Discovery and optimisation of 1-acyl-2-benzylpyrrolidines as potent dual orexin receptor antagonists

Article abstract of DOI:10.1039/c5md00074b

Starting from a thienopiperidine lead compound with high intrinsic clearance in rat and human liver microsomes and low aqueous solubility, a novel series of 1-acyl-2-benzylpyrrolidines were discovered as potent and competitive dual orexin receptor antagon

Full text of DOI:10.1039/c5md00074b

MedChemComm
View Article Online
View Journal
CONCISE ARTICLE
Discovery and optimisation of 1-acyl-2-
benzylpyrrolidines as potent dual orexin receptor
Cite this: DOI: 10.1039/c5md00074b
antagonists†
*
Jodi T. Williams, John Gatfield, Catherine Roch, Alexander Treiber, Francois Jenck,
Martin H. Bolli, Christine Brotschi, Thierry Sifferlen, Bibia Heidmann and
Christoph Boss
Starting from a thienopiperidine lead compound with high intrinsic clearance in rat and human liver
microsomes and low aqueous solubility, a novel series of 1-acyl-2-benzylpyrrolidines were discovered as
potent and competitive dual orexin receptor antagonists. Metabolic stability was improved to afford oral
exposure, and aqueous solubility was increased by twenty fold, providing compounds suitable for pre
clinical evaluation. Compound 27 showed insurmountable antagonism at both orexin 1 and orexin 2 recep-
tor sub types and displayed a comparable sleep-promoting effect in the rat to almorexant and suvorexant.
Received 20th February 2015,
Accepted 14th April 2015
DOI: 10.1039/c5md00074b
www.rsc.org/medchemcomm
Intracerebellar orexin A and orexin B infusion in rats results
in enhanced arousal, delayed onset of REM sleep, and main-
Introduction
The vertebrate orexin neuropeptide–receptor system plays a
pivotal role in the regulation of sleep and wake states as well
as emotional states related to stress or reward. The neuropep-
tides orexin A and orexin B are synthesised by a small number
of neurons in the lateral hypothalamus, a brain region
involved in arousal, emotional and metabolic regulation and
motivated behaviours, e.g. feeding.1–4 Orexin peptides are
released at axonal terminals and – pre- or post-synaptically –
bind and activate two closely related G protein coupled recep-
tors (GPCR): orexin receptor type 1 (OX1) and orexin receptor
type 2 (OX2).5–7 Activated neuronal orexin receptors couple to
the Gq/phospholipase C/protein kinase C pathway resulting
in cellular depolarisation and increased cytosolic Ca2+ concen-
trations.8 Thus, orexin receptor signaling is excitatory by
enhancing synaptic transmission.
The distribution of OX1 and OX2 in mammalian brains is
indicative of their important role in the regulation of vigi-
lance states and circadian activity. Orexin-secreting neurons
of the lateral hypothalamus project to the basal forebrain,
corticolimbic structures, and brainstem, especially to those
regions related to sleep/wake regulation (locus coeruleus,
raphe nucleus, tuberomammillary nucleus), regions activated
in anxiety/stress-related conditions (paraventricular nucleus)
as well as regions involved in reward processing and drug
abuse (nucleus accumbens, ventrotegmental area).1,2,9–18
tenance of cortical activation. Pharmacological inhibition of
the orexin system in animal models of insomnia, stress/anxi-
ety and drug abuse has demonstrated a central role of an
overactive orexin system in these pathologies and suggests
orexin receptors as therapeutic targets in insomnia, stress/
anxiety-related disorders and addiction.19–30 Extensive clinical
trials with two dual orexin receptor antagonists (DORAs) have
demonstrated that targeting the orexin system is an effective
strategy in treating sleep disorders. In insomnia patients,
both almorexant and suvorexant dose-dependently increased
sleep efficiency by decreasing latency to persistent sleep and
wake after sleep onset.31–33 Suvorexant received FDA market-
ing authorization in 2014 and represents the first-in-class
dual orexin receptor antagonist for the treatment of insomnia
characterised by difficulties with sleep onset and/or sleep
maintenance.34,35 Studies into the respective contributions of
pharmacologically blocking OX1 and OX2 on sleep–wake
states have revealed a more important role for OX2.36 Recent
reports describing OX2 subtype-selective antagonists suggest
that blockade of OX2 alone may be suitable for the treatment
of sleep disorders, and future clinical studies should shed
light on whether there is an advantage of selectively blocking
OX2 over dual OX1 and OX2 blockade.37,38
The ideal profile of a sleep drug from a pharmacokinetic
perspective is difficult to achieve. Metabolic stability in par-
ticular needs to be carefully assessed; too stable runs the risk
of overshooting with the pharmacodynamic effect and could
lead to “hangover” phenomena. Too unstable, on the other
hand, will lead to lower bioavailability and higher doses may
be required to achieve the required duration of action. The
Drug Discovery Chemistry, Actelion Pharmaceuticals Ltd., Gewerbestrasse 16,
CH-4123, Switzerland. E-mail: jodi.williams@actelion.com
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c5md00074b
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun.

View Article Online
Concise Article
MedChemComm
latter is then associated with a greater risk for safety findings. Results and discussion
It is increasingly recognised that the rate at which a drug
The efforts described herein began with racemic
N-acylthienopiperidine 1 (Scheme 1), a dual orexin receptor
antagonist that we disclosed earlier.40 DORA 1 showed single
digit nM potency at both OX receptors but suffered from poor
in vitro ADME and physicochemical properties (Table 1). Our
initial investigation centred on reducing the lipophilicity of
DORA 1 as a means of modulating the high in vitro metabolic
clearance as determined experimentally in the presence of rat
and human liver microsomes (RLM, HLM).
As we considered the thiophene moiety present in 1 a met-
abolic liability,41 we eliminated this ring leading to com-
pound 2 (Scheme 1). The removal of the thiophene ring
resulted in only a modest 5-fold loss in potency at OX1; how-
ever, a more pronounced 70-fold loss in potency at OX2 was
observed. In the thienopiperidine series, the indole moiety
was required to obtain low nM potency at both OX receptors.
Attempts to replace it with an ortho-biaryl motif, a common
feature of many orexin receptor antagonists,42–45 to give com-
pound 3, led to a 24-fold and a 300-fold loss in potency at
OX1 and OX2, respectively. Having removed the thiophene
from 1, we explored the possibility that we would have
greater flexibility in terms of acylating the piperidine
associates with and dissociates from its target receptor – its
binding kinetics – directly affects the drug's efficacy and
safety. In the context of orexin receptor antagonism and
sleep, a desired compound profile may be characterised by
relatively slow receptor dissociation kinetics eliciting insur-
mountable antagonism to allow it to also be effective with
increasing orexin peptide concentrations. These two indepen-
dent features of a potential novel sleep drug, namely pharma-
cokinetic profile and binding kinetics, need to be carefully
evaluated to identify the appropriate balance.
In this manuscript, we describe the structural optimisa-
tion of our initial lead compound 1 into the in vivo active
compound 27. During the optimisation process, we focus on
OX2 potency and monitor the lipophilic ligand efficiency
(LLE) metric to assess design progress.39 DORA 27 shows
single digit nM potency at both OX receptors and exhibits
insurmountable antagonism in calcium release assays at
OX1 and OX2. An investigation into the binding kinetics of 27
reveals an estimated receptor occupancy half-life of 1–5 min
at both orexin receptor subtypes. DORA 27 decreases wakeful-
ness and increases sleep efficiency in male Wistar rats after
oral administration of 100 mg kg−1.
Scheme 1 Initial strategy for reducing lipophilicity of lead compound 1.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2015

View Article Online
MedChemComm
Concise Article
Table 1 Properties of racemic thienopiperidine lead compound 1
a
hOX1 IC50
hOX2 IC50
5.4 nM
1.4 nM
MW
460
5.5
3.5
a
c log P
HLM CLint
RLM CLint
>1250 μL min−1 mg−1
>1250 μL min−1 mg−1
3 μM
LLEd
b
b
Solubilitye
7 μg mL−1 (pH 4.1)
CYP3A4c IC50
a
b
Geomean of at least 3 experiments, standard deviation is <50% in all cases. Intrinsic metabolic clearance with liver microsomes, at 1 μM.
c
e
Testosterone 6β-hydroxylation used as a P450 isoform-specific marker. d Calculated from hOX2 IC50
.
Water, unbuffered.
nitrogen and we prepared the piperidine analog of com-
pound 3. To our delight, this modification resulted in DORA
4 showing single digit nM potency at both OX receptors.
The LLE metric is used to capture the efficiency of
improved in vitro potency in relation to change in lipophilic-
ity: LLE = pIC50 − c log P. It has been proposed that an accept-
able target LLE for a drug candidate lies in the region >5.46
The modifications leading to 4 clearly were steps in the right
direction in terms of addressing the poor physicochemical
properties highlighted for DORA 1. LLE was improved (4.8 vs.
3.5), molecular weight was lower (420 vs. 460) and there was
a measurable improvement in intrinsic clearance in HLM
(CLint 616 vs. >1250).
illustrated by compound 6. Intrinsic clearance in HLM was
measured for compounds 5 and 6 and offered no advantage
over 4. Given the apparent preference for benzyl over phene-
thyl in terms of potency, we chose to keep the benzyl group
in further optimisation efforts.
All compounds that have been described so far were pre-
pared as racemates. The next modification that was investi-
gated was ring contraction of the piperidine to the correspond-
ing pyrrolidine. This modification gave us the opportunity to
prepare the desired analogs in enantiomerically pure form as
the required enantiomerically pure starting material was read-
ily available (Scheme 2). We envisaged that DORA 4 would have
a similar receptor binding mode to the known (S)-pyrrolidine
SB-674042 disclosed by Smithkline Beecham,47 and we there-
fore set out to first prepare the (S)-enantiomer of our
pyrrolidine.48
We next explored modifications of the benzyl group at the
2-position of the piperidine ring (Table 2). Extending to the
phenethyl analog 5 led to a 10-fold and a 5-fold loss in
potency at OX1 and OX2, respectively. Subsequent removal of
the 4-methoxy substituent led to a further loss in potency as
Reaction of prolinol with sulfuryl chloride in pyridine and
dichloromethane at −78°C provided the bicyclic sulfamate
Table 2 SAR exploration of the benzyl group
a
IC50 (nM)
Cmpd
R
n
OX1
OX2
HLMb
( ) 4
( ) 5
( ) 6
OMe
OMe
H
1
2
2
2.7
28
108
3.3
16
243
616
817
1173
a
Geomean of at least 3 experiments, standard deviation is <50% in all cases. b Intrinsic metabolic clearance (μL min−1 mg−1) with human liver
microsomes, at 1 μM.
Scheme 2 Representative synthetic route to 1-acyl-2-benzylpyrrolidines. Reagents and conditions: (a) SO2Cl2, Py, DCM, −78 °C, 43–71%; (b) nBuLi,
THF, −78 °C, X = Br or I, 10–95%; (c) H+, EtOH, 95 °C, 10–95%; (d) TBTU, DIPEA, DMF, RT, 30–90%. For variations of R1 see Tables 3 & 5 and for var-
iations of R2 see Tables 3 & 6.
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun.

View Article Online
Concise Article
MedChemComm
Table 3 Potency, microsomal stability, and LLE of selected 1-acyl-2-benzylpyrrolidines
a
IC50 (nM)
Cmpd
R
OX1
8.5
OX2
8.4
HLMb (RLMb )
LLE
4.7
7
356 (>1250)
8
16
45
16
3.5
65
104 (>1250)
>1250
5.4
3.9
4.5
9
10
4.7
102
11
12
13
60
23
3.5
71
17
3.8
4.5
3.3
4.0
3.1
158
338 (>1250)
14
33
12
129
4.0
a
b
Geomean of at least 3 experiments, standard deviation is <50% in all cases. Intrinsic metabolic clearance (μL min−1 mg−1) with liver
microsomes, at 1 μM.
intermediate, that was subsequently reacted with the lithium
anion of the chosen benzene derivative. Acidic hydrolysis of
the isolated sulfamic acid at an elevated temperature
furnished the corresponding 2-benzyl pyrrolidine with intact
stereochemical integrity. Amide formation with the required
o-biaryl carboxylic acid gave the desired compound.
This ring contraction led to compound 7 (Table 3), show-
ing a 6-fold loss in potency at both OX1 and OX2 (assuming
racemic 4 would be about two times more potent as its active
enantiomer). Although the in vitro clearance of 7 in RLM
remained high (CLint >1250), this structural modification
was encouragingly accompanied by a further improvement in
metabolic stability in HLM (CLint 356 vs. 616). Rat pharmaco-
kinetics were assessed for compound 7, and in line with the
measured in vitro metabolic stability, this compound had
high clearance (55 mL min−1 kg−1) and low oral bioavailabil-
ity (<10%). Nevertheless, physicochemical characteristics
were found to be favourable for compound 7, log D was deter-
mined to be 3, LLE was 4.7, and aqueous solubility was mea-
sured as 160 μg mL−1 (pH 4.1, unbuffered). Consequently,
pyrrolidine 7 demonstrated that the optimisation of physico-
chemical properties was on track. With the aim to improve
pharmacokinetic parameters, the pyrrolidine ring was kept
constant in the next round of optimisation and attention
turned to the biaryl carboxamide, in particular the tolyl methyl
group which was considered a likely metabolic soft spot.
Table 4 Plasma, brain, and CSF exposure 3 h after po dosing to male
Wistar rats (n = 3, 100 mg kg−1)
Cmpd Plasma (ng mL−1
)
Brain (ng g−1
)
CSF (ng mL−1
)
B/P ratio
8
114
737
225
129
45
134
89
12
113
22
0.39
0.18
0.40
0.97
10
12
13
125
6
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2015

View Article Online
MedChemComm
Concise Article
Table 5 Potency, microsomal stability, and LLE of selected 1-acyl-2-benzylpyrrolidines
a
IC50 (nM)
Cmpd
R
OX1
35
OX2
39
HLMb (RLMb )
1040
LLE
2.0
15
16
17
64
40
124
798
1.9
3.1
3.4
4.3
18
19
20
3.8
26
33
1.8
16
18
>1250 (>1250)
>1250 (>1250)
996 (>1250)
2.8
1.9
1.8
21
22
13
22
10
66
>1250 (>1250)
>1250
1.6
0.6
a
b
Geomean of at least 3 experiments, standard deviation is <50% in all cases. Intrinsic metabolic clearance (μL min−1 mg−1) with liver
microsomes, at 1 μM.
Removal of the tolyl methyl group from 7 had little impact
on orexin receptor affinity, and in line with expectation, com-
pound 8 showed improved stability in HLM. Disappointingly
however, this observation did not translate across species
and 8 remained highly unstable in RLM. We next removed
one of the triazole nitrogens leading to pyrazole 9 and this
resulted in lower potency at both OX receptors. Replacement of
the triazole by 2-pyrimidine was well tolerated (10 vs. 8).
Attempts to introduce further polarity such as pyridine 11
resulted in higher microsomal stability albeit with lower
potency. Additional heteroaromatic rings were also explored
and found to be well tolerated, particularly by OX2 (12–14).
Several structural modifications of the biaryl carboxamide
led to an improved stability in HLM, and compounds with
CLint values below 100 μL min−1 mg−1 were even considered
too stable for a potential sleep drug. Our next move was to
assess central nervous system (CNS) distribution in the rat
after oral administration. To circumvent the persistent rat
microsomal instability observed so far for this compound
class, we opted for high dose experiments with the intention
of saturating metabolic processes. Compound concentration
was determined in plasma, brain, and cerebrospinal fluid
(CSF), the latter being used as a surrogate for brain free con-
centration. Four compounds were investigated for CNS pene-
tration and all were found to reach low plasma and conse-
quently brain concentrations (Table 4). In terms of brain/
plasma ratio, 13 appeared to be the most promising although
this compound had the lowest CSF concentration suggesting
it was the most highly protein bound.49,50 Encouraged by the
CNS penetration potential of 13, we selected this compound
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun.

View Article Online
Concise Article
MedChemComm
Table 6 Potency, microsomal stability, and LLE of selected phenylthiazole carboxamides
a
IC50 (nM)
Cmpd
R
OX1
111
OX2
62
HLMb (RLMb )
216
LLE
2.2
23
24
25
26
9.1
11
17
6.3
7.8
10
612
300
210
2.7
3.3
3.2
27
28
11
6.6
8.0
197 (>1250)
3.2
3.0
8.3
235
a
b
Geomean of at least 3 experiments, standard deviation is <50% in all cases. Intrinsic metabolic clearance (μL min−1 mg−1) with liver
microsomes, at 1 μM.
to explore the dimethoxyphenyl moiety which had largely
remained constant during our SAR investigation.
promising when considering both potency and microsomal
stability. It is conceivable that the polarity conveyed by the
two methoxy groups contributes to lowering the overall lipo-
philicity of the compound and replacing just one of them is
detrimental for metabolic stability.51 Subsequent attempts to
modulate the metabolism were directed back towards the
phenylthiazole moiety (Table 6). The 4-fluoro substituted phe-
nyl derivative 23 was not well tolerated by either receptor,
resulting in a 32-fold and a 21-fold loss in potency at OX1
and OX2, respectively. This motif is present in SB-649868, a
clinical compound from GSK which reached phase II trials.43
The fact that this decoration was not tolerated in our
pyrrolidine series implied a different receptor binding mode
to the one of the GSK compound. The 3-chloro derivative 24
showed a similar potency to 13; however, switching methyl
for chlorine contributed little to reducing overall lipophilicity.
We next explored the 3- and 4-methoxy derivatives 25 and 26,
respectively. Both positions were able to accommodate a
methoxy substituent retaining good potency and reasonable
stability in HLM. The related piperonyl derivative 27 showed
the same profile as either methoxy derivative and replace-
ment of one of the piperonyl oxygen atoms by carbon was
well tolerated as illustrated by 28.
Tethering the two methoxy groups to give compound 15
led to at least a 10-fold loss in potency at both OX receptors
and the same was observed for the corresponding piperonyl
derivative 16 (Table 5). Relocating the 4-methoxy substituent
to the 5-position was well tolerated by the OX receptors but
had little influence on microsomal stability as illustrated by
17. Replacement of the 3-methoxy substituent by chlorine
was also well tolerated by both OX receptors albeit with a
significant loss of microsomal stability, indicating that 18
had an in vitro half-life of less than one minute in the pres-
ence of both RLM and HLM. Other structural analogs bear-
ing
a chloromethoxy di-substitution pattern as well as
3-chloro,4-methyl and 3,4-dichloro derivatives furnished
inferior DORAs with LLE clearly heading in the wrong direc-
tion (19–22).
Although oxidative demethylation of the dimethoxyphenyl
moiety was considered a likely culprit involved in the meta-
bolic turnover of 13, our SAR investigation into the nature
and position of substituents on the phenyl ring provided us
with no handle with which to tune metabolic stability. The
3,4-dimethoxy pattern actually turned out to be the most
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2015

View Article Online
MedChemComm
Concise Article
Fig. 1 Effect of 27 on OxA-induced calcium release. (A) CHO cells expressing either the human OX1 or human OX2 receptor were pre-incubated
with a dilution series of 27 for 120 min followed by the addition of a dilution series of OxA. Calcium release was recorded, peak fluorescence
values were exported, and concentration–response curves were generated. IC50 values at 1.6 nM OxA were determined and used as a basis to cal-
culate the apparent Kb via the generalised Cheng–Prusoff equation. Representative experiment of n = 4 independent experiments is shown. Values
represent arithmetic mean of duplicates SD. (B) Experiment performed as described in (A) but with simultaneous addition of pre-mixed antagonist
and OxA instead of the pre-incubation. Representative experiment of n = 2 independent experiments is shown. Values represent arithmetic mean
of duplicates SD.
Having made initial efforts to explore the SAR around the
5-phenyl ring of the phenylthiazole moiety, it became clear
that in terms of potency several modifications were possible.
Piperonyl 27, showing reasonable human microsomal stabil-
ity and having an LLE of 3.2, was also measured in RLM and
metabolic turnover remained too high to be exactly deter-
mined under the chosen assay conditions. The aqueous solu-
bility of 27 was measured as 147 μg mL−1 (pH 4.3,
unbuffered) indicating that the improvement in solubility for
this series of compounds, as already demonstrated for com-
pound 7, was maintained. Compound 27 was selected for a
blood–brain barrier (BBB) experiment under the same condi-
tions already used for 13 (Table 4). To our surprise, plasma
were determined in vitro to be 2.7% and 4%, respectively, for
27. As in vitro binding data yielded a brain/plasma partition
ratio of 0.68, 27 was suspected to be a P-gp substrate in rats.
Indeed, 27 exhibited polarised transport in a human P-gp
transport assay with an efflux ratio of 7. A brain free fraction
of 4% for 27 translated into a free compound concentration
of 60 nM, well above its OX2 IC50 value and in line with a
measured CSF concentration of 133 ng mL−1.
The mode of antagonism of 27 at OX1 and OX2 was
assessed in more detail using Ca2+ release assays and stably
transfected Chinese hamster ovary (CHO) cells recombinantly
expressing human or rat OX1 or OX2 receptors. Orexin A con-
centration–response curves (CRC) were generated in the pres-
ence of increasing concentrations of 27 (Fig. 1A). The com-
pound induced rightward shifts of the OxA CRCs and a
suppression of the maximal OxA response demonstrating
insurmountable antagonism. Apparent Kb values (an approxi-
mation of the inhibitory constant if IC50 values generated at
low agonist concentrations are used) were thus calculated
concentrations of 27 at 3 h were found to be 2328 ng mL−1
,
18-fold higher than those determined for 13. Brain concentra-
tions were found to be 698 ng g−1, only 6-fold higher than
those of 13, resulting in a total brain/plasma ratio of 0.3 and
suggesting that 27 partitions less readily into the CNS as
compared to 13. Unbound fractions in rat plasma and brain52
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun.

View Article Online
Concise Article
MedChemComm
Fig. 2 Dissociation kinetics of 27 determined by calcium release assays in CHO cells expressing either human OX1 or human OX2. Cells were pre-
incubated for 120 min with antagonist dilution series and then either directly stimulated with OxA IJEC50–EC70) or subjected to a compound wash-
out procedure followed by stimulation with OxA at the indicated time points after washout. The peak calcium responses were used to calculate
IC50 values which were transformed into apparent Kb values as described in the ESI.† The time-dependent changes in apparent Kb after washout of
three independent experiments (fine lines) and their geometric mean (bold line) are shown. Kb values after washout that are significantly different
(p < 0.05) from the Kb at 0 min are indicated with an asterisk. Test: one-way ANOVA, Dunnett's post test.
Fig. 3 Effect of 27, suvorexant, and almorexant on the time spent in sleep and wake stages (% of total time) during the first 6 h of the night active
period post administration in male Wistar rats. Rats were administered a single oral dose of vehicle (PEG 400) or 100 mg kg−1 of compound. Data
are expressed as mean SEM. *p < 0.05, **p < 0.01, ***p < 0.001 (n = 8 for 27, n = 12 for almorexant and n = 14 for suvorexant).
with the help of the generalised Cheng–Prusoff equation as
described in the ESI.† Apparent Kb values for 120 min antago-
nist pre-incubation were calculated to be Kb = 5.3 nM (human
OX1) and 1.4 nM (human OX2). No differences in potency,
selectivity, or surmountability between human and rat recep-
tors were detected.53
Insurmountable antagonism can be caused by competitive
antagonists with slow dissociation rates as well as by alloste-
ric antagonists that do not compete with the natural ligand
for binding but change the affinity for the natural ligand
through binding to an allosteric site. These two mecha-
nisms causing insurmountability can be differentiated by
performing the previously described curve shift experiments
with simultaneous addition of the antagonist dilutions
together with the OxA dilutions. In this experimental setting,
a competitive binding mode results in antagonist-induced
rightward shifts of the OxA CRC without suppression of maxi-
mal response, while an allosteric mode of antagonism still
results in an insurmountable antagonist profile. Fig. 1B
shows the CRCs for OxA in the presence of increasing con-
centrations of 27 on human OX1 and OX2. Clearly, 27 caused
rightward shifts in OxA CRCs without suppression of maxi-
mal responses demonstrating the competitive nature of its
binding mode.
Next, the receptor occupancy half-life (t1/2) of 27 at human
OX1 and human OX2 was assessed using antagonist washout
assays in combination with calcium mobilisation assays.
CHO cells expressing the recombinant human OX1 or OX2
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2015

View Article Online
MedChemComm
Concise Article
receptors were dyed with fluo-4 and then subjected to a dilu-
tion series of 27. After 120 min of incubation at room temper-
ature, cells were extensively washed to remove all unbound
antagonist and then after different recovery times (0–30 min)
stimulated with EC50–70 of OxA to analyse for residual recep-
tor blockade. The residual blockade was quantified by deter-
mining the IC50 value at the different time points and then
calculating the apparent Kb via the generalised Cheng–
for suvorexant and by +22% for almorexant ( p < 0.01 for 27
and suvorexant and p < 0.001 for almorexant, paired t-test).
Finally, as observed with other DORAs,44,55–59 the time spent
in REM sleep was also increased compared to vehicle-treated
rats, significantly for 27 and almorexant (+84 and +50%,
respectively, p < 0.01, paired t-test) and non-significantly for
suvorexant (+20%, p > 0.05, paired t-test).
Prusoff equation using the EC50 and slope of the OxA CRC Conclusions
determined at every time point. Three independent experi-
In summary, the structural optimisation of a thienopiperidine
ments were performed. Fig. 2 shows the development of the
apparent Kb values after 27 washout. Kb values of the three
individual experiments are displayed as fine lines, and their
geometric mean is displayed as a bold line. For both recep-
tors, the three experiments yielded similar increases in Kb
after washout with a statistically significant change in Kb
value versus Kb (0 min) occurring at 5 min (OX1) and at 20
min (OX2). The fold-change in Kb at these time points was
used to calculate an approximate t1/2 assuming first-order dis-
sociation kinetics yielding half-lives of 1.1 min (human OX1)
and 5.4 min (human OX2). For comparison, the half-life of
suvorexant was reported to be 79 min at human OX2.54 Thus,
27 displays a rather short half-life at OX1 and OX2 but long
enough to result in insurmountable antagonism in calcium
release assays, which represent – due to their rapid response
time – the most sensitive assays to detect insurmountable
antagonism.
To evaluate the pharmacodynamics of compound 27, it
was tested in a rat sleep model. Male Wistar rats implanted
with radiotelemetry probes recording continuously electroen-
cephalogram/electromyography (EEG/EMG) and locomotor
activity were administered with a single oral dose of 100 mg
kg−1 at the beginning of the nocturnal active phase, when
endogenous orexin levels increase. DORA 27 showed a com-
parable sleep-promoting effect to almorexant and suvorexant
when tested under the same experimental conditions (Fig. 3).
Behaviourally, it significantly decreased home cage activity by
47% over the first 6 h of the night period following adminis-
tration compared to vehicle-treated animals (−42% and −31%
vs. vehicle for almorexant and suvorexant at 100 mg kg−1 po,
respectively, p < 0.001 for 27 and almorexant and p < 0.01
for suvorexant, paired t-test). Electrophysiologically, it
decreased significantly the time spent in active wake com-
pared to vehicle-treated animals (−29%, −20% and −16% vs.
vehicle for 27, almorexant, and suvorexant, respectively, p <
0.001 for 27 and almorexant and p < 0.01 for suvorexant,
paired t-test). Compound 27 increased the time spent in
quiet wake by 17% compared to vehicle-treated animals ( p <
0.05, paired t-test). However, the impact on this parameter is
not predictive for sleep quality and depends on the com-
pound tested (−8% for almorexant and +9% for suvorexant
vs. vehicle, p < 0.05 and p > 0.05, respectively, paired t-test).
As a consequence of the decrease in time spent in wake
stages, the time spent in non-REM (rapid eye movement)
sleep was significantly increased vs. vehicle over the 6 h
period following administration, by +18% for 27, by +19%,
lead compound led to the discovery of a novel 1-acyl-2-
benzylpyrrolidine series of competitive dual orexin receptor
antagonists with improved physicochemical properties. Com-
pound 27 showed insurmountable antagonism in calcium
release assays at OX1 and OX2, combined with a rather short
estimated receptor occupancy half-life of 1–5 min at both
orexin receptor subtypes. Compound 27 showed a comparable
sleep-promoting effect to almorexant and suvorexant when
tested under the same experimental conditions. Whilst the
potential of the 1-acyl-2-benzylpyrrolidines to deliver in vivo
active compounds has been demonstrated, further studies
shall aim at improving CNS distribution.
Acknowledgements
For expert technical support, the authors gratefully thank
Christian Barten, Pauline Schrepfer, Daniel Trachsel, Katalin
Menyhart, Celia Müller-Grandjean, Thomas Sasse, Stephane
Delahaye, Isabelle Weber, Fabienne Drouet, Eric Soubieux,
Rolf Wuest and Julia Friedrich.
Notes and references
1 T. Sakurai, Nat. Rev. Neurosci., 2007, 8, 171–181.
2 R.-J. Liu, A. N. van den Pol and G. K. Aghajanian,
J. Neurosci., 2002, 22, 9453–9464.
3 T. S. Kilduff and C. Peyron, Trends Neurosci., 2000, 23, 359–365.
4 L. De Lecea, T. S. Kilduff, C. Peyron, X. B. Gao, P. E. Foye,
P. E. Danielson, C. Fukuhara, E. L. F. Battenberg, V. T.
Gautvik, F. S. Bartlett II, W. N. Frankel, A. N. Van Den Pol,
F. E. Bloom, K. M. Gautvik and J. G. Sutcliffe, Proc. Natl.
Acad. Sci. U. S. A., 1998, 95, 322–327.
5 T. Matsuki and T. Sakurai, Results Probl. Cell Differ.,
2008, 46, 27–55.
6 K. N. Nilaweera, P. Barrett, J. G. Mercer and P. J. Morgan,
Neuroscience, 2003, 119, 713–720.
7 T. Sakurai, A. Amemiya, M. Ishii, I. Matsuzaki, R. M.
Chemelli, H. Tanaka, S. C. Williams, J. A. Richardson, G. P.
Kozlowski, S. Wilson, J. R. S. Arch, R. E. Buckingham, A. C.
Haynes, S. A. Carr, R. S. Annan, D. E. McNulty, W.-S. Liu,
J. A. Terrett, N. A. Elshourbagy, D. J. Bergsma and M.
Yanagisawa, Cell, 1998, 92, 573–585.
8 J. Gatfield, C. Brisbare-Roch, F. Jenck and C. Boss,
ChemMedChem, 2010, 5, 1197–1214.
9 R. Spinazzi, P. G. Andreis, G. P. Rossi and G. G. Nussdorfer,
Pharmacol. Rev., 2006, 58, 46–57.
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun.

View Article Online
Concise Article
MedChemComm
10 K. Yoshida, S. McCormack, R. A. Espana, A. Crocker and
T. E. Scammell, J. Comp. Neurol., 2006, 494, 845–861.
11 M. G. Lee, O. K. Hassani and B. E. Jones, J. Neurosci.,
2005, 25, 6716–6720.
12 J. C. Geerling, T. C. Mettenleiter and A. D. Loewy,
Neuroscience, 2003, 122, 541–550.
13 J. Fadel and A. Y. Deutch, Neuroscience, 2002, 111, 379–387.
14 A. Yamanaka, N. Tsujino, H. Funahashi, K. Honda, J.-l. Guan,
Q.-P. Wang, M. Tominaga, K. Goto, S. Shioda and T. Sakurai,
Biochem. Biophys. Res. Commun., 2002, 290, 1237–1245.
15 E. Eggermann, M. Serafin, L. Bayer, D. Machard, B. Saint-
Mleux, B. E. Jones and M. Muhlethaler, Neuroscience,
2001, 108, 177–181.
J. Zeitlhofer, S. Berg, M. Partinen, C. L. Bassetti, B. Hoegl,
I. O. Ebrahim, E. Holsboer-Trachsler, H. Bengtsson, Y.
Peker, U. M. Hemmeter, E. Chiossi, G. Hajak and J.
Dingemanse, Clin. Pharmacol. Ther., 2012, 91, 975–985.
33 D. Michelson, E. Snyder, E. Paradis, M. Chengan-Liu, D. B.
Snavely, J. Hutzelmann, J. K. Walsh, A. D. Krystal, R. M.
Benca, M. Cohn, C. Lines, T. Roth and W. J. Herring, Lancet
Neurol., 2014, 13, 461–471.
34 L. Citrome, Int. J. Clin. Pract., 2014, 68, 1429–1441.
35 US Belsomra Package Insert. United States Food and Drug
Administration product information. Available from: http://
www.accessdata.fda.gov/scripts/cder/drugsatfda/index.cfm?
fuseaction=Search.Label_ApprovalHistory#labelinfo.
16 J. J. Hagan, R. A. Leslie, S. Patel, M. L. Evans, T. A. Wattam,
S. Holmes, C. D. Benham, S. G. Taylor, C. Routledge, P.
Hemmati, R. P. Munton, T. E. Ashmeade, A. S. Shah, J. P.
Hatcher, P. D. Hatcher, D. N. C. Jones, M. I. Smith, D. C.
Piper, A. J. Hunter, R. A. Porter and N. Upton, Proc. Natl.
Acad. Sci. U. S. A., 1999, 96, 10911–10916.
17 C. Peyron, D. K. Tighe, A. N. van den Pol, L. de Lecea, H. C.
Heller, J. G. Sutcliffe and T. S. Kilduff, J. Neurosci., 1998, 18,
9996–10015.
36 C. Dugovic, J. E. Shelton, S. Yun, P. Bonaventure, B. T.
Shireman and T. W. Lovenberg, Front. Neurosci., 2014, 8, 28.
37 A. J. Roecker, S. P. Mercer, J. D. Schreier, C. D. Cox, M. E.
Fraley, J. T. Steen, W. Lemaire, J. G. Bruno, C. M. Harrell,
S. L. Garson, A. L. Gotter, S. V. Fox, J. Stevens, P. L.
Tannenbaum, T. Prueksaritanont, T. D. Cabalu, D. Cui, J.
Stellabott, G. D. Hartman, S. D. Young, C. J. Winrow, J. J.
Renger and P. J. Coleman, ChemMedChem, 2014, 9, 311–322.
38 S. P. Mercer, A. J. Roecker, S. Garson, D. R. Reiss, C.
Meacham Harrell, K. L. Murphy, J. G. Bruno, R. A. Bednar, W.
Lemaire, D. Cui, T. D. Cabalu, C. Tang, T. Prueksaritanont,
G. D. Hartman, S. D. Young, C. J. Winrow, J. J. Renger and
P. J. Coleman, Bioorg. Med. Chem. Lett., 2013, 23, 6620–6624.
39 C. W. Murray, D. A. Erlanson, A. L. Hopkins, G. M. Keserü,
P. D. Leeson, D. C. Rees, C. H. Reynolds and N. J.
Richmond, ACS Med. Chem. Lett., 2014, 5, 616–618.
40 M. Bolli, C. Boss, C. Brotschi, B. Heidmann, T. Sifferlen, D.
Trachsel and J. T. Williams, WO2012114252A1, 2012.
41 D. Gramec, L. Peterlin Masic and M. Sollner Dolenc, Chem.
Res. Toxicol., 2014, 27, 1344–1358.
42 P. J. Coleman, J. D. Schreier, C. D. Cox, M. J. Breslin, D. B.
Whitman, M. J. Bogusky, G. B. McGaughey, R. A. Bednar, W.
Lemaire, S. M. Doran, S. V. Fox, S. L. Garson, A. L. Gotter,
C. M. Harrell, D. R. Reiss, T. D. Cabalu, D. Cui, T.
Prueksaritanont, J. Stevens, P. L. Tannenbaum, R. G. Ball, J.
Stellabott, S. D. Young, G. D. Hartman, C. J. Winrow and J. J.
Renger, ChemMedChem, 2012, 7, 415–424.
18 A. S. Jansen, X. V. Nguyen, V. Karpitskiy, T. C. Mettenleiter
and A. D. Loewy, Science, 1995, 270, 644–646.
19 C. J. Winrow, K. Q. Tanis, D. R. Reiss, A. M. Rigby, J. M.
Uslaner, V. N. Uebele, S. M. Doran, S. V. Fox, S. L. Garson, A. L.
Gotter, D. M. Levine, A. J. Roecker, P. J. Coleman, K. S. Koblan
and J. J. Renger, Neuropharmacology, 2010, 58, 185–194.
20 G. C. Harris and G. Aston-Jones, Trends Neurosci., 2006, 29,
571–577.
21 G. C. Harris, M. Wimmer and G. Aston-Jones, Nature,
2005, 437, 556–559.
22 E. Merlo Pich and S. Melotto, Front. Neurosci., 2014, 8, 26.
23 P. L. Johnson, W. Truitt, S. D. Fitz, P. E. Minick, A. Dietrich,
S. Sanghani, L. Traskman-Bendz, A. W. Goddard, L. Brundin
and A. Shekhar, Nat. Med., 2010, 16, 111–115.
24 M. A. Steiner, H. Lecourt and F. Jenck, Int. J. Neuropsychopharmacol.,
2013, 16, 417–432.
25 M. A. Steiner, H. Lecourt and F. Jenck, Psychopharmacology,
2012, 223, 465–475.
26 T. M. Furlong, D. M. L. Vianna, L. Liu and P. Carrive, Eur. J.
Neurosci., 2009, 30, 1603–1614.
27 B. Boutrel and L. de Lecea, Physiol. Behav., 2008, 93, 947–951.
28 S. G. Nair, S. A. Golden and Y. Shaham, Br. J. Pharmacol.,
2008, 154, 406–416.
29 C. Brisbare-Roch, J. Dingemanse, R. Koberstein, P. Hoever, H.
Aissaoui, S. Flores, C. Mueller, O. Nayler, J. van Gerven, S. L. de
Haas, P. Hess, C. Qiu, S. Buchmann, M. Scherz, T. Weller, W.
Fischli, M. Clozel and F. Jenck, Nat. Med., 2007, 13, 150–155.
30 M. Narita, Y. Nagumo, S. Hashimoto, M. Narita, J. Khotib,
M. Miyatake, T. Sakurai, M. Yanagisawa, T. Nakamachi, S.
Shioda and T. Suzuki, J. Neurosci., 2006, 26, 398–405.
31 P. J. Coleman, C. D. Cox and A. J. Roecker, Curr. Top. Med.
Chem., 2011, 11, 696–725.
43 R. Di Fabio, A. Pellacani, S. Faedo, A. Roth, L. Piccoli, P.
Gerrard, R. A. Porter, C. N. Johnson, K. Thewlis, D. Donati,
L. Stasi, S. Spada, G. Stemp, D. Nash, C. Branch, L. Kindon,
M. Massagrande, A. Poffe, S. Braggio, E. Chiarparin, C.
Marchioro, E. Ratti and M. Corsi, Bioorg. Med. Chem. Lett.,
2011, 21, 5562–5567.
44 C. D. Cox, M. J. Breslin, D. B. Whitman, J. D. Schreier, G. B.
McGaughey, M. J. Bogusky, A. J. Roecker, S. P. Mercer, R. A.
Bednar, W. Lemaire, J. G. Bruno, D. R. Reiss, C. M. Harrell,
K. L. Murphy, S. L. Garson, S. M. Doran, T. Prueksaritanont,
W. B. Anderson, C. Tang, S. Roller, T. D. Cabalu, D. Cui,
G. D. Hartman, S. D. Young, K. S. Koblan, C. J. Winrow, J. J.
Renger and P. J. Coleman, J. Med. Chem., 2010, 53, 5320–5332.
45 C. L. Branch, W. N. Chan, A. Johns, C. N. Johnson, D. J.
Nash, R. Novelli, J.-P. Pilleux, R. A. Porter, R. E. A. Stead and
G. Stemp, WO2003002561A1, 2003.
32 P. Hoever, G. Dorffner, H. Benes, T. Penzel, H. Danker-
Hopfe, M. J. Barbanoj, G. Pillar, B. Saletu, O. Polo, D. Kunz,
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2015

View Article Online
MedChemComm
Concise Article
46 P. D. Leeson and B. Springthorpe, Nat. Rev. Drug Discovery,
2007, 6, 881–890.
54 R. Mould, J. Brown, F. H. Marshall and C. J. Langmead, Br.
J. Pharmacol., 2014, 171, 351–363.
47 C. J. Langmead, J. C. Jerman, S. J. Brough, C. Scott, R. A.
Porter and H. J. Herdon, Br. J. Pharmacol., 2004, 141,
340–346.
48 G. F. Cooper and M. G. Martin, US5130432A, 1992.
49 T. T. Mariappan, V. Kurawattimath, S. S. Gautam, C. P.
Kulkarni, R. Kallem, K. S. Taskar, P. H. Marathe and S.
Mandlekar, Mol. Pharmaceutics, 2014, 11, 477–485.
50 J. Watson, S. Wright, A. Lucas, K. L. Clarke, J. Viggers, S.
Cheetham, P. Jeffrey, R. Porter and K. D. Read, Drug Metab.
Dispos., 2009, 37, 753–760.
55 C. J. Winrow, A. L. Gotter, C. D. Cox, P. L. Tannenbaum, S. L.
Garson, S. M. Doran, M. J. Breslin, J. D. Schreier, S. V. Fox,
C. M. Harrell, J. Stevens, D. R. Reiss, D. Cui, P. J. Coleman
and J. J. Renger, Neuropharmacology, 2012, 62, 978–987.
56 T. Sifferlen, C. Boss, E. Cottreel, R. Koberstein, M. Gude, H.
Aissaoui, T. Weller, J. Gatfield, C. Brisbare-Roch and F.
Jenck, Bioorg. Med. Chem. Lett., 2010, 20, 1539–1542.
57 H. Aissaoui, R. Koberstein, C. Zumbrunn, J. Gatfield, C.
Brisbare-Roch, F. Jenck, A. Treiber and C. Boss, Bioorg. Med.
Chem. Lett., 2008, 18, 5729–5733.
51 An LCMS retention time shift of 0.07 min was observed for
compound 18 vs. compound 13, see the ESI.†
52 K. D. Read and S. Braggio, Expert Opin. Drug Metab. Toxicol.,
2010, 6, 337–344.
53 For a comparison of human vs. rat Kb values please see
Table S1 in the ESI.†
58 C. Boss, C. Brisbare-Roch, F. Jenck, H. Aissaoui, R.
Koberstein, T. Sifferlen and T. Weller, Chimia, 2008, 62,
974–979.
59 T. Sifferlen, R. Koberstein, E. Cottreel, A. Boller, T. Weller, J.
Gatfield, C. Brisbare-Roch, F. Jenck and C. Boss, Bioorg.
Med. Chem. Lett., 2013, 23, 3857–3863.
This journal is © The Royal Society of Chemistry 2015
Med. Chem. Commun.