Pharm. Des. 1997, 3, 615.
3. 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.; Manzandarani, H.; Bailey, T. J. Med. Chem. 1998, 41,
2442.
4. For general principles see Yalkowsky, S. H. Solubility and
Solubilization in Aqueous Media; Oxford University Press:
New York, 1999.
bMixture of amophorus and crystalline material.
cCitrate salt.
5. Venkatesan, A.; Albright, J. D.; Grosu, G. T.; Chan, P. S.;
Coupet, J.; Saunders, T.; Ru, X.; Manzandarani, H. Bioorg.
Med. Chem. Lett. 1999, 9, 1733.
6. Venkatesan, A.; Albright, J. D.; Grosu, G. T.; Delos San-
tos, E. G.; Chan, P. S.; Coupet, J.; Ru, X.; Saunders, T.;
Manzandarani, H. Bioorg. Med. Chem. Lett. 1999, 9, 1737.
7. See Trybulski, E. J.; Molinari, A. J.; Bagli, J. F.; Ashwell,
M. A.; Caggiano, T. J. US 5880122 for experimental details,
1999.
8. See Albright, J. D.; Venkatesan, A. M.; Dusza, J. P.; Sum,
F. US 5753648 for experimental details.
9. Brandage, S.; Luning, B.; Lundin, C. Acta. Chem. Scand.
1973, 27, 433.
10. von Troxler, F.; Stoll, Andre, P.; Niklaus, P. Helv. Chim.
Acta 1968, 51, 1870.
11. Gould, P. L. Inter. J. Pharmaceuticals 1986, 33, 201.
12. Albright, J. D.; Delos Santos, E. G.; Dusza, J. P.; Chan, P.
S.; Coupet, J. C.; Ru, X.; Manzandarini, H. Bioorg. Med.
Chem. Lett. in preparation.
they display low nanomolar binding and good oral
activity. It was disappointing, however, to again ®nd
that the physical properties of these [((10-biphenyl)-2-
ylcarbonyl)amino]-2-chlorobenzoic derived sca€olds
hampered their crystallization and hence solubility
determination. An approximate upper limit for the
solubility of 4a was established with partially crystalline
material (entry 4a in Table 3) and as this was below the
target value it indicated that further increases in solubi-
lity would be needed. It was postulated that this
increase in solubility could be achieved by the intro-
duction of additional hydrogen bonding groups. A ser-
ies of substituent modi®cations were investigated at R1
and R2, and, indeed the required increase in solubility
occurred when oxygen replacements were incorporated