Discovery of PF-184563, a potent and selective V1a antagonist for the treatment of dysmenorrhoea. the influence of compound flexibility on microsomal stability

Article abstract of DOI:10.1016/j.bmcl.2011.08.038

The V1a receptor has emerged as an attractive target for a range of indications including Raynaud's disease and dysmenorrhoea. As part of an effort to discover a new class of orally active V1a antagonist, we optimised a highly lipophilic, metabolically unstable lead into a range of potent, selective and metabolically stable V1a antagonists. In this communication, we demonstrate the series-dependent effect of limiting the number of rotatable bonds in order to decrease Cytochrome P450-mediated metabolism. This effort culminated in the discovery of PF-184563, a novel, selective V1a antagonist with excellent in vitro and in vivo properties.

Full text of DOI:10.1016/j.bmcl.2011.08.038

Bioorganic & Medicinal Chemistry Letters 21 (2011) 5684–5687
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of PF-184563, a potent and selective V1a antagonist
for the treatment of dysmenorrhoea. The influence of compound flexibility
on microsomal stability
Patrick S. Johnson a, Thomas Ryckmans a, , Justin Bryans a, Dave M. Beal a, Kevin N. Dack a, Neil Feeder b,
⇑
Anthony Harrison c, Mark Lewis c, Helen J. Mason a, James Mills a, Julie Newman a, Christelle Pasquinet a,
Dave J. Rawson a, Lee R. Roberts a, Rachel Russell d, Deborah Spark d, Alan Stobie a, Toby J. Underwood a,
Robin Ward d, Simon Wheeler a
a Pfizer Worldwide Medicinal Chemistry, Ramsgate Road, Sandwich Kent CT139NJ, UK
b Pfizer Pharmaceutical Development, Ramsgate Road, Sandwich Kent CT139NJ, UK
c Pfizer Department of Pharmacokinetics, Distribution and Metabolism, Ramsgate Road, Sandwich Kent CT139NJ, UK
d Pfizer Discovery Biology Ramsgate Road, Sandwich Kent CT139NJ, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The V1a receptor has emerged as an attractive target for a range of indications including Raynaud’s dis-
ease and dysmenorrhoea. As part of an effort to discover a new class of orally active V1a antagonist, we
optimised a highly lipophilic, metabolically unstable lead into a range of potent, selective and metabol-
ically stable V1a antagonists. In this communication, we demonstrate the series-dependent effect of lim-
iting the number of rotatable bonds in order to decrease Cytochrome P450-mediated metabolism. This
effort culminated in the discovery of PF-184563, a novel, selective V1a antagonist with excellent
in vitro and in vivo properties.
Received 11 July 2011
Accepted 6 August 2011
Available online 12 August 2011
Keywords:
V1a antagonists
Triazoles
Ó 2011 Elsevier Ltd. All rights reserved.
LiPE
Lead optimisation
Number of rotatable bonds
Vasopressin (AVP) is a small peptide that exerts its effects
through the V1a, V1b, V2 and Oxytocin (OT) receptors.1,2 The V1a
receptor is expressed inter alia in smooth muscles, and its over-
activation by elevated levels of AVP has been linked to both Ray-
naud’s disease3 and dysmenorrhoea. Clinical evidence shows that
blockade of the V1a receptor by Relcovaptan 1 alleviates the symp-
toms of both these conditions.3,4 As part of an effort to discover a
new agent for the treatment of dysmenorrhoea, the triazole 2 at-
tracted our interest as a novel V1a antagonist. An analog of this
compound had been disclosed as a V1a antagonist.5 (Fig. 1).
Despite modest potency and very high lipophilicity, compound
2 displayed a relatively high Ligand Efficiency6 (LE) owing to its
low molecular weight. Furthermore this chemotype is syntheti-
cally enabled, for example, each substituent on the triazole core
can be introduced from widely available monomers (carboxylic
acids and amines) using well-established methods. Early optimisa-
tion efforts on this lead focused on reducing the lipophilicity of the
series, as high lipophilicity has been linked to a range of issues in
development candidates, such as low solubility, high metabolic
clearance, promiscuity and overall toxicity.6,7
The first key modification in this series was the replacement of
the biaryl moiety of 2 with a 2-piperidyl pyridine group, resulting
in compound 3. Pleasingly, this compound proved to be only
slightly less potent than 2 with a similar LE, and a cLogP reduced
by 3 units. This translated into a much higher LiPE (4.4), a measure
of binding efficiency per unit of lipophilicity and of overall lead
quality.8 Using this optimised 2-piperidyl pyridine scaffold, deriv-
atives were prepared to further explore the SAR of the substituents
at the 3- and 5-positions of the triazole.
Chemistry: The synthesis of compound 3 (R1 = H) exemplifies
the route to our final compounds 7a–f and 8a–j. Formation of
bis-hydrazide 5 from the acid 4 was followed by a dehydrative cyc-
lisation to oxadiazole 6. Reaction of this oxadiazole with aniline
formed the final triazole 3 in good yields. The full experimental de-
tails for the preparation of these compounds have been dis-
closed.9,10 (Scheme 1).
Compounds 8a–d were similarly obtained from the correspond-
ing benzylamines; compound 8e was prepared using butyryl
hydrazide in step (a), while compounds 8f–j were prepared by
nucleophilic substitution from the corresponding 5-chloromethyl
⇑
Corresponding author. Tel.: +44 1304 643 735; fax: +44 1304 651 817.
E-mail address: Thomas.ryckmans@gmail.com (T. Ryckmans).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.08.038

P. S. Johnson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5684–5687
5685
2
N
1
N
Cl
HO
N
N
5
3
Cl
N
N
N
N
4
N
O
NH2
N
S
O
O
O
1 Relcovaptan
hIC50 13nM
MW 621
2
3
O
hIC50 200nM
MW 311
hIC50 345nM
MW 319
O
cLogP 4.4
LE 0.30
cLogP 5.3
LE 0.39
cLogP 2.3
LE 0.38
LiPE 4.5
LiPE 1.4
LiPE 4.4
Figure 1. Structure of Relcovaptan 1 and lead 2.
oxadiazole 9 (see Scheme 2) followed by step (c) using benzyl-
amine.9–11
trend) were not successful, possibly because of the limited range
of lipophilicity exemplified (85% of prepared compounds had cLogP
values between 1 and 3). Another molecular descriptor linked to
microsomal clearance is the number of rotatable bonds (NRB);
compounds with high NRB (high flexibility) are more likely to
adopt a conformation recognised by Cytochrome P450. Indeed, plot-
ting HLM clearance binned between low (Clint < 10), medium
(10 6 Clint 6 20) and high (Clint > 20) values against NRB values
showed that all stable compounds bar one (with a low cLogP of
0.6) had a NRB count below 5. Conversely, all compounds with a
NRB > 7 displayed a high clearance, even at low cLogP values.
(Fig. 2)
This analysis, albeit based on a limited dataset, led us to design
a series of compounds in which conformational restriction was
achieved by tethering the 3- and 5-substituents of the triazole.
Reaction of chloromethyl oxadiazole 9 with bis nucleophiles 10
afforded the adduct 11, which was cyclised under acidic conditions
to give tethered compounds 12 and 13. Full experimental details
for the preparation of these compounds have been disclosed.11–13
Cyclic ethers 12 proved mostly equipotent to their non-cyclic ana-
logs, but generally displayed a trend toward higher metabolic sta-
bility (compare 12a and 3, 12c and 7a, 12d and 7f). This trend can
be attributed to either (or both) the reduction in cLogP or NRB
count for compounds 12a–d compared with compounds 3 and
Results and discussion: Selected examples are presented here. For
N-phenyl derivatives 7, substituents such as 4-methoxy (7a),
4-methyl (7c) and 4-chloro (7f) proved beneficial to potency at the
V1a receptor, with a concomitant increases in LiPE. The 4-CN and
4-CF3 groups were poorly tolerated. Disappointingly, our most po-
tent compound 7f suffered from significant metabolic instability
as measured in Human Liver Microsomes (HLM). Next, a series of
N-benzyl derivatives was prepared. In this case, the 3-chloro com-
pound (8c) displayed interesting potency but with similar metabolic
liability to 7f. Modification of the side chain at the 5-position of the
triazole demonstrated that while basic (8h) and acidic (8i) groups
were poorly tolerated, both lipophilic and polar side chains were
beneficial to potency (compare 8a with 8e and 8f). The best LiPE va-
lue was achieved with compound 8j. Combination of the features of
8c (R1 = 3-Cl) and 8c (R2 = 2H-[1,2,3]triazolyl–) in 8k proved to be
additive in terms of potency, but also resulted in high metabolic
instability. All potent derivatives 8 proved metabolically unstable
despite their low cLogP values (Table 1).
In order to define a strategy to reduce microsomal clearance, we
sought to find a correlation between HLM clearance and physico-
chemical properties using a larger dataset (60 compounds). At-
tempts to correlate cLogP with HLM clearance (a well established
O
N
O
N
H
N
a
b
OH
O
N
H
N
N
O
N
N
N
N
6
4
5
c
N
N
N
N
N
N
R2
N
N
N
N
N
N
N
N
R1
R1
N
R1
7a-f
3 R1 = H
8e-8j R1=H
8k R1=Cl
8a-d
O
N
O
N
N
N
N
R2
O
*
N
N
O
N
N
OH
*
*
*
*
*
*
8e
8f
8g
8h
8i
8j
8k
Scheme 1. Typical procedure for the synthesis of compound 6. Reagents and conditions (a) oxalyl chloride (2 equiv), DCM, 25 °C, 16 h then acetylhydrazide (2 equiv), DCM,
25 °C, 16 h, 39%; (b) POCl3 (excess), 100 °C 2 h 95%; (c) aniline or benzylamine (4 equiv), anhydrous MgCl2 (0.3 equiv), 150 °C 18 h 75%, or xylene, pTSA (0.01 equiv), reflux
18 h.

5686
P. S. Johnson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5684–5687
N
Cl
N
N
N
NH2
X
O
O
N
a
N
X
N
R1
N
N
N
H2
9
10
11
R1
b
N
N
N
N
N
R2
X
O
N
N
N
N
N
N
N
2
R1
R1
12a-f
R1
13a-h
5
N
N
N
3
4
Scheme 2. Typical procedure for the synthesis of compounds 12 and 13. Reagents and conditions (a) X = OH: NaH, THF, 25 °C, 3 h, 72%, X = NH2, THF, 50°C, 18 h, 75%; (b)
xylene, pTSA (0.01 equiv), reflux 18 h, X = OH: 65%, X = NH:76%.
Table 1
Table 2
Antagonist activitya at the human V1a receptor
Antagonist activitya of cyclic derivatives 12 and 13 at the human V1a receptor
Compound
R1
hV1a
IC50 (nM)
cLogP
LiPE
HLM
Clint
Compound
R1
R2
hV1a
IC50 (nM)
cLogP
LiPE
HLM
Clint
b
b
2
3
—
H
200
345
133
199
71
2310
580
9
117
83
29
1450
17
20
5.3
2.3
2.4
2.8
2.8
1.8
3.2
3.1
1.8
2.3
2.5
2.3
3.4
1.7
1.3
1.8
0.8
1.1
1.8
1.4
4.4
4.8
4.2
4.6
4.2
3.3
5.3
5.4
5.1
5.4
3.9
4.7
6.3
6.7
4.7
5.1
7.0
6.9
n/a
<8
11
33
40
<8
n/a
23
<8
23
29
19
154
23
34
n/a
<8
12a
12b
12c
12d
12e
12f
13a
13b
13c
13d
13e
13f
—
—
—
—
—
—
—
H
H
345
20
135
17
5380
904
4480
110
272
6.7
1.0
1.5
1.0
1.7
1.7
1.7
1.0
1.7
1.4
2.1
2.7
1.3
0.7
2.5
5.8
6.5
6.1
6.4
3.9
4.7
4.7
5.6
5.5
6.3
6.1
8.3
7.9
6.7
<8
11
<8
15
<8
32
n/a
n/a
n/a
14
n/a
182
9
4-Me
4-OMe
4-Cl
3-Cl
5-Cl
—
Cl
—
Cl
Cl
7a
7b
7c
7d
7e
7f
8a
8b
8c
8d
8e
8f
8g
8h
8i
4-OMe
2-Me
4-Me
4-CN
4-CF3
4-Cl
H
3-Me
3-Cl
4-Me
H
H
H
H
H
H
Cl
Me
Me
CO2Me
SO2NMe2
COMe
SO2Me
3.2
0.5
4.8
1.3
Cl
Cl
Cl
13g
13h
n/a
25
a
Values are geometric means of a minimum of two experiments.
In vitro Human Liver Microsomal Intrinsic Clearance, expressed in
609
2460
18
b
l
L/min/mg.
8j
8k
19
104
3
a
These results prompted us to prepare cyclic amines 13. Unsub-
Values are geometric means of a minimum of two experiments.
In vitro Human Liver Microsomal Intrinsic Clearance, expressed in
b
lL/min/mg.
stituted derivative 13a was surprisingly weak at the V1a receptor,
but addition of a 4-Chloro substituent improved potency 40-fold
(13b). N-Methylation (13c) also increased potency 16-fold. Gratify-
ingly, the combination of these features in 13d proved additive,
with a 670-fold increase of potency (16 ꢀ 40 = 640) between 13a
and 13d. This compound also displayed adequate metabolic stabil-
ity. Further N-substituted derivatives 13e–h were prepared, result-
ing in a range of very potent compounds (Table 2).
The single crystal X-ray structures of 13d and 13g were solved,
the two structures differing mainly in torsion angles across the two
rotatable bonds14 (Fig. 3).
Compounds 13d, 13e, 13g, and 13h were further evaluated in a
panel of in vitro assays. No significant inhibition of Cytochrome
P450 (IC50 values >30
or blockade of the hERG channel15 were observed. The compounds
showed little binding to V1b or OT receptors (>10 M) (Table 3).
lM for CYP 1A2, 2C9, 2C19, 2D6 and 3A4)
l
While compounds 13d, 13g, and 13h displayed weak binding to
the hV2 receptor, 13e and 13h proved to be potent agonists of this
receptor and not progressed further. The 40-fold V2-V1a selectivity
window for 13d and 13g, while not optimal, was deemed accept-
able at the time. Good cell permeability was predicted for both
13d and 13g using the Caco-2 model (Papp A/B and B/A of 29/28
and 17/27 10ꢁ6 cm/sec respectively). Compounds 13d and 13g
were profiled over 70 targets (Cerep Panel), producing less than
Figure 2. Relationship between NRB and HLM Clint. Individual points have been
jittered to avoid superposition. Legend: Blue: cLogP < 1.5, Green, cLogP between 1.5
and 3, Red: cLogP>3.
7a–f. 4-substitution was beneficial to potency, but chlorine atoms
in the 3- or 5-positions were poorly tolerated (12d vs 12d and 12f)
(Scheme 2).
30% inhibition of binding at 10 lM on all targets. Finally, both

P. S. Johnson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5684–5687
5687
N-oxide, but minor N-demethylation was also observed. Human
PK estimations using single-specie scaling17 based on rat PK data
predicted good absorption (F = 75%) and half-life of 3.5 h,18 suitable
for oral administration using controlled release (Table 4).
In conclusion, we have reported the discovery of PF-184563, a
potent, selective V1a antagonist suitable for human studies.
Modifications of a highly lipophilic substrate allowed us to identify
a new class of polar V1a antagonists. A relationship between
molecular flexibility (as measured by the number of rotatable
bonds) and in vitro metabolic stability was identified, resulting in
the design of a class of rigid, tethered compounds 13. This effort
culminated in the discovery of PF-184563 (13d), a potent, selective
V1a antagonist suitable for human clinical studies. The full phar-
macological profile of PF-184563 will be presented separately.
Supplementary data
Figure 3. Single crystal X-ray structures of 13d (green) and 13g (cyan).
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.08.038.
Table 3
Activity of cyclic derivatives 13 at the hV1a receptor, the hV2 receptor (in agonist and
References and notes
antagonist modes) and the hERG channel
Compound
hV1a
IC50 (nM)
hV2
EC50 (nM)
hV2
IC50 (nM)
hERG
IC50 (nM)
1. Ryckmans, T. Curr. Opin. Drug Disc. Dev. 2010, 13, 538.
2. Ryckmans, T. Annu. Rep. Med. Chem. 2009, 44, 129.
a
a
a
b
3. Hayoz, D.; Bizzini, G.; Noël, B.; Depairon, M.; Burnier, M.; Fauveau, C.; Rouillon,
A.; Brouard, R.; Brunner, H. R. Rheumatology 2000, 39, 1132.
4. Brouard, R.; Bossmar, T.; Fournié-Lloret, D.; Chassard, D.; Åkerlund, M. BJOG:
Int. J. Obstetrics Gynaecol. 2000, 107, 614.
5. Kakefuda, A.; Suzuki, T.; Tobe, T.; Tsukada, J.; Tahara, A.; Sakamoto, S.;
Tsukamoto, S. J. Med. Chem. 2002, 45, 2589.
6. Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc. 2007, 6, 881.
7. Hughes, J. D.; Blagg, J.; Price, D. A.; Bailey, S.; DeCrescenzo, G. A.; Devraj, R. V.;
Ellsworth, E.; Fobian, Y. M.; Gibbs, M. E.; Gilles, R. W.; Greene, N.; Huang, E.;
Krieger-Burke, T.; Loesel, J.; Wager, T.; Whiteley, L.; Zhang, Y. Bioorg. Med.
Chem. Lett. 2008, 18, 4872.
8. Ryckmans, T.; Edwards, M. P.; Horne, V. A.; Correia, A. M.; Owen, D. R.;
Thompson, L. R.; Tran, I.; Tutt, M. F.; Young, T. Bioorg. Med. Chem. Lett. 2009, 19,
4406.
9. Armour, R. D.; Baxter, A. D.; Bryans, J. S.; Dack, K. N.; Johnson, P. S.; Lewthwaite,
R. A.; Newman, J.; Rawson, D. J.; Ryckmans, T. Application: WO 2004037809;
2004.
10. Bryans, J. S.; Johnson, P. S.; Ryckmans, T.; Stobie, A. WO 92,005,079,808, 2005.
11. Ellis, D.; Johnson, P. S.; Nortcliffe, A.; Wheeler, S. Synth. Commun. 2010, 40,
3021.
12. Beal, D. M.; Bryans, J. S.; Johnson, P. S.; Newman, J.; Pasquinet, C.; Peakman, T.
M.; Ryckmans, T.; Underwood, T. J.; Wheeler, S. Submitted to Tetrahedron Lett.
2011.
13. Bryans, J. S.; Johnson, P. S.; Ryckmans, T.; Stobie, A. WO2004074291, 2004.
14. The following crystal structures have been deposited at the Cambridge
Crystallographic Data Centre and allocated the deposition numbers CCDC
836796 & 836797.
13d
13e
13g
13h
6.7
3.2
4.8
1.3
265
31
215
13
8490
n/a
4500
715
17,900
47,300
35,200
30,700
a
Values are geometric means of a minimum of two experiments.
Values are obtained from a single experiment.
b
Table 4
In vivo PK parameters of PF-184536 (13d) following iv and po administration of 2 mg/
kg (n = 2, Sprague–Dawley rats) and single-specie scaling predictions for human PK
parameters
Parameter
Measured values (rat)
Predicted values (human)
Tmax (h)
Cmax (ng/mL)
F (%)
Cl (mL/min/kg)
Vd (L/kg)
t1/2 (h)
0.8
97
34
—
—
75%
6.4
2.0
3.5
77%a
51
2.7
1.8
69%
Ppb (% bound)
a
Measured human plasma protein binding value.
15. Singleton, D. H.; Boyd, H.; Steidl-Nichols, J. V.; Deacon, M.; de Groot, M. J.; Price,
D.; Nettleton, D. O.; Wallace, N. K.; Troutman, M. D.; Williams, C.; Boyd, J. G. J.
Med. Chem. 2007, 50, 2931.
16. Aubrecht, J.; Ku, W. W.-T.; Osowski, J. J. Application: EP 1217077; 2002.
17. Hosea, N. A.; Collard, W. T.; Cole, S.; Maurer, T. S.; Fang, R. X.; Jones, H.; Kakar, S.
M.; Nakai, Y.; Smith, B. J.; Webster, R.; Beaumont, K. J. Clin. Pharmacol. 2009, 49,
513.
18. Single-specie scaling of rat unbound clearance was carried out using
correlations of human/rat unbound clearance values for 44 drugs using the
following equation:
compounds were evaluated in the Ames16 and the CHO in vitro
micronucleus assays. While compound 13g was positive in the
micronucleus assay, 13d (PF-184536) did not induce microbial
gene mutations or chromosomal abnormalities in vitro. The phar-
macokinetic properties of this compound were measured in the
rat (2 mg/kg), demonstrating the favorable profile of PF-184536.
There was no significant biliary or renal clearance of PF-184563
in bile duct cannulated rats. The major metabolic route in rat
was N-demethylation to 13b. The main metabolite of PF-184536
by human hepatocytes and microsomes was the inactive pyridine
Log ðClu humanÞ ¼ 0:9285 ꢀ LogðClu ratÞ ꢁ 0:6171ðr2 ¼ 0:89Þ
The human unbound volume is assumed to be the same as the animal species in
which the pharmacokinetics were measured.