The Quest for the Best Dual Orexin Receptor Antagonist (Daridorexant) for the Treatment of Insomnia Disorders

Article abstract of DOI:10.1002/cmdc.202000453

Since its discovery in 1998, the orexin system has been of interest to the research community as a potential therapeutic target for the treatment of sleep/wake disorders, stress and anxiety disorders, addiction or eating disorders. It consists of two G protein-coupled receptors, the orexin 1 and orexin 2 receptors, and two neuropeptides with agonistic effects, the orexin A and orexin B peptides. Herein we describe our efforts leading to the identification of a promising set of dual orexin receptor antagonists (DORAs) which subsequently went through physiology-based pharmacokinetic and pharmacodynamic modelling>^[1] and finally led to the selection of daridorexant, currently in phase 3 clinical trials for the treatment of insomnia disorders.

Full text of DOI:10.1002/cmdc.202000453

Accepted Article
Title: The quest for the best dual orexin receptor antagonist
(daridorexant) for the treatment of insomnia disorders
Authors: Christoph Boss, John Gatfield, Christine Brotschi, Bibia
Heidmann, Thierry Sifferlen, Markus von Raumer, Gunther
Schmidt, Jodi T. Williams, Alexander Treiber, and Catherine
Roch
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemMedChem 10.1002/cmdc.202000453
Link to VoR: https://doi.org/10.1002/cmdc.202000453
01/2020

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
The quest for the best dual orexin receptor antagonist
(daridorexant) for the treatment of insomnia disorders
Christoph Boss*[a], John Gatfield [a], Christine Brotschi [a], Bibia Heidmann [a], Thierry Sifferlen [a],
Markus von Raumer [a], Gunther Schmidt [a], Jodi T. Williams [a], Alexander Treiber [a], and Catherine
Roch [a]
[a]
Dr. C. Boss, Dr. J. Gatfield, Dr. C. Brotschi, Dr. B. Heidmann, Dr. T. Sifferlen, Dr. M. von Raumer, Dr. G. Schmidt, Dr. J.T. Williams, Dr. A. Treiber, Dr. C.
Roch
Drug Discovery and Preclinical Research & Development
Idorsia Pharmaceuticals Ltd
Hegenheimermattweg 91; CH-4123 Allschwil / BL (Switzerland)
E-mail: christoph.boss@idorsia.com
Supporting information for this article is given via a link at the end of the document.
[3]
Abstract: Since its discovery in 1998, the orexin system is of interest
to the research community as a potential therapeutic target for the
treatment of sleep/wake disorders, stress and anxiety disorders,
addiction or eating disorders. It consists of two G protein-coupled
all over the CNS suggesting a role in various functions. Many
regions express both Ox1R and Ox2R, but some express mainly
one of the two, suggesting different functions of the two receptor
[8]
subtypes.
The orexin system is indeed involved in various
receptors, the orexin
1
and the orexin 2-receptor and two
physiological processes. It is a key player in the control of arousal,
more specifically for the maintenance of wakefulness and for
sleep-to-wake transitions, taking the homeostatic, circadian and
motivational necessities into account. [9] In addition, the role of the
orexin system in the control of cardiovascular, behavioral,
metabolic and thermoregulatory function led to the hypothesis
that it integrates sleep-wake neurophysiology with energy
metabolism to optimize resource utilization and energy
conservation.[10] It is thus assumed, that modulation of the orexin
system may have an impact on various CNS disorders such as
sleep-wake disorders, addiction, eating disorders, stress and
anxiety disorders. Some polymorphisms in the orexin ligand and
receptor genes have been described in the human population,
some of which might be linked to disease phenotypes, however
neuropeptides with agonistic effects, the orexin A and the orexin B-
peptide. We describe our efforts leading to the identification of a
promising set of dual orexin receptor antagonists (DORAs) which
subsequently went through physiology-based pharmacokinetic and
[1]
pharmacodynamic modelling
and finally led to the selection of
daridorexant (93) currently in phase 3 clinical trials for the treatment
of insomnia disorders.
Introduction
The orexin neuropeptides, orexin A (OxA) and orexin B (OxB),
also known as hypocretin-1 and hypocretin-2, were discovered in
1998 by two independent research groups.[2] The two
neuropeptides are secreted by a small population of excitatory
neurons from the lateral hypothalamus area.[3] Orexin A and
orexin B were found to bind to previously identified orphan G
protein-coupled receptors (GPCR) the orexin 1 (Ox1R) and the
[11]
with little knowledge on the mechanism of their involvement.
Over the last two decades, small molecule modulators of the
orexin system were identified thanks to the efforts of several
research teams. Antagonists of the orexin receptors, exhibiting
different selectivity profiles (dual orexin receptor antagonists
(DORA) or selective orexin receptor antagonists for one or the
other receptor), were broadly used to investigate
pharmacologically the impact of orexin receptor inhibition in
physiological and pathological conditions. Orexin receptor
antagonists with differential binding kinetics (surmountable versus
insurmountable) properties were also identified. [12] More recently,
reports on selective orexin 2 receptor agonists were published,
and showed the expected wake promoting effects in healthy
rodents or monkeys and wake promotion and decrease of
cataplexy events in narcoleptic mice. [13]
Another important result to the field was the publication of the
crystal structure of the human Ox2R, a member of the b-branch
of the rhodopsin GPCR family, bound to suvorexant (1).[14] The
authors succeeded to solve the structure at a resolution of 2.5
Angstroem, by applying lipid-mediated crystallization and protein
engineering with a fusion chimera. A horseshoe resembling
conformation is adopted by suvorexant (1) allowing optimal p-
stacking and enabling binding deeply into the orthosteric receptor
binding pocket. Further reports have been published describing
orexin
2 receptor (Ox2R) (also termed hypocretin 1 and
hypocretin 2 receptor). Both orexin neuropeptides are produced
from prepro-orexin by a cascade of enzymatic reactions.[4] Orexin
A, a 33 amino-acid peptide, activates both Ox1 and Ox2 receptors
with similar potencies. Orexin B, which consists of 28 amino-acids,
activates the Ox2R with higher potency than the Ox1R.[5] In
neurons, Ox1R and Ox2R signal via various cascades to increase
intracellular calcium levels pre- and/or post-synaptically, thus
augmenting synaptic transmission (Figure 1).[6] The orexin system
is conserved across mammalian species. OxA, containing two
disulfide bridges, is conserved in rat, mouse, pig, dog and man.
The linear, non-lipophilic OxB peptide from rat, mouse, pig and
dog differs by maximally two amino acids (S18N) from human
OxB.[7] As structural and functional homology between rat, dog
and human orexin receptors and peptides are high, reliable
translations of pharmacological outcomes can be made among
species. While orexin-producing neurons are localized exclusively
in the lateral hypothalamus, the receptors are broadly expressed
1
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
the crystal structure of the human Ox2R (hOx2R) and the human
Ox1R (hOx1R) bound to small molecule ligands,[15] and more
recently the hOx2R bound to the selective Ox2R antagonist
EMPA.[16] An additional recent publication [17] discussed 14 novel
structures with 10 different novel ligands in one or both the orexin
receptors. This wealth of bio-structural information helps to
understand the structural features resulting in selectivity for either
receptor or duality of compounds and might even influence the
quest for orexin receptor agonists.
Figure 1. Orexin enhances synaptic transmission in orexin-responsive neurons of the central nervous system. A small group of orexinergic neurons in the lateral
hypothalamus synthesizes prepro-orexin (131 amino acids) which is proteolytically processed by signal peptidase and pro-hormone convertases into the
neuropeptides orexin A (33 amino acids, two disulfide bridges) and orexin B (28 amino acids) and stored in synaptic vesicles. Upon activation of orexinergic neurons,
orexin A and orexin B are released and activate pre- and post-synaptically located OX1 and OX2 receptors that are expressed in multiple areas of the central
nervous system. OX1 and OX2 receptors are G protein-coupled receptors whose activation in neurons – via various signaling cascades often involving
phospholipase C (PLC) – leads to elevated intracellular calcium concentrations- and enhanced synaptic transmission.
[24]
Patients suffering from insomnia disorders are dissatisfied with
their sleep quantity or quality and suffer from clinically significant
negative consequences on daytime performance. It is estimated
that in the US, about 50% of the adult population is dissatisfied
with its sleep in at least a few nights per week. [18] Sleep problems
result in successful outcomes.
The former classes of
pharmacological treatments for insomnia include either targeting
gamma-aminobutyric acid (GABA)-A, histamine, serotonin or
melatonin receptors. The non-benzodiazepine GABA-A receptor
modulators, such as zolpidem and eszopiclone are considered as
standard-of-care insomnia treatments.
allosterically enhancing the effect of the most important inhibitory
neurotransmitter (GABA) and result in general CNS depression to
provoke sleep. [26] Such treatments are most efficient in initiating
sleep. Due to their short duration of action, they are less effective
in maintaining sleep. This was improved with slow-release
[25]
are highly prevalent even though only about 10% of the patients
These drugs work by
[19]
fulfill the formal criteria for insomnia diagnosis.
Insomnia is
defined by either difficulties in initiating and/or maintaining sleep
or when the sleep that is obtained is non-refreshing or of poor
quality. It is often associated with one or more of the following
symptoms: fatigue, low energy, difficulty concentrating, mood
disturbances, and decreased performance in work or at school.
formulations. Moreover, the approach of general CNS depression
[20] [21]
[27]
,
The cost for society due to lost work time, increased
results in various safety concerns for patients.
Increased
accident risk and higher occurrence of chronic health problems
are significant. Estimates for the US published in 1994 and 2015
nighttime parasomnias (sleepwalking, eating) and increased
occurrence of driving accidents due to next-day somnolence have
[28]
are above 100 billion USD per year.[22]
,
[23] This, and the serious
been associated with the use of GABA modulators.
This
consequences on human health prove the high medical need for
safe and effective insomnia treatments. Diagnosis of sleep
disorders is often complex due to difficulties identifying the cause
of the problem. This results in the fact that treatment of sleep
resulted in modifications of labeling of GABA modulators and
requests by the US Food and Drug Administration (FDA) for dose
lowering for sedative hypnotic drugs in 2013. In addition,
tolerance and rebound insomnia have been associated with
zolpidem treatment. This led to the recommendation of
intermittent use for only short periods of time for patients suffering
from insomnia. [26] Novel insomnia treatment options devoid of the
disorders may be highly complex. One option is
a
nonpharmacological intervention in the form of cognitive-
behavioral therapy (CBT). CBT requires increased contact with
health care providers and asks for high patient compliance to
2
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
[37]
drawbacks of general CNS depression would therefore represent
an important step forward to treat this life-impairing condition.
Several, structurally and biochemically diverse orexin receptor
antagonists have been clinically investigated over the past
decade (Figure 2), resulting in a new class of pharmacological
treatment with the FDA approval of suvorexant (1) from Merck &
residual effects, as for example sedation.
We therefore
embarked on a new program to identify potent, dual, and brain
penetrant orexin antagonists starting from ACT-462206 (7),[38]
which was in human phase I clinical trials and now is in
development in veterinary medicine. We herein describe the
details of our work resulting in the identification of daridorexant
(ACT-541468, 93) currently in phase 3 clinical trials for the
treatment of insomnia disorders.
Co in August 2014 and of lemborexant (2) from Eisai in December
[29]
2019 for the treatment of insomnia.
The first dual orexin
receptor antagonist (DORA) with clinical development history in
From a lead optimization perspective, the discovery of an
insomnia drug poses specific challenges with respect to its
pharmacological profile in man. Rapid sleep onset and sleep
maintenance for the entire night period of about 8 hours are key
attributes. Beyond target potency, brain penetration and safety
aspects, particular attention was therefore drawn to the
pharmacokinetic optimization in man. As outlined in our previous
publication, physiology-based pharmacokinetic modelling was
used to identify those compounds with the best PK-PD profile in
man[1]. In practical terms, only broad acceptance criteria were
defined for the progression of compounds for those assays driving
human PK, i.e. metabolic stability, plasma protein binding,
solubility in biological media, logD and pKa. In contrast, more
stringent criteria were employed for other optimization parameters
like hERG, MDR-1/P-gp efflux, or P450 inhibition.
insomnia was almorexant (3). [30] The compound was stopped due
[31]
to tolerability issues.
The piperidine-based compound GSK-
649868 (4) was studied in insomniac patients. It was put on hold
due to preclinical toxicology findings although it was reported to
promote sleep. The back-up compound from Merck, filorexant (5),
also a piperidine based DORA, showed sleep promoting effects
[29b, 32]
in insomniac patients in phase 2 clinical trials.
Finally, a
selective Ox2 receptor antagonist (SO2RA), seltorexant (6) found
in a research program at Janssen Pharmaceutica / JnJ, [33] is also
studied in clinical trials by Minerva Biosciences focusing on the
treatment of insomnia with or without major depressive disorder.
Results and Discussion
Compound 7 (Figure 3) exhibited excellent properties as a potent,
dual and brain penetrant orexin receptor antagonist. It was not a
P-glycoprotein (Pgp) substrate (Efflux ratio = 1.1) based on a
human multidrug resistant protein transporter assay (MDR-1), it
showed human plasma protein binding (hPPB) as well as rat
plasma protein binding (rPPB) of 99.3% and metabolic stability in
human liver microsomes (HLM) of 140 µL min-1 mg-1. The
compound showed no significant cytochrome P450 (P450)
inhibition nor time-dependent cytochrome P450 inhibition. The
pharmacokinetic parameters in rats and dogs showed a short half-
life (t1/2) of 1.9 h and 1.7 h respectively. Compound 7 was
Figure 2. Structures of some orexin receptor antagonists with clinical history.
The regulation of the sleep-wake cycle is highly complex and
involves different neuronal pathways. The orexin system is a key
component of arousal via its excitatory projections to all the wake-
promoting brain areas. Its role as a stabilizer of wakefulness,
allowing long periods of wake, is well recognized. [34] Pre-clinically,
DORAs have shown sleep promotion in healthy animals while
maintaining sleep architecture, an advantage as compared to
[35]
former sleep drugs with less selective mechanisms of action.
Animal model data translated into efficacy in patients suffering
[36]
from insomnia.
However, suvorexant (1), the first DORA
entering the market, was approved by the FDA with limitations in
terms of dose due to concerns about dose-related, next-morning
Figure 3. Where we started: ACT-462206; the quest for anilide replacements.
3
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
investigated in male Wistar rats and in male Beagle dogs
implanted with radio telemetry probes recording continuously the
electroencephalogram (EEG)/ the electromyogram (EMG) and
locomotor activity. It dose-dependently increased both the time
attractive starting point for further medicinal chemistry work. As
depicted in Figure 3, our initial efforts targeted the replacement of
the anilide moiety in 7. Non-aromatic amides were investigated,
exemplified by 8, resulting in an almost total loss of orexin
spent in rapid-eye-movement (REM) and non-REM (NREM) sleep. receptor antagonistic activity.
Sleep architecture was conserved, as the relative proportion of
More promising results were obtained by heteroaromatic
NREM and REM sleep over total sleep time were not significantly
replacements of the anilide moiety by an oxadiazole unit in 9 or a
benzimidazole group in 10, resulting in activities interesting
enough to continue the optimization.
[38]
changed.
Compound 7 had certain limitations in its
physicochemical behavior, but the overall profile made 7 an
Figure 4. SAR investigation of the benzimidazole proline-sulfonamides: Substitution on the benzimidazole moiety (compounds are racemic).
In Figure 4, a summary of the first efforts on the SAR of
substituents attached to the phenyl ring of the benzimidazole
group is depicted. The mono-methyl-substituted compounds 10
(Figure 3) and 11 showed similar activities on the Ox1 receptor.
A methyl-substituent in position 4 (11) resulted in improved
potency on the Ox2 receptor as compared to 10. Bis-methyl
substituted benzimidazoles 12 to 14 all exhibit significantly
improved antagonist activities on both, the Ox1 and Ox2 receptor.
Replacing the methyl-substituent in 11 by a trifluoromethyl-
substituent (15) resulted in a 10-fold improved Ox1 potency and
only a slight improvement in Ox2 potency whereas replacement
of both methyl groups from 12 with CF3-groups in 16 resulted in a
complete loss of potency. Compound 17 is the trifluoromethyl
analog of 10, exhibiting similar activity and 18 proves that
halogen-substitution is tolerated. Consequently, the compounds
Figure 5. Sulfonamide SAR on the phenyl-oxadiazole series (compounds are racemic).
4
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
depicted in Figure 4 clearly showed that the substitution pattern
on the benzimidazole moiety may significantly influence the
orexin receptor antagonist potency.
A first aspect we clarified, by performing the core modification
exercise summarized in Figure 7, the relevance of the proline-
template.
We also wanted to get a first idea on the potential of the 3-phenyl
substituted oxadiazole as an anilide replacement. We therefore
prepared a small series of substituted phenyl-sulfonamides
combined with the oxadiazole based isosteres, summarized in
Figure 5 by compounds 19 to 25. The next step was to replace
the substituted phenyl sulfonamide moiety connected to the
proline ring N-atom by bis-aryl-amide groups previously identified
by us and others in other approaches toward orexin receptor
antagonists. [4]
The initial series of compounds (30 to 34) contained a 2-triazolo-
5-methyl-benzoyl substituent attached to the ring nitrogen atom of
the different scaffolds, combined with a 4-methoxy-benzimidazole
moiety in α-position to the ring nitrogen atom. It showed that
heterocyclic scaffolds such as thiomorpholine (33) or morpholine
(34) resulted in a significant loss in orexin receptor antagonistic
activity. Comparing 33 and 34 to 36 and 37 confirmed the
influence of the substitution pattern at the benzimidazole moiety
on activity. Moving from a pyrrolidine core (30) to a piperidine core
(31) did not have a positive impact on activity either. The 2-
azabicyclo [2.2.1] heptane based 32 with 1R,3S,4S-chirality,
which can be looked at as 2-bridged pyrrolidines, with the bridge
from C3 to C5, resulted in a very potent DORA. Further
characterization of this compound was performed. Human liver
microsomal (HLM) stability measurement resulted in a value of
202 µL min-1 mg-1. The compound did not show significant P450
inhibition (3A4T IC50 = 18 µM (T = testosterone used as marker);
3A4M IC50 = 14 µM (M = midazolam used as a marker); 2C9 IC50
= 41 µM; 2D6 IC50 >50 µM) and had an acceptable IC50 shift in the
time dependent CYP3A4 assay of 3.9-fold. In our in-house
standard in vivo experiment where blood-brain-barrier (BBB)
penetration was assessed and the compound was dosed orally at
100 mg kg-1 to male Wistar rats, plasma exposure measured 3h
post administration was high with 6093 ng mL-1. However, the
corresponding brain exposure was only 682 ng g-1, resulting in a
B/P-ratio of 16%. Similar behavior was observed with other
derivatives of this series (low B/P-ratio), indicating a possible
brain penetration issue potentially due to Pgp-mediated efflux.
Compound 32 was confirmed to be a Pgp substrate in the human
MDR1 assay where the efflux ratio was 53 and PappAB was 1.1
10-6 cm.s-1. Even though the orexin receptor antagonist activity
was good with many derivatives (see also 35) it was decided to
abandon the template and return to the pyrrolidine core for further
optimization work. Another advantage of the pyrrolidine as a
template, was the commercial availability of additionally
substituted derivatives. A selected overview of this work is
depicted in Figure 8 with compounds 38 to 45, and with 38 serving
as the reference.
The introduction of a methylene substituent in position 4 (39) had
no influence on orexin receptor antagonist activity. Reducing the
methylene to a methyl group and having the R1 exit vector and
the substituent in a cis arrangement (40) was as well neutral with
respect to activity. Increasing the size of the substituent in position
4 to phenyl (42) resulted in a slightly more dual compound,
although the effect on antagonist activity was still limited.
Compounds 44 and 45, both containing a fluorine attached to
position 4 but of opposite chirality, showed no positive effect of
fluorine and indicated the limited influence of the chirality in this
position on activity. The same observation was made when
comparing 40 to 41, even though not only the sense of chirality
was different, but the substituents varied from methyl to methoxy.
The introduction of an additional methyl group in position 2 (43),
resulting in a quaternary carbon atom showed an almost 10-fold
increase in Ox1R antagonist potency combined with only a 2-3-
fold potency loss on the Ox2R. Furthermore, this additional methyl
stabilized the chiral center and prevented racemization in the
chemical transformations performed to construct the
benzimidazole moiety onto the 2-methyl-proline template.
Figure 6. Switch from sulfonamide to amide.
Compounds 26 and 27 represent early examples where
substituents present in 1 and 4 are differently linked to the proline
template (Figure 6). The orexin receptor antagonist activities
obtained were surprisingly moderate. The breakthrough came
when the bis-aryl-amide moieties attached to the ring N-atom of
the template were combined with the benzimidazole groups or the
phenyl-oxadiazole units. The results obtained with the
benzimidazole derivative 28 triggered our interest for a systematic
investigation of this class of orexin antagonists. Comparison of 28
with 11, bearing in mind the influence on orexin receptor
antagonist activity of the benzimidazole substitution pattern,
indicated the potential to identify promising molecules within this
group of compounds. Compound 29 confirmed the potential of an
aryl substituted oxadiazole moiety attached to position 2 of the
template. This part of our work was described separately. [39]
5
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
Figure 7. Core modification in the benzimidazole series of orexin receptor antagonists.
We checked for the influence of the chirality on orexin receptor
antagonist potency and found the S-enantiomer as being the
more potent one, in line with earlier observations. [38]
The results depicted in Table 1 gave a first insight into the effects
of mono-substitution at the benzimidazole part in positions 4 and
5 (or 6 and 7; depending on the benzimidazole tautomer drawn)
compared to the 5-position. In case of the methoxy substitution,
the 5-position resulted in the more active compound 49 as
compared to 48. The effect was more pronounced on the Ox2R
and almost negligible on the Ox1R. A single fluorine substituent
in either position (50 vs 51) resulted in derivatives of lower activity,
whereas a single chlorine- (52 and 53), bromine- (54 and 55) or a
(Compounds 46 to 62). The chirality of all compounds was (S) and, trifluoromethyl-substituent (56 and 57) in either position always
in order to put the results in the broadest possible perspective, the
substituent connected to the proline-ring nitrogen was chosen to
be the 2-triazolo-5-methyl-benzoyl moiety. For compounds 46
and 47 bearing one methyl substituent, substitution in the 4-
position resulted in a significantly more active compound as
resulted in derivatives more active toward the Ox1R. The
Ox2R/Ox1R selectivity ratio was between approximately 4-fold for
54 to 15-fold for 53 with best potencies on the Ox1R being around
11 nM.
Figure 8. Substitutions at the pyrrolidine-core.
6
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
Table 1. Modification of the benzimidazole substitution pattern
IC50 hOx1
(nM)[a]
19
IC50 hOx2
(nM)[a]
54
Compound
R1
R2
R3
R4
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
CH3
H
OCH3
H
F
H
Cl
H
Br
H
CF3
H
i-Pr
H
H
H
H
CH3
H
OCH3
H
F
H
Cl
H
Br
H
CF3
H
t-Bu
SO2CH3
OCF3
CN
CH3
F
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH3
F
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
OCH3
H
110
112
90
295
189
39
29
23
13
11
24
16
6
389
35
389
4
19
3
1.7
54
25
9
40
3.7
31
515
15
14
6
390
390
46
2030
931
197
413
83
129
76
129
47
30
1030
229
733
19
19
10
7
638
29
19
656
29
H
CH3
CH3
CH3
CH3
Cl
Br
F
F
[b]
[b]
H
H
H
H
H
H
H
H
H
CH3
F
Cl
Cl
31
OCH3
OCH3
[c]
[c]
1270
134
32
22
63
14
19
14
81
F
F
Cl
CF3
CF3
H
H
H
H
H
Cl
Br
CH3
F
5
1.8
14
8
27
6
H
Cl
CH3
CH3
Br
Cl
OCH3
CH3
CH3
Br
F
CF3
H
31
740
1.4
44
0.7
CH3
Br
[a] Data are the geometric mean (+- 2-fold) of at least three independent experiments. [b] Compound 68: R1 and R2 forming a benzodioxane moiety. [c]
Compound 73: R2 and R3 forming a benzodioxane moiety.
7
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
Larger alkyl substituents in positions 4 or 5, as shown for 58 and
59, resulted in very interesting potencies on the Ox1R as well as
on the Ox2R, asking for further investigations. Introducing a polar
substituent in 5-position, such as the methylsulfone group in 60,
yielded generally less active antagonists. The activity of the
trifluoromethoxy substituted 61 was comparable to the
trifluoromethyl analog 53. Another polar and electron withdrawing
substituent such as nitrile, resulted also in lower orexin receptor
antagonist activity (62). As a next step, di- and tri-substituted
benzimidazoles were prepared and profiled (Compounds 63 to
84).
The 4,5-dimethyl substitution pattern, as given in 63, surprisingly
resulted in a more active compound as one would expect from the
results obtained with 46 and 47. These results suggested that
substituent effects were not purely additive in this series (in both,
the positive or negative direction). Fixing the substituent in
position 4 to methyl and combining it with a halogen in position 5
(64 to 66) resulted in perfect duality in case of fluorine (64) and a
further gain in potency, especially in case of bromine (66). This
derivative and the chlorine analog 65 were further characterized
and are described in Table 2, Figure 11, and Table 6.
First in vitro characterization of the most interesting compounds
65, 66, 78, and 80 on human and rat liver microsomal stability and
P450 inhibition (Table 2) showed acceptable results with HLM
between 98 and 206 µL min-1 mg-1. The lowest P450 inhibition
measured on all 3A4T, 3A4M, 2C9 and 2D6 was with an IC50
=
4.5 µM and there was no major time dependent increase in
inhibition (max. 2.4-fold). In standard brain penetration
experiments where compounds were tested orally at 100 mg kg-1,
all 4 compounds showed good exposure (concentration in the
brain above 1000 ng mL-1) and brain penetration properties
(Brain/Plasma (B/P) ratio between 93% and 185%) at 3h post
administration (Table 2).
Table 2. Metabolic stability, cytochrome P450 inhibition and brain penetration
properties in rats of 65, 66, 78 and 80.
Parameters
65
66
186 / 947
11
78
98 / 276
7.9
80
155 / 879
8.4
HLM/RLM
clearance in
µL min-1 mg-1
206 /
>1250
Combining two fluorine substituents (67) did not have a similarly
beneficial effect as combining two methyl substituents (63).
Compound 68 bearing a 5,6-dihydro-benzodioxine moiety was a
balanced dual orexin receptor antagonist but not attractive
enough to be further investigated due to the benzodioxine moiety.
The results obtained by putting substituents in positions 5 and 6
are summarized by examples 69 to 78. Compound 69 exhibited
the same orexin receptor antagonist potency as its 3,4-analog 63.
The same was observed for the two bis-fluoro substituted analogs
67 and 70. Two chlorines (71) resulted in better antagonistic
activity but an almost 10-fold preference for the Ox1R remained.
In case of the 5,6-dihydro-benzodioxine moiety, the analog 73
was significantly less active as compared to 68. 5,6-dimethoxy-
substitution (72) resulted in perfect duality with borderline potency.
Still compound 72 was further characterized: HLM stability was
186 µL min-1 mg-1; P450 inhibition showed IC50 values >50 µM for
3A4T, 3A4M, 2C9 and 2D6, and time dependent increase in 3A4
inhibition was > 2.4-fold. The brain penetration experiment
performed in rats with sampling 3 h following oral administration
of 10 mg kg-1, 30 mg kg-1and 100 mg kg-1 suggested that 72 was
a Pgp substrate in rats. Indeed, the B/P-ratio increased from 13%
at 10 mg kg-1 ([P]=188 ng mL-1, [B]=26 ng g-1) to 44% at 100 mg
kg-1 ([P]=11667 ng mL-1, [B]=5060 ng g-1) reflecting a potential
saturation of the Pgp with increasing exposure. The efflux ratio in
the MDR1 assay with 72 was 49 and the permeability PappAB was
0.9 10-6 cm.s-1, showing that this compound was in fact also a
human Pgp substrate. The compound was abandoned.
Compounds 74 to 77 show further substituent combinations in
5,6-positions. These four derivatives showed excellent potency
toward the Ox1R and were less potent on the Ox2R, and therefore
not further characterized. From the subgroup of 4,6-disubstituted
benzimidazole moieties, 78 and 80 turned out to be the most
interesting derivatives and are extensively described in Table 2,
Figure 11, and Table 6. The 4,7-dimethoxy derivative 83 was not
active enough to be further investigated. In contrast, the 4,5-
dimethyl-6-bromo substitution pattern (84) yielded one of the most
potent orexin receptor antagonists identified in this study.
Unfortunately, the HLM stability was limited with 266 µL min-1 mg-
3A4T in µM
3A4 shift
11
2.4
12
1.7
2.1
2
3A4M in µM
2C9 in µM
2D6 in µM
17
>50
4.5
17
7.2
39
13
5.7
39
13
33
Brain penetration experiments in rats at 100 / 30 mg kg-1,
3h post administration [a]
2513 ± 770
/ 1190 ±
313
[plasma] [b]
In ng mL-1
1219 ± 142
/ 125 ± 48
1677 ± 253
/ nd
2095 ± 234
/ 723 ± 108
2348 ± 478
/ 1632 ±
608
3880 ±
1362 / 753
± 144
[brain] [b]
in ng g-1
2201 ± 221
/ 145 ± 47
2379 ± 561
/ nd
B/P ratio
181 / 117
142 / nd
49 / nd
93 / 137
185 / 104
[CSF] [c]
50 ± 3 /
3 ± 3
58 ± 27 /
28
49 ± 1 /
14 ± 5
in ng mL-1
[a] formulated in 100 % PEG400. [b] values given as mean ± SEM of n=3. [c]
values for n=1-3; nd: not determined
These compounds were pursued and tested in freely moving male
Wistar rats implanted with telemetry devices to continuously
record EEG/EMG as previously described. [38] First compounds to
be tested were 65 and 80. Our initial approach was to test
compounds orally at 100 mg kg-1 in a polyethylene glycol 400
(PEG400) formulation and, if efficacy was observed, to decrease
the dose to 30 mg kg-1. At 100 mg kg-1, over the 12 h night period
following administration, compounds 65 and 80 increased
significantly the time spent in NREM sleep by 29 min (p=0.0011)
and by 37 min (p=0.0004), respectively (Figure 9) compared to
matched vehicle. Both also increased the time spent in REM sleep
(+9 min and +30 min respectively), although statistically
significantly only for 80 (p=0.0082). At 30 mg kg-1, a similar pattern
was seen with a significant increase of the time spent in NREM
sleep of 17 min and 35 min for 65 (p=0.0448) and 80 (p=0.0051),
1
and the human plasma protein binding was very high with
>99.9%. Therefore 84 was as well abandoned.
8
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
respectively and a significant increase of REM sleep time for 80
(+13 min, p=0.0332) (Figure 9). Because efficacy was observed
at 30 mg kg-1 for both compounds, we decided to change our
starting dose for in vivo testing in EEG/EMG from 100 mg kg-1 to
30 mg kg-1.
6h period analysis, respectively). It suggested that the evaluation
and comparison of the subsequent compounds could be
performed at 30 mg kg-1 over the 6h period following
administration.
In this first series, all 4 compounds (65, 80, 66 and 78)
significantly increased the time spent sleeping following
administration of 30 mg kg-1, from 22 min (compound 65) to 45
min (compound 80) (Figure 11). Compound 80 showed the largest
increases in NREM sleep (+31 min vs vehicle) whereas
compound 78 had limited efficacy on NREM sleep (+13 min vs
vehicle).
The next step was the optimization of the amide substituent with
a few selected benzimidazoles. Results are summarized in Table
3Error! Reference source not found.. For this investigation and
based on previous in vivo results, we had selected the 4-methyl-
5-chloro-, the 5-trifluoromethyl-6-chloro-, the 5,6-dimethoxy- and
the 5-trifluoromethoxy-substitution patterns on the benzimidazole
part and combined them with a series of bi-aryl-carboxylic acid
derivatives with a focus on 2-triazolo substitution combined with
one or two additional substituents on the phenyl ring.
a
compound 65
30 mg/kg
100 mg/kg
50
40
**
30
*
20
10
0
NREM
REM
b
com pound 80
3 0 m g /k g
50
40
30
20
10
0
1 0 0 m g /k g
**
***
**
NREM sleep proportion in
% of total sleep time
77%
80%
82%*
73%**
80
60
40
20
0
*
NREM sleep
REM sleep
**
&&
**
*
*
N R E M
R E M
&&
&
#
##
##
Figure 9. Effect of 30 and 100 mg kg-1 oral administration of 65 and 80 on sleep
stages. Summary of the 12 h night active period following oral administration of
compound 65, a) and compound 80, b) formulated in 100% PEG400. Data are
expressed as mean ± SEM. *p<0.05, **p<0.01, ***p<0.001 vs matched vehicle,
paired t-test. n=7-8 per group. REM, rapid eye movement; NREM, non-REM.
##
65
80
66
78
Figure 11. Effect of 30 mg kg-1 of 65, 80, 66, and 78 on sleep stages. Summary
of the 6 h night active period following oral administration of compounds
formulated in 100% PEG400. Data are expressed as mean ± SEM. n=7-8 per
group. NREM sleep proportion is given for the compound treated group. REM,
rapid eye movement; NREM, non-REM. For total sleep time: *p<0.05, **p<0.01;
for NREM sleep: &p<0.05, &&p<0.01; for REM sleep: #p<0.05, ##p<0.01 each
vs matched vehicle, paired t-test.
100
100 mg kg-1
30 mg kg-1
**
80
60
40
20
0
**
**
**
In the group of compounds 85 to 88, wherein the 2-triazolo-5-
methyl-benzoyl moiety was replaced by a 2-triazolo-5-fluoro-
benzoyl moiety, 85 (containing the 4-methyl-5-chloro pattern on
the benzimidazole part), was clearly the most potent dual orexin
receptor antagonist and was further characterized as described in
Table 4, Figure 12 and Table 6 together with all the most
interesting compounds from Table 3. In the following group
bearing the 2-triazolo-5-chloro-benzoyl substituent, 92 was
abandoned due to low antagonist potency, and 90 because of its
preferential activity towards the Ox1R. Compound 89 which is
again the combination with the 4-methyl-5-chloro pattern in the
benzimidazole part was further characterized. Compound 91 was
further characterized in vitro. Stability in HLM was 90 µL min-1 mg-
1; P450 inhibition was 3A4T IC50 = 41 µM; 2C9 IC50 >50 µM; 2D6
IC50 >50 µM and time dependent increase in 3A4 inhibition was >
2.7-fold. In the in vivo brain penetration -experiment at 100 mg kg-
12h
period
6h
period
12h
6h
period
period
Figure 10. Effect of 80 on total sleep time. Summary of the 6 or 12 h night active
period following oral administration of compound 80 formulated in 100%
PEG400. Data are expressed as mean ± SEM. **p<0.01vs matched vehicle,
paired t-test. n=7-8 per group. REM, rapid eye movement; NREM, non-REM.
In addition, as expected with compound 80 (Figure 10), by
decreasing the dose, the duration of action decreased. Whereas
at 100 mg kg-1, a larger increase in sleeping time was seen when
looking at the data over the 12 h period (+67 min vs vehicle)
compared to the 6 h period (+44 min), at 30 mg kg-1, a very similar
increase was seen (+48 and +45 min vs vehicle for the 12h and
1
p.o., plasma and brain exposure were very high at 3h post
administration with 16907 ng mL-1 and 18827 ng g-1, respectively,
9
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
Table 3. Best benzimidazoles R1 combined with a set of benzoic acids as R2
IC50 (Ox1) in nM [a]
IC50 (Ox2) in nM [a]
Cmpd No
R2
R1
4.9
13
85
5.6
34
86
200
154
87
138
720
88
N.D.
N.D.
3
8
89
2.3
21
90
10
23
91
99
119
92
1.4
3.2
93
25
26
94
46
199
95
N.D.
N.D.
N.D.
1.9
3.1
96
1.5
4.5
97
15
25
98
5
16.5
99
12
25
100
11
45
101
872
458
102
529
1600
103
N.D.
N.D.
10
13
104
8.9
36
105
414
1200
106
197
1870
107
2.5
10
108
5.6
25
109
316
114
110
N.D.
N.D.
N.D.
N.D.
5.6
31
111
7.3
82
112
426
104
113
6
13
118
3.6
7.6
114
3
7.8
115
36
16
116
34
380
117
[a] Data are the geometric mean (+- 2-fold) of at least three independent experiments.
resulting in a B/P-ratio of 111%. However, the efflux ratio of 25
and the PappAB of 2.5 10-6 cm.s-1 in the MDR-1 assay, showed
that this compound was a Pgp substrate in human, confirming the
high likelihood of compounds combining two aromatic methoxy-
substituents of being human Pgp substrates, as previously
observed with 72. Compound 91 was therefore discarded. In the
next group containing the 2-triazolo-5-methoxy-benzoyl
substituent, 95 was not further characterized due to its relatively
low potency. Compound 93 (ACT-541468, Daridorexant), again
part of the 4-methyl-5-chloro-benzimidazole subgroup was very
promising and was further characterized. The remaining
compound 94 was further characterized since the 5,6-dimethoxy-
benzimidazole moiety resulted, so far, always in very good
exposure data in the rat. This was also true for this compound.
The in vivo brain penetration -experiment performed at 100 mg
kg-1 p.o. in rats resulted, 3h post administration, in the highest
exposure seen during the project so far, with plasma
concentration of 35640 ng mL-1 but a brain concentration of 11673
10
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
ng g-1, giving a B/P ratio of only 33%. This nurtured the suspicion
that, as previously seen in 5,6-dimethoxy-benzimidazole
containing compounds, 94 was a Pgp-substrate also in human.
Indeed, in the MDR-1 assay the efflux ratio was 63 and PappAB
chloro-benzimidazole member of this subgroup (104) resulted in
interesting antagonist potency on both orexin receptors and was
further characterized. As a next investigation, the phenyl ring of
the benzoyl substituents was changed into a pyridine ring,
was 0.8 10-6 cm.s-1, and as a result 94 was not further investigated. resulting in the 6-methyl-3-triazolyl-2-picolinoyl unit reflected in
Otherwise 94 was very stable in HLM with 14 µL min-1 mg-1 and
the IC50 values for all P450 investigated were > 50 µM. The
introduction of the 2-triazolo-4,5-dimethyl-benzoyl moiety resulted
in a subgroup 96 to 99, all exhibiting very good antagonist potency
on both the Ox1R and the Ox2R. Compound 99 showed a decent
HLM stability of 205 µL min-1 mg-1 and 3A4T inhibition of 26 µM.
The 5,6-dimethoxy-benzimidazole derivative 98 resulted in HLM
stability of 39 µL min-1 mg-1, did not inhibit any of the P450
measured (IC50 > 50 µM), but showed a time dependent increase
in inhibition of 3A4 > 9. It was therefore stopped. The 5-
trifluoromethyl-6-chloro-benzimidazole derivative 97 showed very
promising dual orexin antagonistic activity and was further
108 to 110. The results obtained in combination with the 4-methyl-
5-chloro-benzimidazole unit (108) were very interesting and also
this compound was further characterized. Compound 109
exhibited interesting receptor potency, was very stable in HLM (12
µLmin-1 mg-1)), and showed a reasonable P450 inhibition profile
with 3A4T IC50= 9.9 µM, 2C9 IC50= 4.5 µM and 2D6 IC50= 23 µM.
No time dependent increase in P450 inhibition was detected. The
compound was abandoned for further characterization as it was a
Pgp-substrate (Efflux ratio = 5.9; PappAB = 8.2 10-6 cm.s-1).
Compound 110 was as well not further investigated as it was not
potent enough on the orexin receptors. Combining the pyridine
ring with a pyrazole substituent resulted in 111 to 113 containing
characterized.
Compound
96,
the
4-methyl-5-chloro-
a
6-methyl-3-pyrazolyl-picolinyl substituent. Again, the
benzimidazole derivative resulted in HLM stability of 456 µL min-1
mg-1 and the 3A4T inhibition was at 6.8 µM. In the in vivo brain
penetration experiment in rats at 100 mg kg-1, the compound
could neither be detected in plasma nor in the brain 3h post
administration. The results of the MDR1 assay did not highlight a
major Pgp issue for human with an efflux ratio of 3.4 and PappAB
of 19.10-6 cm.s-1. This result indicated how subtle structural
changes can influence PK behavior. The next subgroup of
compounds containing the 2-triazolo-6-fluoro-benzoyl substituent
100 to 103 showed less antagonist potency on the Ox2R and was
not further characterized. Adding a methoxy group to obtain the
combination with the 4-methyl-5-chloro-benzimidazole moiety
resulted in the very promising compound 111 which will be
discussed in detail. Based on the Ox2R potency, 112 and 113
were not further characterized. The last group of compounds 114
to 118 contain a 5-methyl-2-pyrimidinyl-benzoyl moiety attached
to the proline ring nitrogen. Compounds 114 and 118, both very
potent dual orexin receptor antagonists, were further
characterized. Compound 116 was not further characterized
under the assumption it was a strong Pgp-substrate as the other
derivatives from the 5,6-dimethoxy-benzimidazole sub-series.
Compound 117 was abandoned based on its low orexin receptor
antagonist potency. In this investigation we found that the 4-
methyl-5-chloro-benzimidazole unit resulted in the most
promising dual orexin receptor antagonists in combination with
various substituents attached by an amide bond to the pyrrolidine
ring-nitrogen atom of the template.
2-triazolo-5-methoxy-6-fluoro-benzoyl
moiety, resulted in
compounds 104 to 107. The derivatives 106 and 107 exhibited
low orexin receptor antagonist potency and were not further
characterized. The 5-trifluoromethyl-6-chloro-benzimidazole
derivative 105 showed good HLM stability (48 µL min-1 mg-1) and
P450 inhibition in the 3A4T assay was 6.5 µM. The 4-methyl-5-
Table 4. Metabolic stability, cytochrome P450 inhibition and brain penetration properties in rats of the most interesting compounds from Table 3
Parameters
85
89
93
97
104
108
111
114
118
HLM/RLM
clearance in
µL min-1 mg-1
138 /
>1250
96 / >1250
125 / 645
138 / 990
197 / 618
64 / 1180
111 / 437
106 / 572
69 / 988
3A4T in µM
3A4 shift
5.6
1.3
3.6
8.1
33
6.2
1.3
8.5
5.5
23
4.1
2.1
7.3
14
6.1
1.9
> 50
5.7
21
8.4
2.1
14
16
> 1.7
11
15
1.2
12
12
1.5
28
18
2.1
41
3A4M in µM
2C9 in µM
2D6 in µM
7.7
30
24
30
19
7.5
29
31
> 50
>50
> 50
Brain penetration experiments in rats at 100 / 30 mg kg-1, 3h post administration [a]
9816 ±
2244 / 371
± 95
[plasma][b] in
ng mL-1
1354 ±152 /
nd
1205 ±307 /
273 ± 63
1280 ± 31 /
309 ± 42
786 ±220 /
321 ± 31
1144 ± 60 /
nd
2029 ± 513
/ 520 ± 135
1087 ± 103
/ 236 ± 21
3560 ± 602
/ 513 ± 65
13005 ±
4455/ 97 ±
33
5880 ±
1201 / 301
± 34
3507 ±
1133 / nd
2895 ± 595
/ 536 ± 136
1808 ± 116
/ 300 ± 63
1537 ±471 /
495 ± 37
1984 ± 256
/ nd
1999 ± 685
/ 293 ± 100
2230 ± 265
/ 153 ± 12
[brain][b] in ng g-1
B/P ratio
259 / nd
240 / 196
141 / 97
196 / 154
173 / nd
99 / 56
132 / 26
205 / 65
165 / 59
118 ± 36 /
nd
30 ± 5 /
5 ± 3
63 ±3 / 24 ±
5
10 ± 5 /
6 ± 2
242 ± 63 /
27
1862 /
7.2 ± 0.2
54 ± 7 /
25 ± 1
105 ± 37 /
10
[CSF] [c] in ng mL-1
49 ± 4 / nd
[a] formulated in 100 % PEG400. [b] values given as mean ± SEM of n=3. [c] values for n=1-3. nd: not determined.
11
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
First in vitro characterization of the most interesting compounds,
summarized in Table 4, were performed on human and rat liver
microsomal stability and P450 inhibition.
of the time spent in NREM (+7 to +12 min depending on
compounds) and REM sleep (+4 to +8 min) over the 6h period
following administration. This trend translated to a statistically
significant increase of the total time spent asleep for compound
111 only (+15 min of sleep vs vehicle). All the other compounds
(93, 97, 104, 114 and 118) showed statistically significant
increases of the time spent sleeping (from +25min for compound
104 to +53 min for compound 108 ( vs vehicle)) with a significant
increase of the time spent in NREM sleep (from +15 min for
compound 97 to +34 min for compound 108). In the 5 compounds,
only compound 104 did not increase in a statistically significant
manner the time spent in REM sleep (+10 min vs vehicle).
All compounds showed results on HLM between 64 and 197 µL
min-1 mg-1 and lowest P450 inhibition on all 3A4T, 3A4M, 2C9 and
2D6 at 3.6 µM with no time dependent inhibition of 3A4 (max. 2.1-
fold). In standard brain penetration experiments, where
compounds were tested orally at 100 mg kg-1, all 4 compounds
showed good exposure (concentration in the brain above 1500 ng
mL-1) and brain penetration properties (B/P ratio between 99 and
259%) at 3h post administration (Table 4).
When tested in an EEG experiment at 30 mg kg-1 (Figure 12), 85,
89 and 111 showed a trend but no statistically significant increase
NREM sleep proportion
in % of total sleep time
80
60
40
20
0
81%
77%
75%*
81%
76%
77%**
77%
74%*
81%
NREM sleep
REM sleep
***
**
***
**
&&
&&&
**
&&
**
&&&
&&&
*
&
###
##
##
#
#
85
89
93
97
104
108
111
114
118
Figure 12. Effect of 30 mg kg-1 of 85, 89, 93, 97, 104, 108, 111, 114, and 118 on sleep stages. Summary of the 6 h night active period following oral administration
of compounds formulated in 100% PEG400. Data are expressed as mean ± SEM. n=7-8 per group. NREM sleep proportion is given for the compound treated group.
REM, rapid eye movement; NREM, non-REM. For total sleep time: *p<0.05, **p<0.01, ***p<0.001; for NREM sleep: &p<0.05, &&p<0.01, &&&p<0.001; for REM
sleep: #p<0.05, ##p<0.01, ###p<0.001 each vs matched vehicle, paired t-test.
Based on the very interesting results obtained in the profiling of
93 (see Table 4, Figure 12, and Table 6), it was decided to
investigate a few additional benzimidazole substitution patterns in
combination with the 2-triazolo-5-methoxy-benzoyl moiety
attached via an amide linkage to the template ring nitrogen atom.
The results are depicted in Figure 13. All the compounds 119 to
123 showed decent potency, especially 119 and 120, wherein the
chloro substituent attached to the 5-position of the benzimidazole
system was replaced by bromo or fluoro, respectively.
Methylation at the benzimidazole moiety resulted in the very
potent dual orexin antagonist 123. By NMR-experiments, it was
shown that methylation or alkylation (see Figure 16)
predominantly results in the drawn regioisomer. Due to their lower
orexin receptor antagonist potency, 121 and 122 were not further
characterized. Compound 119 was further characterized and is
discussed in detail in Table 5, Figure 14, and Table 6. Further
characterization of 120 resulted in good HLM stability of 71 µL
min-1 mg-1), human PPB of 99.8% and an efflux ratio in the MDR1
assay of 4 with a permeability PappAB of 14 10-6 cm.s-1. With this
efflux ratio and because we had compounds with overall better
profile, the further evaluation of 120 was stopped. This targeted
investigation confirmed that the 4-methyl-5-chloro-pattern was the
best choice on the benzimidazole unit when combined with the 2-
triazolo-5-methoxy-benzoyl moiety.
Figure 13. Further substituent variations in combination with «triazolo-methoxy-benzoic acid».
12
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
penetration properties (Table 5). However, when tested in the
EEG experiment at 30 mg kg-1, the sleep promoting effect was
limited with no significant increase of the total sleep time (+19 min
vs vehicle), NREM (+12 min) or REM (+ 7 min)) (Figure 14).
Table 5. Metabolic stability, cytochrome P450 inhibition and brain penetration
properties in rats of compounds 119 and 139
Parameters
119
139
NREM sleep proportion
in % of total sleep time
80
60
40
20
0
77%
78%***
HLM/RLM clearance
in µL min-1 mg-1
134 / 658
149 / 835
3A4T in µM
3A4 shift
5.8
1.4
10
8.9
2.9
4.4
30
NREM sleep
REM sleep
**
3A4M in µM
2C9 in µM
2D6 in µM
&&
15
###
> 50
50
Brain penetration experiment in rats at 100 / 30 mg kg-1,
3h post administration [a]
119
139
[plasma][b] in ng mL-1
2196 ± 139 / nd
3173 ± 476 / nd
145
823 ± 73 / 222 ± 51
Figure 14. Effect of 30 mg kg-1 of compounds 119 and 139 on sleep stages.
Summary of the 6 h night active period following oral administration of
compound formulated in 100% PEG400. Data are expressed as mean ± SEM.
n=8. NREM sleep proportion is given for the compound treated group. REM,
rapid eye movement; NREM, non-REM. For total sleep time: **p<0.01; for
NREM sleep: &&p<0.01; for REM sleep: ###p<0.001 each vs matched vehicle,
paired t-test.
[brain][b] in ng g-1
B/P ratio
1009 ± 45 / 69 ± 31
123 /31
[CSF] [c] in ng mL-1
66 ± 13 / nd
42 / 1.4 ± 1.4
[a] formulated in 100 % PEG400. [b] values given as mean ± SEM of n=3-4. [c]
values for n=1-4. nd: not determined.
In Figure 15 we summarize further modifications of the carboxylic
acid moieties in combination with the 4-methyl-5-chloro-
benzimidazole.
Compound 119, the only compound further characterized from
Figure 13 showed an HLM stability at 134 µL min-1 mg-1. P450
inhibition was acceptable and in the brain penetration experiment
at the dose of 100 mg kg-1 the compound showed good brain
Figure 15. Further exploration of benzoic- and heteroaryl carboxylic acid moieties in combination with the 4-methyl-5-chloro-benzimidazole
13
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
First, we investigated substituted biphenyl groups as benzoyl-
substituents. Orexin antagonist activities, especially on the Ox2R,
showed a tendency towards weaker potency. Human liver
microsomal stabilities of 116 µL min-1 mg-1 and 103 µL min-1 mg-1
were determined for 124 and 125, respectively. P450 inhibition in
the 3A4T assay resulted in IC50 of 3.6 µM and 2.3 µM for 124 and
125, respectively. This, combined with the suboptimal Ox2R
antagonist activity, resulted in the decision to stop further work
with these compounds. Compound 126 did not show sufficient
orexin receptor antagonist potency to be of further interest.
Compound 127 showed HLM stability of 194 µL min-1 mg-1. In the
P450 assays, IC50 values were 7 µM for 3A4T, 2.6 µM for 2C9
and 6.6 µM for 2D6. Unfortunately, the time dependent increase
in 3A4 inhibition was very high with 23-fold. The 5-phenyl-
thiazoloyl derivative 128 was a very potent DORA with HLM
stability of 296 µL min-1 mg-1. For this compound, the P450
inhibition assays resulted in IC50 of 6.2 µM for 3A4T, 6.3 µM for
2C9 and 12 µM for 2D6. Unfortunately, here as well, the time
dependent increase in 3A4 inhibition was very high with 12.8-fold.
Another potent DORA was 129 with an additional chloro-
substituent on the phenyl ring. Human liver microsomal stability
was 285 µL min-1 mg-1, P450 inhibition was at the lower,
acceptable end (3A4T IC50 = 4.2 µM; 3A4M IC50= 14 µM; 2C9
IC50= 4.8 µM; 2D6 IC50 = 13 µM) and time dependent increase in
3A4 inhibition resulted in 2.9-fold. In the in vivo brain penetration
-experiment at 100 mg kg-1, plasma concentrations of 558 ng mL-
1 and brain concentrations of 607 ng g-1 were observed 3h post
administration, resulting in a B/P-ratio of 109%. Because of the
relatively limited exposure compared to other compounds tested
up to now, this compound was tested at 100 mg kg-1 p.o. to assess
its sleep promotion potential in rats. Compound 129 increased the
time spent in NREM sleep by 29 ± 12 and in REM sleep by 17 ±
4 min over the 12h post administration period. However, statistical
significance was reached only for REM sleep (p = 0.0520 and p =
0.0039 compared to vehicle for NREM and REM sleep,
respectively, paired t-test). In addition, it showed a very high
human plasma protein binding of >99.9%.
Compounds 130 to 133, containing a 2,5-diphenyl-thiazol-4-oyl
substituent, represent rather Ox1R preferential or selective
antagonists. Therefore, these compounds were not further
characterized in this study.
In Figure 16, a summary of selected N-alkylated benzimidazoles
is depicted. Based on its very good dual orexin receptor
antagonist potency, 134 was further characterized. HLM stability
resulted in 99 µL min-1 mg-1. P450 inhibition was acceptable
(3A4T IC50 = 9.1 µM; 3A4M IC50 = 12 µM; 2C9 IC50 = 17 µM; 2D6
IC50 = 49 µM) and time dependent increase in 3A4 inhibition was
1.7-fold. In the in vivo brain penetration experiment at 100 mg kg-
1
p.o., plasma concentrations were 1303 ng mL-1 and brain
concentrations 508 ng g-1 (ratio of 39%) indicating the potential of
this compound being a Pgp-substrate. In addition, human PPB
was > 99.9%.
Compounds 135 to 138 were not further characterized due to
limited orexin receptor antagonist potency. Compound 140, a
highly potent DORA resulted in an inacceptable time dependent
increase of 3A4-inhibition of 5 and was therefore abandoned as
well.
Finally, compound 139 (Figure 16), showed an HLM stability at
149 µL min-1 mg-1. P450 inhibition was acceptable and the in vivo
brain penetration experiment at the dose of 100 mg kg-1 resulted
in good brain penetration properties (Table 5). When tested at 30
mg kg-1 in the EEG experiment, 139 significantly increased the
total sleep time (+ 40 min vs vehicle) via an increase of 24 min of
NREM sleep and 16 min REM sleep (Figure 14).
Out of the investigations summarized so far, the interesting “15
good compounds” which were investigated in the EEG/EMG
experiment at 30 mg kg-1 were further characterized in order to
select the best one for preclinical and clinical development.
Figure 16. N-alkylated benzimidazoles (see also cpd 118).
Whereas many compounds showed promising in vivo efficacy in
rats, the overall profile of the compounds was looked at to select
the most promising drug candidate for clinical development.
Physiology-based pharmacokinetic and pharmacodynamic
(PBPK/PD) modeling was then used to identify those compounds
with the best PK/PD profile in man. [1]
14
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
Table 6. The 15 compounds investigated for their sleep promotion potential in rats
Parameters
65
66
78
80
85
89
93
97
Kb hOx1R / Kb hOx2R
in nM
0.8 / 1.1
22
0.5 / 1
33**
3.1 / 2.8
29*
8.8
/ 13
2.5 / 3.5
14
1.1 / 1.6
19
0.7 / 1.2
37**
3.4 / 5.4
[a] Increase in total sleep time in
min
45**
30**
*
MDR1 assay:
Efflux
2.7
21
3.4
16
2.3
28
2.2
31
2.1
30
2.3
33
3.3
24
2.3
18
PappAB in 10-6 cm.s-1
PPB in %
Human / Rat / Dog
99.94 / 99.4 / 99.6
>99.9 / 99.2 /99.5
99.92 / 99.6 / 99.6
99.8 / 99.3 / 99.7
99.6 / 99.2 / 99.7
99.92 / 99.4 / 99.7
99.9 / 98.3 / 99.4
>99.9 / 99.8
hPXR activation:
EC50 in µM
3.6
5.3
2.3
3.3
1.6
6
3.5
5.4
1.8
7.1
1.9
7
4
5.2
0.7
9.7
Emax in %
Stability in liver microsomes
in µL min-1 mg-1
Human / Rat / Dog
206 / >1250 / 335
186 / 947 / 267
98 / 276 / 55
155 / 879 / 80
96 / >1250 / 88
125 / 645 / 87
138 / 990 / 235
197 / 618 / n.d.
n.d.
i.v. PK in rat: 1 mg kg-1
AUC in ng*h/mL
CL in mL/(min*kg)
Vss in L/kg
325
51
3.8
3.2
113
150
7.9
0.9
461
36
11
498
33
2.2
2.2
553
30
3.6
2.7
196
85
3.9
1.3
Not performed
based on EEG
inactivity
T1/2 in h
6.2
[b] oral PK in rat: 10 mg kg-1
Cmax in ng mL-1
Tmax in h
AUC in ng*h/mL
F in %
i.v. PK in dog: 1 mg kg-1
AUC in ng*h/mL
CL in mL/(min*kg)
Vss in L/kg
126
0.5
208
6.4
261
0.5
240
319
0.5
1590
34
400
0.5
562
11.3
415
0.5
592
11
174
0.5
186
9.4
Not performed
based on EEG
inactivity
n.d.
n.d.
1690
9.9
3.7
1500
11
2.5
2070
8.1
5
1930
8.6
3.6
1160
14
5.4
1750
9.5
2.4
5
Not performed
based on EEG
inactivity
T1/2 in h
6.5
5.4
9.7
6.4
6.8
[b] oral PK in dog: 30 mg/dog
Cmax in ng mL-1
Tmax in h
247
0.3
1090
38
278
0.5
690
23
68
2
626
20
384
0.3
1560
42
Not performed
based on EEG
inactivity
85
1.0
355
17
239
1
910
35
n.d
1.1
AUC in ng*h/mL
F in %
Brain Fu in %
0.7
0.3
0.2
0.5
0.9
0.3
1
[a] 30 mg kg-1 p.o. formulated in 100% PEG400. [b] formulated in MC0.5%. *p<0.05, **p<0.01, ***p<0.001 vs matched vehicle, hPXR: human pregnane X receptor, n.d.: not determined
15
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
Table 6. continued
Parameters
104
108
111
114
119
118
139
Kb hOx1R / Kb hOx2R
in nM
5.2 / 8.1
25**
4.0 / 4.6
53***
3.8 / 4.5
15*
7.7 / 1.9
35**
1.2 / 1.4
19
1.6 / 5.5
36***
1.0 / 1.9
40**
[a] Increase in total sleep time
in min
MDR1 assay:
Efflux
2.9
16
8.4
9.7
13
4.5
2.5
28
2.2
23
1.4
34
4.7
20
PappAB in 10-6 cm.s-1
PPB in %
Human / Rat / Dog
99.6 / 98.2 / 99.4
99.7 / 97.3 / 97.5
99.8 / 97.2 / 98.6
>99.9 / 98.8 / 99.2
>99.9 / 98.8 / 99.2
99.2 / 99.1 / 99.2
99.97 / 98.7 / 98.9
hPXR activation:
EC50 in µM
2.6
5.1
2.3
5.2
1.6
6.7
1.9
7.1
0.9
4.1
0.4
4
4.6
5.5
Emax in %
Stability in liver microsomes
in µL min-1 mg-1
Human / Rat / Dog
i.v. PK in rat: 1 mg kg-1
AUC in ng*h/mL
CL in mL/(min*kg)
Vss in L/kg
64 / 1180 / 156
n.d.
111 / 437 / 65
106 / 572 / 68
n.d.
138 / >1250 / 182
134 / 658 / 185
69 / 988 / 89
149 / 835 / 212
767
22
0.6
0.4
616
27
1.2
1.3
461
36
3.2
1.7
524
32
0.79
0.4
n.d.
T1/2 in h
#
[b] oral PK in rat: 10 mg kg-1
Cmax in ng mL-1
Tmax in h
846
0.3
1170
15
464
0.3
965
16
121
0.5
212
n.a.
1050
0.5
1990
43
346
1.0
911
?
n.d.
n.d.
n.d.
n.d.
AUC in ng*h/mL
F in %
i.v. PK in dog: 1 mg kg-1
AUC in ng*h/mL
CL in mL/(min*kg)
Vss in L/kg
1900
8.8
1.2
2160
7.7
2.5
4220
3.9
2.2
2260
7.4
1.1
n.d.
T1/2 in h
1.5
5.5
8.5
2.9
[b] oral PK in dog: 30 mg/dog
Cmax in ng mL-1
Tmax in h
747
0.3
1620
49
211
0.8
818
20
736
0.25
2750
39
774
0.5
2130
28
AUC in ng*h/mL
F in %
n.d.
1.1
n.d.
2.5
n.d.
n.d.
Brain Fu in %
3
0.8
0.7
0.8
[a] 30 mg kg-1 p.o. formulated in 100% PEG400. [b] formulated in MC0.5%. *p<0.05, **p<0.01, ***p<0.001 vs matched vehicle, hPXR: human pregnane X receptor, n.d.: not determined, # formulated in 100% PEG400
16
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
Based on limited efficacy on either total sleep time, NREM sleep
or REM sleep time at 30 mg kg-1 compared to the other
compounds, 78, 85, 89, 104, 111, and 119 were discarded. The
lack of efficacy of 111 was associated with a strong decrease of
exposure and a decrease in B/P ratio in the brain penetration
experiment when decreasing the dose from 100 to 30 mg kg-1,
suggesting this compound could be a Pgp substrate in rats.
Looking at the MDR1 data, 111 was a clear Pgp substrate in
human with an efflux ratio of 13. Compound 108 was also
classified as a Pgp substrate in human based on the MDR1 data
showing an efflux ratio of 8.4 and was discarded. 139, with is
efflux ratio of 4.7 in combination with very a high hPPB of 99.97%
was discarded. Other liabilities appeared for some compounds
when tested for instance for in the human pregnane X receptor
(hPXR) assay reflecting induction of cytochrome P450 and
playing a role in detoxification. Indeed, compound 97, 119 and
118 were discarded based on an EC50 < 1 µM in this assay.
Compounds with non-measurable hPPB, as observed with
compound 66 or compound 114 were not favored either. In the
following, the detailed pharmacological characterization of our
remaining frontrunner 93 is described.
episode lasting at least 30 s) were decreased (one-way ANOVA,
p < 0.0001 and p = 0.0065, respectively). Latency to NREM sleep
decreased from 61 min for the vehicle-treated rats to 17 min at 10
mg kg-1 (p = 0.0006) and down to 11 min at 300 mg kg-1 (p <
0.0001) and latency to REM sleep from 63 min to 25 min at 30 mg
kg-1 (shorter latency observed, p=0.0037). Non-REM and REM
sleep increased in physiological proportion, i.e. under 93 x HCl
treatment, sleep architecture was conserved, as the relative
proportions of non-REM and REM sleep over the total sleep time
were not significantly changed (71 to 78% NREM, one-way
ANOVA, p = 0.1689).
Compound 93 x HCl was also evaluated in freely moving male
Beagle dogs implanted with telemetry transmitters for EEG/EMG
recording. Dogs received a single oral dose of 0, 10, 30, or 90 mg
of 93 x HCl as an amorphous solid dispersion in gelatin capsules
during their day active phase (Figure 18).
active w ake
quiet w ake
N R E M sleep
R E M sleep
5 9 .5
2 8 .7
8 .5
2 .2
p la ce b o
1 0 m g
3 0 m g
9 0 m g
*
A dose-response of the hydrochloride salt of 93 was performed in
male Wistar rats implanted with telemetry transmitters for
EEG/EMG recording to evaluate its effect on the sleep-wake cycle
(Figure 17). Compound 93 x HCl was formulated in 0.5%
methylcellulose and single oral doses of 93 x HCl, equivalent to 0,
10, 30, 100, or 300 mg kg-1 of the free base were administered by
gavage at the beginning of the nocturnal active phase.
4 9 .1
5 1 .5
2 8 .7
1 7 .3
4 .5
5 .1
*
1 9 .2
2 3 .4
2 7 .5 * *
***
1 1 .0
4 6 .4
2 0 .3
2 2 .6
9 .9
1 6 .7
2 2 .7
4 8 .8
1 1 .4
P hysio lo g ica l nig ht
active wake
quiet wake
NREM sleep
REM sleep
Figure 18. Effect of 93 x HCl on the time spent in sleep and wake stages (% of
total time) in dogs. Summary of the 6 h day active period following oral
administration of 93 x HCl formulated as an amorphous solid dispersion and
given in gelatin capsules. Data are expressed as mean ± SEM. One-way
ANOVA followed by the post-hoc Dunnett's multiple comparisons test: * p < 0.05,
** p < 0.01, *** p < 0.001 compared to placebo-treated dogs. n = 6-10/dose
group. MC, methycellulose; REM, rapid eye movement; NREM, non-REM.
vehicle
10 mg kg-1
30 mg kg-1
100 mg kg-1
300 mg kg-1
39.9
35.3
35.9
35.5
35.1
34.2
30.1
18.0
6.2
7.4
*
21.7
***
***
**
31.1
24.0
9.8
***
2*7*.5*
***
1*1.*0*
27.3
***
*
27.3
33.4
9.2
Over the 6 h period following administration, 93 x HCl dose-
dependently decreased total wake time (one-way ANOVA; p =
0.0001) by up to 77 min. This decrease in wakefulness was
accompanied by increases in both REM (up to 28 min) and NREM
(up to 51 min) sleep times (one-way ANOVA; p = 0.0004 and p =
0.002). The effect of the 10 mg/dog dose was statistically
significant vs placebo on NREM sleep variables (+ 32 min, Figure
18). In addition, the latency to the first persistent episode of NREM
sleep (the first NREM sleep episode lasting at least 60 s) was
decreased (one-way ANOVA, p = 0.0448) by 51 to 68 minutes
depending on the dose. The decrease only reached statistical
significance for the 30 mg dose where the latency decreased from
104 min for the placebo-treated dogs to 35 min (p=0.0451).
Similarly to observations in rats, the relative proportions of NREM
and REM sleep over the total sleep time were not significantly
changed (71 to 81% NREM, one-way ANOVA, p = 0.6058).
Compound 93 was tested in a panel of over 130 enzymes and
radioligand binding assays for established central and peripheral
pharmacological targets. Compound 93 showed no relevant
activity against these targets at the tested concentration of 10 µM.
Based on the excellent pharmacology results obtained in rats and
dogs, the outcome of the first safety screen and the results of the
PB-PKPD modelling work [1], it was decided to develop compound
93 (daridorexant).
15.3
33.5
40.4
10.8
physiological day
Figure 17. Effect of 93 x HCl on the time spent in sleep and wake stages (% of
total time) in rats. Summary of the 6 h night active period following oral
administration of 93 x HCl formulated in MC0.5%. Doses calculated for free
base. Data are expressed as mean ± SEM. One-way ANOVA followed by the
post-hoc Dunnett's multiple comparisons test: * p < 0.05, ** p < 0.01, *** p <
0.001 compared to vehicle-treated rats. n = 8/dose group. MC, methycellulose;
REM, rapid eye movement; NREM, non-REM.
Over the 6h period following administration, compound 93 x HCl
dose-dependently decreased total wake time (one-way analysis
of variance (one-way ANOVA; p < 0.0001) by up to 66 min, while
increasing REM (up to 17 min) and NREM (up to 55 min) sleep
times (one-way ANOVA; p = 0.0008 and p < 0.0001, respectively).
The effect of the 10 mg kg-1 dose was statistically significant vs
vehicle on NREM sleep variables (+13 min, Figure 17); the lowest
dose that was statistically significant on all variables was 30 mg
kg-1. In addition, the latency to the first persistent episode of
NREM sleep (the first NREM sleep episode lasting at least 60 s)
and the first persistent episode of REM sleep (the first REM sleep
17
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
Synthesis of 93 (daridorexant) and analogs (Figure 19):
(daridorexant) required the synthesis of the specific 2-triazolo-5-
methoxy-benzoic acid H. This was achieved in 76% yield, by
reacting the commercially available 2-iodo-5-methoxy-benzoic
acid in a copper catalyzed coupling in the presence of cesium
carbonate as base, in DMF at 80°C with common 1,2,3-triazole.
In the final step, 2-triazolo-5-methoxy-benzoic acid (H) is coupled
with intermediate E in dichloromethane, with DIPEA as the base
and HATU as the coupling reagent to yield 93 in 63%. The HCl-
salt of 93 was prepared by adding 1 equivalent of 1.6 M HCl in
dioxane solution to a slurry of 93 in dioxane followed by
evaporation of the solvent and drying of the final material under
high vacuum conditions.
In the first step, the medicinal chemistry synthesis of compound
89 starts with commercially available N-Boc protected (S)-2-
methyl-proline (A) which is reacted with 3-methly-4-chloro
phenylenediamine (B) under standard amide bond forming
conditions in dichloromethane using DIPEA as a base and HATU
as the coupling reagent. The reaction runs at room temperature
over night to give the intermediate C as two regioisomers C1 and
C2 in quantitative yield, which upon cyclization under acidic
conditions at 100°C in 1 hour resulted in the same benzimidazole
derivative D in 49% yield. Boc deprotection by applying standard
conditions (HCl in dioxane, 1.6M) at room temperature resulted in
the HCl-salt of key intermediate E. The preparation of 93
Figure 19. Synthesis of benzimidazole-proline based dual orexin receptor antagonists: Synthesis of 93 (daridorexant) is given as exemplary access to the
compounds; all the other compounds described before can be prepared accordingly by adapting the starting materials. Detailed synthetic information for 93 on kg-
scale can be found in the supporting information and general procedures in WO 2013/182972.
receptor antagonists. The major structural changes from 7 to the
new series, involved replacing the anilide moiety by
a
benzimidazole unit and the sulfonamide moiety by triazolyl-
substituted benzoyl unit. Extensive medicinal chemistry work,
combined with significant characterization efforts in in vivo
pharmacology of a large number of potent dual orexin receptor
antagonists, further in vitro characterization for absorption,
Conclusion
Taking ACT-462206 (7) [38], a proline based dual orexin receptor
antagonist, as a starting point for our medicinal chemistry program,
we identified a novel series of benzimidazole based orexin
18
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
[6]
a) C. Wang, Q. Wang, B. Ji, Y. Pan, C. Xu, B. Cheng, B.
Bai, J. Chen, Front Mol Neurosci 2018, 11, 220; b) J. P.
Kukkonen, C. S. Leonard, Br J Pharmacol 2014, 171, 314-
331.
I. Z. Shahid, A. A. Rahman, P. M. Pilowsky, Vitam Horm
2012, 89, 159-184.
a) J. N. Marcus, C. J. Aschkenasi, C. E. Lee, R. M.
Chemelli, C. B. Saper, M. Yanagisawa, J. K. Elmquist, J
Comp Neurol 2001, 435, 6-25; b) P. Trivedi, H. Yu, D. J.
MacNeil, L. H. Van der Ploeg, X. M. Guan, FEBS Lett
1998, 438, 71-75; c) L. de Lecea, Prog Brain Res 2012,
198, 15-24.
a) A. Eban-Rothschild, L. Appelbaum, L. de Lecea,
Neuropsychopharmacology 2018, 43, 937-952; b) N.
Nevarez, L. de Lecea, F1000Res 2018, 7.
B. Latifi, A. Adamantidis, C. Bassetti, M. H. Schmidt, Front
Neurol 2018, 9, 790.
M. D. Thompson, H. Xhaard, T. Sakurai, I. Rainero, J. P.
Kukkonen, Front Neurosci 2014, 8, 57.
a) G. E. Callander, M. Olorunda, D. Monna, E.
Schuepbach, D. Langenegger, C. Betschart, S.
Hintermann, D. Behnke, S. Cotesta, M. Fendt, G. Laue, S.
Ofner, E. Briard, C. E. Gee, L. H. Jacobson, D. Hoyer,
Front Neurosci 2013, 7, 230; b) D. Hoyer, T. Durst, M.
Fendt, L. H. Jacobson, C. Betschart, S. Hintermann, D.
Behnke, S. Cotesta, G. Laue, S. Ofner, E. Legangneux, C.
E. Gee, Front Neurosci 2013, 7, 235.
distribution, metabolism and excretion (ADME) and safety
assessments, resulted in 15 promising DORA compounds.
Identification of the best compound out of the 15 frontrunners was
[1]
finally performed by implementing a PB-PKPD model
and
[7]
[8]
assessing additional safety aspects in vitro. The efforts
culminated in the selection of compound 93, now called
daridorexant, for preclinical followed by clinical development.
Compound 93 dose-dependently promotes sleep in rats and dogs.
The sleep architecture (=relative proportions of REM vs NREM
sleep) obtained following administration of 93 is comparable to
the architecture of physiological sleep. Compound 93 is in Phase
3 clinical trials for the treatment of insomnia. Top line results of
the first pivotal phase 3 study were released on April 20th, 2020.
The study demonstrated that daridorexant improved overall sleep
and daytime performance of patients with insomnia. Results of the
second pivotal study clearly confirmed the results from the first
pivotal study.
[9]
[10]
[11]
[12]
Experimental Section
Experimental Details are given in the supporting information.
[13]
a) T. Nagahara, T. Saitoh, N. Kutsumura, Y. Irukayama-
Tomobe, Y. Ogawa, D. Kuroda, H. Gouda, H. Kumagai, H.
Fujii, M. Yanagisawa, H. Nagase, J Med Chem 2015, 58,
7931-7937; b) Y. Irukayama-Tomobe, Y. Ogawa, H.
Tominaga, Y. Ishikawa, N. Hosokawa, S. Ambai, Y.
Kawabe, S. Uchida, R. Nakajima, T. Saitoh, T. Kanda, K.
Vogt, T. Sakurai, H. Nagase, M. Yanagisawa, Proc Natl
Acad Sci U S A 2017, 114, 5731-5736; c) M. Suzuki, H.
Yukitake, T. Ishikawa, H. Kimura, 2018; d) H. Yukitake, T.
Ishikawa, A. Suzuki, Y. Shimizu, M. Nakashima, T.
Fujimoto, K. Rikimaru, M. Ito, M. Suzuki, H. Kimura, 2018;
e) H. NAGASE, T. SAITOH, M. YANAGISAWA, Y.
IRUKAYAMA, WO2019117148 2019; f) Y. KAJITA, S.
MIKAMI, Y. MIYANOHANA, T. KOIKE, M. DAINI, N.
OYABU, M. OGINO, K. TAKEUCHI, Y. ITO, N.
TOKUNAGA, T. SUGIMOTO, T. MIYAZAKI, T. ODA, Y.
HOASHI, Y. HATTORI, K. IMAMURA, WO2019027058
2019.
Acknowledgements
Cédric Fischer, Hélène Roellinger, Hélène Massinet, Malvina
Louis, Celia Müller-Grandjean, Katalyn Menyhart, Thomas Sasse,
Pascal Rebmann, Daniela Krüsi, Ursula Fusco, Marc Candreia,
Amandine Boller, Audrey Chardonneau, Daniel Hafner, Christian
Barten, Pauline Ligibel, Stephanie Bazire, Jeanine Thanner,
Markus Gude, Viktor Ribic, Stephane Delahaye, Isabelle Weber,
Fabienne Drouet, Eric Soubieux, Rolf Wuest, Julia Khobzaoui,
Aude Weigel, Kristina Kamin, Bruno Capeleto, Pascal Fullhardt,
Elise Jacob, Christine Metzger, Sandrine Chopinet, Amandine
Fraichard, Michelle Stutz, Maria Ferrand, Patric Dörrwächter,
Marco Tschanz, Alice Prud’homme for expert technical support
and Francois Jenck, Martin Bolli, Thomas Weller, Martine Clozel
and Cedric Vailant for expert discussions..
[14]
[15]
J. Yin, J. C. Mobarec, P. Kolb, D. M. Rosenbaum, Nature
2015, 519, 247-250.
J. Yin, K. Babaoglu, C. A. Brautigam, L. Clark, Z. Shao, T.
H. Scheuermann, C. M. Harrell, A. L. Gotter, A. J.
Roecker, C. J. Winrow, J. J. Renger, P. J. Coleman, D. M.
Rosenbaum, Nat Struct Mol Biol 2016, 23, 293-299.
R. Suno, K. T. Kimura, T. Nakane, K. Yamashita, J. Wang,
T. Fujiwara, Y. Yamanaka, D. Im, S. Horita, H. Tsujimoto,
M. S. Tawaramoto, T. Hirokawa, E. Nango, K. Tono, T.
Kameshima, T. Hatsui, Y. Joti, M. Yabashi, K. Shimamoto,
M. Yamamoto, D. M. Rosenbaum, S. Iwata, T.
Shimamura, T. Kobayashi, Structure 2018, 26, 7-19 e15.
M. Rappas, A. A. E. Ali, K. A. Bennett, J. D. Brown, S. J.
Bucknell, M. Congreve, R. M. Cooke, G. Cseke, C. de
Graaf, A. S. Dore, J. C. Errey, A. Jazayeri, F. H. Marshall,
J. S. Mason, R. Mould, J. C. Patel, B. G. Tehan, M. Weir,
J. A. Christopher, J Med Chem 2020, 63, 1528-1543.
L. Gershell, Nat Rev Drug Discov 2006, 5, 15-16.
D. Riemann, K. Spiegelhalder, B. Feige, U. Voderholzer,
M. Berger, M. Perlis, C. Nissen, Sleep Med Rev 2010, 14,
19-31.
S. Schutte-Rodin, L. Broch, D. Buysse, C. Dorsey, M.
Sateia, J Clin Sleep Med 2008, 4, 487-504.
C. M. Morin, C. L. Drake, A. G. Harvey, A. D. Krystal, R.
Manber, D. Riemann, K. Spiegelhalder, Nat Rev Dis
Primers 2015, 1, 15026.
M. K. Stoller, Clin Ther 1994, 16, 873-897; discussion 854.
E. M. Wickwire, F. T. Shaya, S. M. Scharf, Sleep Med Rev
2016, 30, 72-82.
Keywords: drug design, orexin receptor antagonists, insomnia,
[16]
[17]
G protein-coupled receptors, structure-activity relationship
[1]
[2]
A. Treiber, R. de Kanter, C. Roch, J. Gatfield, C. Boss, M.
von Raumer, B. Schindelholz, C. Muehlan, J. van Gerven,
F. Jenck, J Pharmacol Exp Ther 2017, 362, 489-503.
a) 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. 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, M.
Yanagisawa, Cell 1998, 92, 573-585; b) L. de Lecea, T. S.
Kilduff, C. Peyron, X. Gao, P. E. Foye, P. E. Danielson, C.
Fukuhara, E. L. Battenberg, V. T. Gautvik, F. S. Bartlett,
2nd, W. N. Frankel, A. N. van den Pol, F. E. Bloom, K. M.
Gautvik, J. G. Sutcliffe, Proc Natl Acad Sci U S A 1998,
95, 322-327.
[18]
[19]
[20]
[21]
[3]
C. Peyron, D. K. Tighe, A. N. van den Pol, L. de Lecea, H.
C. Heller, J. G. Sutcliffe, T. S. Kilduff, J Neurosci 1998, 18,
9996-10015.
[4]
[5]
J. Gatfield, C. Brisbare-Roch, F. Jenck, C. Boss,
ChemMedChem 2010, 5, 1197-1214.
T. Matsuki, T. Sakurai, Results Probl Cell Differ 2008, 46,
27-55.
[22]
[23]
19
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
[24]
[25]
[26]
[27]
D. Riemann, M. L. Perlis, Sleep Med Rev 2009, 13, 205-
214.
D. J. Nutt, S. M. Stahl, J Psychopharmacol 2010, 24,
1601-1612.
J. MacFarlane, C. M. Morin, J. Montplaisir, Clin Ther 2014,
36, 1676-1701.
L. C. Toner, B. M. Tsambiras, G. Catalano, M. C.
Catalano, D. S. Cooper, Clin Neuropharmacol 2000, 23,
54-58.
[28]
[29]
a) W. Yang, M. Dollear, S. R. Muthukrishnan, Arch Phys
Med Rehabil 2005, 86, 1265-1266; b) N. Gunja, J Med
Toxicol 2013, 9, 163-171; c) R. H. Farkas, E. F. Unger, R.
Temple, N Engl J Med 2013, 369, 689-691.
a) C. Drahl, Chemical & Engineering News 2014, 92, 52-
55; b) P. Murphy, M. Moline, D. Mayleben, R. Rosenberg,
G. Zammit, K. Pinner, S. Dhadda, Q. Hong, L. Giorgi, A.
Satlin, J Clin Sleep Med 2017, 13, 1289-1299; c) Y.
Yoshida, Y. Naoe, T. Terauchi, F. Ozaki, T. Doko, A.
Takemura, T. Tanaka, K. Sorimachi, C. T. Beuckmann, M.
Suzuki, T. Ueno, S. Ozaki, M. Yonaga, J Med Chem 2015,
58, 4648-4664.
[30]
[31]
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, F.
Jenck, Nat Med 2007, 13, 150-155.
P. Hoever, G. Dorffner, H. Benes, T. Penzel, H. Danker-
Hopfe, M. J. Barbanoj, G. Pillar, B. Saletu, O. Polo, D.
Kunz, J. Zeitlhofer, S. Berg, M. Partinen, C. L. Bassetti, B.
Hogl, I. O. Ebrahim, E. Holsboer-Trachsler, H. Bengtsson,
Y. Peker, U. M. Hemmeter, E. Chiossi, G. Hajak, J.
Dingemanse, Clin Pharmacol Ther 2012, 91, 975-985.
a) K. M. Connor, E. Mahoney, S. Jackson, J. Hutzelmann,
X. Zhao, N. Jia, E. Snyder, D. Snavely, D. Michelson, T.
Roth, W. J. Herring, Int J Neuropsychopharmacol 2016,
19; b) 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, J. J. Renger, ChemMedChem
2012, 7, 415-424, 337.
[32]
[33]
M. A. Letavic, P. Bonaventure, N. I. Carruthers, C.
Dugovic, T. Koudriakova, B. Lord, T. W. Lovenberg, K. S.
Ly, N. S. Mani, D. Nepomuceno, D. J. Pippel, M. Rizzolio,
J. E. Shelton, C. R. Shah, B. T. Shireman, L. K. Young, S.
Yun, J Med Chem 2015, 58, 5620-5636.
[34]
[35]
T. E. Scammell, E. Arrigoni, J. O. Lipton, Neuron 2017, 93,
747-765.
A. Palasz, D. Lapray, C. Peyron, E. Rojczyk-Golebiewska,
R. Skowronek, G. Markowski, B. Czajkowska, M.
Krzystanek, R. Wiaderkiewicz, Int J
Neuropsychopharmacol 2014, 17, 157-168.
[36]
a) P. Hoever, S. de Haas, J. Winkler, R. C. Schoemaker,
E. Chiossi, J. van Gerven, J. Dingemanse, Clin Pharmacol
Ther 2010, 87, 593-600; b) W. J. Herring, E. Snyder, K.
Budd, J. Hutzelmann, D. Snavely, K. Liu, C. Lines, T.
Roth, D. Michelson, Neurology 2012, 79, 2265-2274; c) R.
Rosenberg, P. Murphy, G. Zammit, D. Mayleben, D.
Kumar, S. Dhadda, G. Filippov, A. LoPresti, M. Moline,
JAMA Netw Open 2019, 2, e1918254.
[37]
[38]
L. Citrome, Int J Clin Pract 2014, 68, 1429-1441.
C. Boss, C. Roch-Brisbare, M. A. Steiner, A. Treiber, H.
Dietrich, F. Jenck, M. von Raumer, T. Sifferlen, C.
Brotschi, B. Heidmann, J. T. Williams, H. Aissaoui, R.
Siegrist, J. Gatfield, ChemMedChem 2014, 9, 2486-2496.
a) C. Brotschi, M. H. Bolli, J. Gatfield, B. Heidmann, F.
Jenck, C. Roch, T. Sifferlen, A. Treiber, J. T. Williams, C.
Boss, ChemMedChem 2020, 15, 430-448; b) C. Brotschi,
C. Roch, J. Gatfield, A. Treiber, J. T. Williams, T. Sifferlen,
B. Heidmann, F. Jenck, M. H. Bolli, C. Boss,
[39]
ChemMedChem 2019, 14, 1257-1270.
20
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000453
ChemMedChem
FULL PAPER
The quest for the best dual orexin receptor antagonist (daridorexant) for the treatment of insomnia disorders
Since its discovery in 1998, the orexin system is of interest to the research community as a potential therapeutic target for the treatment of
sleep/wake disorders. We describe our efforts leading to the identification of daridorexant which successfully finished two pivotal phase 3
clinical trials for the treatment of insomnia disorders.
21
This article is protected by copyright. All rights reserved.