Optimisation of pharmacokinetic properties to afford an orally bioavailable and selective V1A receptor antagonist

Article abstract of DOI:10.1016/j.bmcl.2011.05.092

The previously described lead compound 5 is a potent and selective V _1A antagonist with affinity at both the rat and human receptor, but displays poor oral bioavailability and moderate clearance. We report herein the successful optimisation of the pharmacokinetic (PK) properties to afford the potent, selective, orally bioavailable and CNS penetrant compound 15f. A custom optimisation approach was required which demonstrated the value of using early, rapid in vivo PK studies to show improvements in oral exposure. Such assays may be of particular value where low oral bioavailability is anticipated to be multifactorial (e.g., permeability, gut wall metabolism and/or transport) where satisfactory modelling of in vitro data is likely to be difficult within a drug discovery context.

Full text of DOI:10.1016/j.bmcl.2011.05.092

Bioorganic & Medicinal Chemistry Letters 21 (2011) 4622–4628
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Optimisation of pharmacokinetic properties to afford an orally bioavailable
and selective V1A receptor antagonist
William Arbuckle a, James Baker c, David Barn a, Matilda Bingham a, , Angus Brown a, Kirsteen Buchanan a,
⇑
Mark Craighead b, Richard Goodwin a, Susan Goutcher b, Michael Kiczun a, Amanda Lyons a, Rachel Milne b,
b
c
c
a
Brian Montgomery a, Susan Napier a, , Jeremy Presland , Hazel Sloan , Zara Turnbull , Grant Wishart
⇑
a Department of Medicinal Chemistry, MSD, Newhouse, Motherwell, ML1 5SH, United Kingdom
b Department of Molecular Pharmacology, MSD, Newhouse, Motherwell, ML1 5SH, United Kingdom
c Department of Pharmacology, MSD, Newhouse, Motherwell, ML1 5SH, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
The previously described lead compound 5 is a potent and selective V1A antagonist with affinity at both
the rat and human receptor, but displays poor oral bioavailability and moderate clearance. We report
herein the successful optimisation of the pharmacokinetic (PK) properties to afford the potent, selective,
orally bioavailable and CNS penetrant compound 15f. A custom optimisation approach was required
which demonstrated the value of using early, rapid in vivo PK studies to show improvements in oral
exposure. Such assays may be of particular value where low oral bioavailability is anticipated to be mul-
tifactorial (e.g., permeability, gut wall metabolism and/or transport) where satisfactory modelling of
in vitro data is likely to be difficult within a drug discovery context.
Received 5 April 2011
Revised 20 May 2011
Accepted 23 May 2011
Available online 30 May 2011
Keywords:
Peptide GPCR
Vasopressin antagonist
CNS penetration
P-glycoprotein
Ó 2011 Published by Elsevier Ltd.
Arginine vasopressin (AVP) is a 9 amino acid cyclic peptide
produced in the hypothalamus. It is released from the posterior
pituitary gland into the bloodstream or directly into the brain.
AVP exerts its effects through four different G-protein coupled
receptors (GPCRs), the vasopressin V1A, V1B, V2 receptors and the
oxytocin (OT) receptor. V1A receptors are localised in the liver (gly-
cogenolysis), vascular smooth muscle cells (vasoconstriction, uter-
ine blood flow, contraction) and the brain (stress adaptation,
memory formation, temperature, circadian rhythm). V1A receptors
are also localised in blood platelets and are involved in platelet
aggregation. Thus a selective V1A antagonist may have clinical util-
ity in dysmenorrhoea, pre-term labour, hypertension, Raynaud’s
disease, depression, anxiety, hyponatremia and congestive heart
failure.1 Relatively few small-molecule selective V1A antagonists
have progressed into clinical development, reflecting the chal-
lenges associated with the development of ligands for peptide
GPCRs in terms of combining potent receptor affinity and selectiv-
ity with drug-like properties suitable for oral delivery.3 The struc-
tures of representative selective V1A antagonists,2 Relcovaptan (SR
49059, 1)4 from Sanofi-Aventis, JNJ-17308616 (2)5 from Johnson
and Johnson and SRX-251 (3) and SRX-246 (4) from Azevan6 are gi-
ven in Figure 1. Molecular weights range from 554 to 772 with a
combined amide and sulfonamide bond-count of three for each
compound. These are not features typically associated with
drug-like properties nor CNS penetrant compounds however,
JNJ-17308616 (2) and SRX-251 (3) have shown CNS activity in
pre-clinical models.5,6
We have previously described the optimisation of a novel series
of peptide derived V1A antagonists to afford compound 5 (Fig. 2,
Table 1).7 Compound 5 is a potent and selective antagonist at both
the human and rat V1A receptors, however it displays suboptimal
DMPK properties; notably the compound has poor permeability
and is a substrate for P-glycoprotein (P-gp) leading to efflux in
the Caco-2 model (A to B (nm/s) <100, B to A/A to B efflux ratio
(ER) >4) and poor oral bioavailability in vivo. The impact of P-gp
on the brain penetration of 5 was confirmed in brain penetration
studies with P-gp KO mice (brain to plasma ratio increased 4.7-fold
in mdr1 a/b KO compared to wildtype-FVP mice).
The goal of the programme was thus to maintain the favourable
selectivity and affinity profile at the multi-targets, hV1A, rV1A, V1B
,
V2 and OT whilst addressing the unfavourable DMPK profile. We
had significant in-house expertise in the vasopressin area which
gave us confidence that we could understand the SAR required
for maintaining the desired selectivity profile. However, the
pseudopeptide nature of the starting lead, and the poor in vitro
DMPK properties of a number of analogues in the lead series, led
us to consider whether the chemotype would always have inher-
ently poor PK. We therefore decided that a defined and customised
⇑
Corresponding authors. Tel.: +44 1698736000; fax: +44 1698736187 (M.B.).
E-mail addresses: matilda.bingham@virginmedia.com (M. Bingham), susanna-
pier67@googlemail.com (S. Napier).
0960-894X/$ - see front matter Ó 2011 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2011.05.092

W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4622–4628
4623
N
O
NH
Cl
OH
Cl
N
O
NH2
O
N
O
N
N
O
N
N
N
O
N
N
O
O
O
S
N
N
O
O
O
O
F
O
N
O
HN
O
O
F3C
N
H
O
O
SRX-246, 4
SRX-251, 3
Figure 1. Structures of compounds 1–4.
Relcovaptan (SR49059), 1
JNJ-17308616, 2
The outline screening cascade which depicts our approach to-
wards improving oral bioavailability is given in Figure 3. The first
feature is the inclusion of custom physicochemical properties as
a guideline for improved absorption. These were derived from
analysis of Log D, molecular weight and Caco-2 data for a total of
389 compounds from the chemical series and were used in the de-
sign process. The criteria thus obtained (MWt<500; 1.5<c log D 7.4
<5, Fig. 3) were not applied in an absolute manner but as guidelines
to aid the chemistry effort towards increasing the likelihood of
obtaining permeable compounds. It is interesting to note that a re-
cent study of a much larger, structurally diverse dataset also iden-
tified Log D and molecular weight as being the most important
factors in determining permeability.8 The second feature of the
flow chart was the inclusion of the ‘rapid rat’ assay for assessing
oral exposure as a means to prioritise compounds for full PK.9a This
assay provides high throughput information on rat oral exposure
over 6 h and has good predictive power for identifying, and hence
filtering out, compounds which will have poor oral exposure de-
fined as rat F<20% (rapid rat AUC <500 h ng/mL).9b However, it does
not provide mechanistic information for the causes underlying the
low oral exposure, this latter information requires a more defined
follow up PK study (e.g., IV/PO) and in vitro absorption, distribu-
tion, metabolism, and excretion (ADME) studies. As depicted in
the screening cascade (Fig. 3) compounds with improved micro-
somal stability and Caco-2 data, but also any compounds with
hV1A pKi >8.0 were progressed to the rapid rat assay. This prag-
matic approach avoided too heavy a reliance on in vitro DMPK data
when complex permeability, metabolism and transport interplay
was anticipated at the gut wall and provided a more holistic ap-
proach to enable the best compound selection for full rat PK
studies.
A notable requirement of the screening cascade was prepara-
tion of 30–50 mg quantities of each compound at first synthesis.
This was deemed feasible since the synthetic routes into the target
compounds were trivial and the 30–50 mg quantity would still be
amenable to parallel synthesis and purification by LCMS and HPLC,
thus maintaining turnaround times. Scheme 1 illustrates the two
synthetic routes which were used to prepare analogues of the hit
compound 5. Both routes were amenable to parallel synthesis with
the final steps c, and f being carried out on 5–30 member com-
pound arrays. The compounds were prepared from chiral pool
starting materials 6 and 10 purchased from commercial sources.
The enantiomeric excess (ee) of the starting materials was con-
firmed by comparison with literature optical rotation and by chiral
SFC or HPLC comparison with the racemate. The maintenance of
enantiomeric purity in key final compounds (>90% ee) was deter-
mined by chiral analytical SFC or HPLC comparison with racemic
product.10 In examples where the amine R3NH2 was only available
O
O
HN
HN
Cl
N
H
5
Cl
hV1A pKi 9.25
rV1A pKi 8.38
hV1B pKi <5.0
hV2 pKi <5.0
hOT pKi <5.0
Figure 2.
Table 1
DMPK properties of compound 5
5a
CL (mL/min/kg)
Vss (L/kg)
t1/2 (h)
44
17.7
8.2
19
3.5
4
Cmax (ng/mL)
tmax (h)
F (%)
a
5% (v/v) cremophor EL in saline, male
wistar rat.
strategy for optimisation of the series was required with the focus
on early assessment of scope for optimisation of oral bioavailabil-
ity. We outline herein the strategy we implemented which ulti-
mately led to the successful identification of the orally
bioavailable, selective V1A antagonist 15f.
MWt<500 1.5<clogD7.4<5
Physicochemical properties
Caco-2: RLM/HLM: Rat heps
Caco-2 (A-B) >100nm/s
CLint <50 µL/min/mg
hV1A pKi > 8.0
Rapid Rat
F%
(i.v./p.o.)
Figure 3.

4624
W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4622–4628
R1 R2
CO2H
R4
Route A
R1 R2
R4
8
O
O
O
O
BocHN
R3
BocHN
R3
b)
c)
N
HN
OH
a)
H
N
H
7
6
9
R
R
R
R1 R2
CO2H
R1 R2
R1 R2
R4
R4
Route B
R4
O
O
O
O
O
HN
R3
8
HN
N
OMe
H2N
d)
e) f)
H
OMe
9
R
R
11
10
R
Scheme 1. Reagents and condition: (a) R3NH2, HBTU (1 equiv), DIPEA (3 equiv), DCM, rt 16 h, 65–90% (b) TFA:DCM 1:4, 0 °C to rt 16 h, 50–90% (c) HBTU (1 equiv), DIPEA
(3 equiv), DCM, rt 16 h, 30–90% (d) HOBt (1.2 equiv), EDCI (1.2 equiv), Et3N (1.2 equiv) 30–89% (e) TFA:H2O, 110 °C, 12 h 50–96% (f) R3NH2, HOBt (1.2 equiv), EDCI (1.2 equiv),
Et3N (1.2 equiv) 30–80%
and 14g–h. In contrast the alkyl and alcohol derivatives had mod-
erate Caco-2 permeability and lower efflux ratios, but suffered
N
from higher predicted intrinsic clearance. Alcohol 14j had the best
balance of microsomal stability and Caco-2 permeability which
was reflected in a rapid rat AUC of 2308 h ng/mL predictive of good
oral bioavailability. However, the rV1A affinity was universally poor
for the alcohol and alkyl analogues, the addition of the cyclopropyl
R1R2 group (14k) which had worked well to boost potency in the
lead had a modest beneficial effect on rat potency but this came
at the expense of microsomal stability and was detrimental to
hV1A binding potency. A breakthrough came in the identification
of the trans dimethylcyclohexane-1,4-diamine containing 14m,
this compound had excellent rat and human V1A affinity and mod-
erate predicted oral exposure in the rapid rat model.
Compounds 14m and 14j were selected for full PK in rat. In both
cases the rapid rat AUC predicted well for the improved oral expo-
sure compared with the starting lead 5. Compound 14j showed
moderate clearance but had good oral bioavailability (F 43%; CL
40 mL/min/kg; Vss 2.0 L/kg; t1/2 0.74 h, oral dose 10 mg/kg in 0.5%
(w/v) gelatin in 5% (w/v) mannitol, male Sprague–Dawley rat)
whilst 14m showed lower clearance and improved oral bioavail-
ability compared with the lead (F 17%; CL 24 mL/min/kg; Vss 9.0
L/kg; t1/2 5.6 h, oral dose 5 mg/kg in 0.5% (w/v) gelatin in 5%
(w/v) mannitol, male Sprague–Dawley rat). In the latter case it is
notable that 14m did not stand out as likely to have improved
PK based on the in vitro microsomal stability and Caco-2 data,
but was successfully identified in the rapid rat assay and pro-
gressed through the flow chart.
Further optimisation was thus separated into two avenues;
optimisation of alcohol 14j to establish whether there was scope
for optimisation of V1A affinity and microsomal stability, and opti-
misation of 14m to further improve the oral bioavailability
We first turned our attention to the SAR of the central amino
acid and terminal phenylacetic acid region of the molecule
(Table 3). Although diverse central amino acids were investigated,
containing alkyl, cycloalkyl, heterocyclic and various substituted
phenylalanine side chains, only modest changes such as in com-
pounds 15a and 15b were tolerated, with the starting p-chlorophe-
nylalanine remaining the best central amino acid in terms of hV1A
and rV1A potency. Changes to the terminal phenylacetic amide
NH2
a) b)
H2N
BocHN
13
12
Scheme 2. (a) (i) CH2O aq (3 equiv), MeOH:AcOH 40:1 1 h at rt, (ii) NaBH3CN
(2 equiv, portionwise 0 °C) (iii) 20 h rt. 74% (b) TFA:DCM 1:4 0 °C to rt 20 h, 91%.
as the racemate, the final compounds were prepared according to
Route B, with an additional preparative SFC separation of the
diastereomeric, single enantiomer products in the final step. The fi-
nal compounds thus prepared were labeled stereoisomer 1, and
stereoisomer 2 in order of elution. The acids 8 and amines R3NH2
were purchased from commercial suppliers, with the exception
of 13 which was prepared in 2 steps from commercially available
12 by reductive alkylation and Boc deprotection (Scheme 2).
The test compounds were evaluated for their ability to displace
the binding of tritium labelled arginine vasopressin ([3H]AVP) in
CHO cells expressing the human V1A receptor or rat V1A receptor
respectively.11a Binding affinities at human V1B, V2 and OT recep-
tors were measured in [3H]AVP or [3H]OT binding assays in CHO
cells. In all assays pKi values were determined from a minimum
of two independent experiments.
We have previously reported our early structure activity rela-
tionships (SAR) for the lead 5 which demonstrated that the series
was tolerant of broad SAR changes in the tropane region of the
molecule.7 Therefore we elected to prepare an initial array of di-
verse analogues with variation in this region (Table 2) including
amine, alcohol, and alkyl substituents to look at the effects on pre-
dicted oral exposure. Mindful of the literature precedent that there
are often significant differences in hV1A and rV1A SAR,12 all ana-
logues were screened in both binding assays (Table 2).
Representative compounds from the array are given in Table 2.
In general, the amine analogues all showed poor Caco-2 permeabil-
ity with the exception of 14f which contained a tertiary rather than
secondary amide. The rapid rat AUC tracked the Caco-2 with uni-
versally poor predicted oral bioavailability for analogues 14a–e

W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4622–4628
4625
Table 2
V1A binding data for analogues 14a–m
R1
R2
O
O
HN
R3
Cl
Cl
14a-m
R1
H
R2
H
R3
*
hV1A pKi
rV1A pKi
RR AUCb h ng/mL
HLM/RLM CLint
(l
L/min/mg)
Caco-2 A to B Papp (ER)d
a
c
14a
5
9.63
9.25
7.66
8.38
0
184
<12/42
58/55
Low (>2)
Low (2.6)
5
HN
N
H
NH
*
N
H
14b
H
H
9.21
7.23
40
<12/38
Low (>4.6)
Stereoisomer 1
NH
*
14c
14d
H
H
H
H
8.33
7.96
6.50
6.72
0
60/<12
39/142
Low (>3.4)
Low (4.2)
N
H
N
*
311
N
H
*
N
N
N
14e
14f
H
H
9.26
7.56
7.92
153
830
45/142
49/79
Low (>20)
Mod (2.0)
H
N
*
N
H
H
H
H
(82%)
N
14g
14h
6.77
7.05
(71%)
(89%)
305
—
18/47
49/182
Low (1.7)
Low (3.7)
*
N
H
5
*
N
H
14i
H
H
H
H
8.52
6.69
147
126/131
Mod (>1.0)
*
14j
14k
8.15
7.73
(82%)
7.01
2308
—
35/101
>270/>270
Mod (2.0)
Mod (2.0)
N
O H
N
5
5
H
14l
14m
H
H
9.74
10.40
8.09
8.77
161
1411
<12/13
43/82
Low (>20)
Low (10.8)
*
N
H
a
b
c
Values in parentheses are for % inhibition at 10 lM of test compound.
The RR AUC classification of predicted oral bioavailability has recently been described9,13: poor <500 h ng/mL, 500 < moderate < 2000 h ng/mL, Good > 2000 h ng/mL.
HLM, in vitro human liver microsome intrinsic clearance; RLM, in vitro rat liver microsome intrinsic clearance.
d
The following categories based on Caco-2 data were used to describe permeability: low < 100 nm/s; 100 < moderate < 200 nm/s; high > 200 nm/s
region were more productive; in general addition of lipophilic bulk
in R1 improved potency at hV1A with 15c and 15e (Table 3) being
two of the most potent analogues prepared. However, although we
had hoped that the increased c Log D of these two compounds would
also contribute to improved permeability, neither the Caco-2 data
nor the rapid rat AUC, were significantly improved. The best com-
pound from the series proved to be compound 15f which retained
potency at hV1A and rV1A and had a rapid rat AUC of 2916 h ng/mL
classed as good. Cross screening of 15f at the related receptors from
the vasopressin family confirmed it was a selective V1A antagonist
(V2 pKi <5.0, V1B pKi <5 and OT pKi 6.42). Furthermore, the functional
antagonism of compound 15f was confirmed in CHO cells expressing
the human V1A receptor in a calcium flux assay.11b
(Compare 15c and 15e, Table 3 with 16a and 16b Table 4). We
therefore elected to focus in more detail on the alcohol terminus
(Table 5). Modest changes to this region of the molecule had a dra-
matic effect on hV1A; addition of a single methyl group as in 17c
afforded over an order of magnitude improvement in hV1A potency
in line with the goal for the alcohol series. This compound also had
significantly improved rapid rat AUC, but was unfortunately only
weakly active at rV1A precluding further in vivo studies in rat. Con-
straint of the pendant propylalcohol chain into a cyclohexane ring
reduced the number of rotatable bonds and also improved the V1A
affinity (17e–h) relative to the starting compound 14j. The most
potent of these compounds 17h had much improved rV1A relative
to the other alcohols prepared, however the disconnect between
the hV1A and rV1A affinity and poor microsomal stability remained
a feature of the alcohol series which was thus deprioritised in fa-
vour of the amine 15f.
The SAR of the alcohol 14j proved divergent from the amine ser-
ies. Attempts to improve the V1A potency by geminal di-substitu-
tion of the terminal phenylacetic amide were unsuccessful

4626
W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4622–4628
Table 3
V1A binding assay results for compounds 15a–g
R1
HN
O
N
O
N
H
R2
15a-g
R1
R2
hV1A pKi
rV1A pKi
RR AUCa h ng/mL
HLM/RLM CLint
32/85
(lL/min/mg)
Caco-2 A to B Papp (ER)b
*
*
15a
15b
OMe
9.28
7.86
1095
Low (11.3)
Cl
Cl
F
7.20
6.61
8.68
—
38/58
47/46
Low (0.7)
Low (3.4)
*
15c
15d
15e
Cl
10.02
998
Cl
F
*
Cl
Cl
9.60
8.31
8.86
628
798
73/58
-
*
*
10.43
>270/175
Low (2.3)
F
F
15f
Cl
Cl
9.62
8.95
8.64
7.16
2916
306
37/66
Low (3.5)
Low (>20)
H
H
*
15g
<12/<12
F
a
The RR AUC classification of predicted oral bioavailability has recently been described9,13: poor < 500 h ng/mL , 500 < moderate < 2000 h ng/mL, good > 2000 h ng/mL.
The following categories based on Caco-2 data were used to describe permeability: low < 100 nm/s; 100 < moderate < 200 nm/s; high > 200 nm/s.
b
Amine 15f was therefore identified as the best compound from
metabolised by CYP3A4 in a panel of human P450s (data not
shown), so the concern was that lower oral bioavailability would
be seen in monkey compared to rat due to higher gut wall expres-
sion of CYP3A. Moderate systemic clearance was seen in both rat
and monkey and, encouragingly, an acceptable oral bioavailability
(20%) was observed in monkey.
The functional V1A antagonism of 15f was assessed in vivo by
the reversal of V1A mediated AVP induced blood pressure increases
in conscious rats as described in the literature.4a,15 Compound 15f
(4 mg/kg, iv) decreased the AVP pressor response by 4.4%, 91.1%,
76.2% and 80.1% at 0, 15, 30 and 60 min respectively post-drug
(n = 1).
In conclusion, we have described the discovery of compound
15f a potent, selective, CNS penetrant, and orally bioavailable V1A
antagonist, with good rV1A affinity. This compound is thus a useful
molecule for further in vivo target validation studies. The optimisa-
tion of this series of V1a antagonists has demonstrated the value of
using early, rapid in vivo PK studies to show improvements in oral
exposure. Such assays may be of particular value where low oral
the series and was selected for further PK profiling (Table 6). Com-
pound 15f had moderate clearance, which combined with the oral
bioavailability of 30%, suggested predominantly hepatic clearance
with good intestinal absorption and limited gut wall metabolism.
The improvement in apparent in vivo permeability and/or de-
creased intestinal metabolism compared with 5 was suggested
by the increased oral bioavailablity (4% vs 30%) without the lower-
ing of systemic clearance. Serial CSF studies were also performed
with 15f14 and this showed a CSF/free plasma AUC ratio of 0.6 (data
not shown), suggesting no active transport from the CNS. Although
5 and 15f had similar Caco-2 efflux ratios (2.6 and 3.5, respec-
tively), in vivo brain penetration data from either P-gp KO mice
or rat serial CSF studies indicated that the two compounds had dif-
ferential blood–brain barrier efflux. The improved oral bioavail-
ability and CNS penetration of 15f may both therefore be a
consequence of decreased active transport in vivo.
The PK of 15f was also examined in cynomolgus monkey, as
shown in Table 6. 15f had been shown to be predominantly

W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4622–4628
4627
Table 4
V1A binding data for analogues 16a–b
R1
HN
O
O
N
OH
H
Cl
16a-b
R1
hV1A pKi
rV1A pKi
RR AUCb h ng/mL
HLM/RLM CLint
35/101
(
lL/min/mg)
Caco-2 A to B Papp (ER)c
a
*
14j
8.15
(82%)
7.01
2308
Mod (2.0)
Mod (2.0)
Cl
*
14k
7.73
7.55
—
—
>270/>270
180/>270
Cl
*
16a
16b
(94%)
6.89
Mod (>1.23)
Mod (1.20)
Cl
*
8.15
—
>270/>270
F
a
b
c
Values in parentheses are for % inhibition at 10 lM of test compound.
The RR AUC classification of predicted oral bioavailability has recently been described9,13: poor < 500 h ng/mL , 500 < moderate < 2000 h ng/mL , good > 2000 h ng/mL.
The following categories based on Caco-2 data were used to describe permeability: low < 100 nm/s; 100 < moderate < 200 nm/s; high > 200 nm/s.
Table 5
V1A binding assay results for compounds 17a–i
O
O
HN
R1
Cl
Cl
17a-i
R1
hV1A pKi
rV1A pKi
RR AUCb h ng/mL
HLM/RLM CLint
35/101
(lL/min/mg)
Caco-2 A to B Papp (ER)c
a
*
N
H
O H
OH
14j
-
8.15
(82%)
2308
Mod (2.0)
17a
17b
Stereoisomer 1
Stereoisomer 2
8.56
8.68
6.94
6.77
-
-
72/243
77/>270
Mod (2.5)
Mod (2.4)
*
*
N
H
17c
17d
Stereoisomer 1
Stereoisomer 2
9.43
7.84
(84%)
(64%)
5556
-
>270/205
77/>270
Mod (2.0)
Mod (2.4)
N
H
OH
17e
17f
Stereoisomer 1
Stereoisomer 2
8.37
9.17
(76%)
6.89
257
836
>270/>270
190/>270
Mod (2.8)
Low (4.1)
*
N
H
O H
OH
17g
-
8.75
6.89
139/252
Low (3.5)
*
*
N
H
17h
17i
Stereoisomer 1
Stereoisomer 2
9.68
8.03
7.40
(63%)
132
1603
>270/>270
>270/>270
High (1.5)
High (1.3)
N
H
OH
a
b
c
Values in parentheses are for % inhibition at 10 lM of test compound.
The RR AUC classification of predicted oral bioavailability has recently been described9,13: poor < 500 h ng/mL , 500 < moderate < 2000 h ng/mL , good > 2000 h ng/mL
The following categories based on Caco-2 data were used to describe permeability: low < 100 nm/s; 100 < moderate < 200 nm/s; high > 200 nm/s

4628
W. Arbuckle et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4622–4628
Kaldor, S. W.; Skelton, J. J.; Dressman, B. A.; Clay, M. P.; Steinberg, M. I.; Bruns,
Table 6
R. F.; Simon, N. G. Bioorg. Med. Chem. 2007, 15, 2054; (b) Ferris, C. F.; Lu, S.-F.;
Messenger, T.; Guillon, C. D.; Heindel, N.; Miller, M.; Koppel, G.; Bruns, F. R.;
Simon, N. G. Pharmacol. Biochem. Behav. 2006, 83, 169.
PK parameters of compound 15f in male rat and monkey
O
O
N
7. Napier, S.; Wishart, G.; Arbuckle, W.; Baker, J.; Barn, D.; Bingham, M.; Brown,
A.; Byford, A.; Claxton, C.; Craighead, M.; Buchanan, K.; Fielding, L.; Gibson, L.;
Goodwin, R.; Goutcher, S.; Irving, N.; MacSweeney, C.; Milne, R.; Mort, C.;
Presland, J.; Sloan, H.; Thomson, F.; Turnbull, Z.; Young, T. Bioorg. Med. Chem.
Lett. 2011, 21, 3163.
HN
F
N
H
8. Waring, M. J. Bioorg. Med. Chem. Lett. 2009, 19, 2844.
15f
9. (a) Korfmacher, W.; Cox, K. A.; Ng, K. J.; Veals, J.; Hsieh, Y.; Wainhaus, S.; Broske,
L.; Prelusky, D.; Nomeir, A.; White, R. E. Rapid Commun. Mass Spectrum. 2001,
15, 335; (b) Mei, H.; Korfmecher, W.; Morrison, R. The A.A.P.S. Journal 2006, 8,
E493.
Cl
Sprague–Dawley rata
Cynomolgus monkeyb
10. HPLC analysis was performed on Perkin Elmer Series 200 HPLC instrument
with Series 200 DAD detection with either a Chiralpak ADH, Chiralpak ASH,
Chiralpak IA, Chiralcel ODH or Chiralcel OJH column, all 25 ꢀ 0.46 cm diameter
and run at 1 mL/min flow rate. Mobile phases were optimised individually for
each compound from MeOH, EtOH, iPrOH, isohexane, heptane. SFC analysis
was performed on a Berger Minigram SFC instrument from Waters/Thar with
either a Chiralpak ADH, Chiralpak ASH, Chiralpak IA, Chiralcel ODH or Chiralcel
OJH column with detection at 230 nm and flow rate of 4 mL/min at 100 bar
CO2. Co-eluants were optimised individually for each compound from MeOH,
EtOH or iPrOH.
CL (mL/min/kg)
Vss (L/kg)
t1/2 (h)
Cmax (ng/mL)
tmax (h)
F
49
23
11.8
3.1
35
5.8
30%
6.4
4.8
41
2.7
20%
a
IV – 30% DMA in saline, 1 mg/kg; PO – 0.5% (w/v) gelatin in 5% (w/v) mannitol,
5 mg/kg.
b
IV – 30% DMA in saline, 1 mg/kg; PO – 0.5% (w/v) gelatin in 5% (w/v) mannitol,
11. (a) CHO cells (V1A. V1B and OT) or membranes prepared from CHO cells (V2)
expressing the human (hV1A. hV1B and hOT) or rat (rV1A) versions of the
receptors were incubated with tritiated ligand ([3H] AVP for V1A, V1B and V2;
[3H]oxytocin for OT) and increasing concentrations of the compound.
3 mg/kg.
Nonspecific binding was determined in the presence of 10
lM lysine
vasopressin for V1A, V1B and V2; 10 M oxytocin for OT. The concentration of
l
compound was plotted against specific binding and data were analysed by
non-linear regression to calculate affinities of the compound for each receptor.
(b) CHO cells expressing the human V1A receptor were loaded with the
calcium-sensitive fluorescent indicator Fluo-3. Cells were pre-incubated for
30 min with increasing concentrations of test compound prior to the addition
of 10nM AVP and fluorescence detection using a FLIPRtetra. The concentration
of compound was plotted against effect and data were analysed by non-linear
regression to determine the pIC50 of the compound for the receptor.
bioavailability is anticipated to be multifactorial (e.g., permeabil-
ity, gut wall metabolism and/or transport) where satisfactory mod-
elling of in vitro data is likely to be difficult within a drug discovery
context. In such cases, prioritisation of compounds for additional
in vivo studies may be facilitated with screens such as the Rapid
Rat assay.
12. (a) Shinoura, H.; Take, H.; Hirasawa, A.; Inoue, K.; Ohno, Y.; Hashimoto, K.;
Tsujimoto, G. FEBS Lett. 2000, 466, 255; (b) Thibonnier, M.; Coles, P.; Conarty, D.
M.; Plesnicher, C. L.; Shoham, M. J. Pharmacol. Exp. Ther. 2000, 294, 195.
13. The ‘Rapid Rat’ assay was performed by dosing male Sprague–Dawly rats
(n = 2) with test compound at 10 mg/kg orally using 5 mL/kg in 20% HP-b-CD.
Blood was taken at 0.5, 1, 2, 4 and 6 h and resulting plasma was analysed by
LC–MS/MS with quantitation against a standard curve.
14. Male Wistar BRL rats (Harlan) were surgically prepared with a cannula for CSF
sampling from the cisterna magna. Three days after surgery, four rats were
dosed with test compound at 20 mg/kg by oral gavage. Serial CSF samples
References and notes
1. (a) Thibonnier, M.; Coles, P.; Thibonnier, A.; Shoham, M. Annu. Rev. Pharmacol.
Toxicol. 2001, 41, 175; (b) Ring, R. H. Curr. Pharm. Des. 2005, 11, 205.
2. (a) Ryckmans, T. Curr. Opin. Drug Discovery Dev. 2010, 13, 538; (b) Rykmans, T.
Ann. Rep. Med. Chem. 2009, 44, 129.
3. Morphy, R.; Rankovic, Z. J. Med. Chem. 2006, 49, 4961.
4. (a) Serradeil-Le Gal, C.; Wagnon, J.; Garcia, C.; Lacour, C.; Guiraudou, P.;
Christophe, B.; Villanova, G.; Nisato, D.; Maffrand, J. P.; Le Fur, G.; Guillon, G.;
Cantau, B.; Barberis, C.; Trueba, M.; Ala, Y.; Jard, S. J. Clin. Invest. 1993, 92, 224;
(b) Serradeil-Le Gal, C.; Herbert, J. M.; Delisee, C.; Schaeffer, P.; Raufaste, D.;
Garcia, C.; Dol, F.; Marty, E.; Maffrand, J. P.; Le Fur, G. Am. J. Physiol.-Heart 1995,
(30 lL) and tail vein blood samples were taken at 0.5, 1, 2 and 4 h. Additionally,
brains were harvested at 4 h and also at 2 h in a non-surgically prepared
satellite group. Plasma was separated from blood by centrifugation. Brain
samples were homogenized with 3 volumes of ice-cold PBS, pH 7.4, and were
protein precipitated prior to analysis. CSF samples were quantified against
standards prepared in artificial CSF. Analysis was by LC–MS/MS for plasma,
brain and CSF samples.
268, H404; (c) Brouard, R.; Bossmar, T.; Fournié-Lloret, D.; Chassard, D.;
0
ÅAkerlund, M. Br. J. Obstet. Gynaecol. 2000, 107, 614.
5. Bleickardt, C. J.; Mullins, D. E.; MacSweeney, C. P.; Werner, B. J.; Pond, A. J.;
Guzzi, M. F.; Martin, F. D. C.; Varty, G. B.; Hodgson, R. A. Psychopharmacology
2009, 202, 711.
6. (a) Guillon, C. D.; Koppel, G. A.; Brownstein, M. J.; Chaney, M. O.; Ferris, C. F.; Lu,
S.-F.; Fabio, K. M.; Miller, M. J.; Heindel, N. D.; Hunden, D. C.; Cooper, R. D. G.;
15. Tskuda, J.; Tahara, A.; Tomura, Y.; Kusayama, T.; Wada, K.; Ishii, N.; Taniguchi,
N.; Suzuki, T.; Yatsu, T.; Uchida, W.; Shibasaki, M. Vasc. Pharmacol. 2005, 42, 47.