SLEEPJ, 2019, 1–14
`
`doi: 10.1093/sleep/zsz076
`Advance Access Publication Date: 29 March 2019
`Original Article
`
`Original Article
`Preclinical in vivo characterization of lemborexant (E2006), a
`novel dual orexin receptor antagonist for sleep/wake regulation
`Carsten Theodor Beuckmann1,*, Takashi Ueno2, , Makoto Nakagawa1, Michiyuki Suzuki3
`and Shigeru Akasofu1
`
`1Neurology Business Group, Discovery, Eisai Co., Ltd., Tsukuba, Japan, 2Drug Metabolism and Pharmacokinetics, Eisai Co., Ltd., Tsukuba, Japan
`and 3Pharmaceutical Regulatory Affairs Department, Marketing Authorization Group, EA Pharma Co., Ltd., Tokyo, Japan
`
`*Corresponding author. Carsten T. Beuckmann, Neurology Business Group, Discovery, Tsukuba Research Laboratories, Eisai Co., Ltd., 5-1-3 Tokodai, Tsukuba, Ibaraki
`300–2635, Japan. Email: c-beuckmann@hhc.eisai.co.jp.
`
`Abstract
`Study Objectives: To present results from in vivo studies underlying the preclinical development of lemborexant (E2006), a novel dual orexin
`(hypocretin) receptor antagonist for sleep/wake regulation.
`
`Methods: Rodent (wild-type rats and wild-type and orexin neuron-deficient [orexin/ataxin-3 Tg/+] mice) studies were performed to evaluate
`the effects of single-dose oral lemborexant (1–300 mg/kg) on orexin-induced increases in plasma adrenocorticotropic hormone (ACTH),
`locomotor activity, vigilance state measures (wakefulness, nonrapid eye movement [non-REM] sleep, rapid eye movement [REM] sleep),
`ethanol-induced anesthesia, and motor coordination, and the effects of multiple-dose oral lemborexant (30 mg/kg) on vigilance state
`measures. Active comparators were almorexant and zolpidem. Pharmacokinetics were assessed after single-dose lemborexant in mice and
`rats.
`
`Results: Lemborexant prevented the orexin-promoted increase in ACTH in rats, therefore demonstrating inhibition of the orexin signaling
`pathway. Furthermore, lemborexant promoted sleep in wild-type mice and rats. Lemborexant promoted REM and non-REM sleep at an equal
`rate (there was no change in the REM sleep ratio). In contrast, zolpidem reduced REM sleep. The sleep-promoting effect of lemborexant was
`mediated via the orexin-peptide signaling pathway as demonstrated by a lack of sleep promotion in orexin neuron-deficient mice. Chronic
`dosing was not associated with a change in effect size or sleep architecture immediately postdosing. Lemborexant did not increase the
`sedative effects of ethanol or impair motor coordination, showing good safety margin in animals. Pharmacokinetic/pharmacodynamic data
`for mice and rats were well aligned.
`
`Conclusions: These findings supported further clinical evaluation (ongoing at this time) of lemborexant as a potential candidate for treating
`insomnia and other sleep disorders.
`
`Statement of Significance
`Traditional pharmacologic treatments for insomnia, such as benzodiazepines, non-benzodiazepine hypnotics, and sedating antidepressants,
`have varying effectiveness across differing clinical insomnia phenotypes and various safety concerns, which has led to investigation of
`potential treatments with alternative mechanisms of action. The orexin (hypocretin) signaling system is of interest because orexins play
`an important role in sleep/wake regulation by binding to orexin-1 and -2 receptors. Here, we summarize results from in vivo rodent studies
`underlying the preclinical evaluation of lemborexant, a novel dual orexin receptor antagonist for treating insomnia/other sleep disorders. We
`observed that lemborexant effectively promoted sleep without potentiating the sedative effects of ethanol or impairing motor coordination.
`These preclinical findings supported further clinical evaluation of lemborexant for treating insomnia/other sleep disorders.
`
`Key words: antagonist; dual orexin receptor antagonist; E2006; in vivo; insomnia; lemborexant; mouse; orexin; rat; sleep
`
`Submitted: 6 September, 2018; Revised: 4 February, 2019
`© Sleep Research Society 2019. Published by Oxford University Press on behalf of the Sleep Research Society.
`This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial
`License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and
`reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com
`
`1
`
`Page 1 of 14
`
`EISAI EXHIBIT 1033
`
`

`

`2
`
`| SLEEPJ, 2019, Vol. 42, No. 6
`
`Introduction
`
`Insomnia is a prevalent sleep disorder that is associated with
`significant health and economic burdens [1, 2]; pharmaceutical
`treatment plays a key role in relieving these burdens. Currently,
`common treatments for insomnia include benzodiazepines,
`non-benzodiazepine hypnotics, and sedating antidepressants.
`Although these treatments are effective, various safety concerns
`[3] have highlighted the need for alternative treatment options
`that promote sleep via a different mechanism(s) of action.
`One alternative mechanism of action is antagonism of
`the orexin (hypocretin)-mediated pathway. The hypothalamic
`neuropeptides, orexin-A (OXA) and orexin-B (OXB), have been
`demonstrated to activate postsynaptic G-protein-coupled
`orexin-1 and orexin-2 receptors (OX1R and OX2R) located in
`the central nervous system [4, 5]. The importance of orexins
`in sleep/wake regulation has been highlighted in animal
`studies, which have shown that deficits in orexin signaling
`are associated with a phenotype similar to the human sleep
`disorder narcolepsy [6–11] and that central infusion of OXA
`promotes wakefulness
`[12–14]. Further, human patients
`with narcolepsy type-I have been shown to have deficits in
`orexin signaling [15], which are due to degeneration of orexin
`neurons [16, 17]. These neurons, which comprise a relatively
`small population in the lateral hypothalamic area [18], project
`widely throughout the central nervous system, including to
`noradrenergic neurons in the locus coeruleus, serotonergic
`neurons in the dorsal and median Raphe nuclei, cholinergic
`neurons
`in the pedunculopontine/laterodorsal tegmental
`nuclei and basal forebrain, dopaminergic neurons in the
`ventral tegmental area, and histaminergic neurons in the
`tuberomammillary nucleus [6, 19–22]. All of these respective
`monoaminergic neurons express OX1R and/or OX2R [14, 23],
`and are recognized to be involved in sleep/wake control [24,
`25]. Support for the latter comes from several elegant mouse
`optogenetic excitation studies, where activation of orexin
`neurons reduced the latency to wakefulness from both
`nonrapid eye movement (non-REM) and rapid eye movement
`(REM) sleep [26]. In contrast, hyperpolarization/inactivation of
`orexin neurons increased non-REM sleep [27, 28].
`The activity of orexin neurons has been shown to be
`mediated by sleep- and wake-promoting neurons. Specifically,
`sleep-promoting GABAergic neurons from the ventrolateral
`preoptic area have been demonstrated to provide inhibitory
`input [29–31], whereas wake-promoting cholinergic neurons
`from the basal forebrain have been demonstrated to provide
`excitatory input [29, 32, 33]. These findings support the
`concept of orexin being a major controller/stabilizer of the
`“switch” between the mutually regulating states of sleep
`and wakefulness, with orexin stabilizing the switch in the
`“wake position,” in other words, stabilizing wakefulness [34].
`Consistent with this proposed role in stabilizing wakefulness, in
`vivo single-unit recordings have shown that orexin neurons fire
`most rapidly during active wakefulness, slow down during quiet
`wakefulness, and are mostly inactive during REM and non-REM
`sleep [35–38]. Furthermore, recent fiber photometry studies in
`mice indicate that orexin neuron activity is causally linked to
`wakefulness [39] and directly influences the activity of wake-
`controlling neurons in the paraventricular thalamus [40]. Finally,
`OXA concentrations in the cerebrospinal fluid (CSF) of rats [41],
`monkeys [42], and humans [43] were found to exhibit circadian
`variation, increasing during the active phase and decreasing
`
`during the rest phase. This suggests that orexins are factors of
`vigilance control, which are in turn under circadian control.
`In recent years, insomnia disorder has been viewed as less
`of an issue related to sleep and more of an issue related to
`excessive activity of wakefulness-related circuits [44]. Indeed,
`functional neuroimaging has shown that night time glucose
`metabolism is elevated in the brains of patients with insomnia,
`suggesting an inability of arousal mechanisms to decrease at the
`appropriate time [45]. Inhibiting the orexin-signaling pathway
`to dampen one of the major arousal-promoting circuits would
`therefore appear to be a reasonable pharmacological approach
`for treating insomnia. To this end, animal studies demonstrated
`that dual orexin receptor antagonists (DORA) and selective OX2R
`antagonists (selective orexin-2 receptor antagonist [2-SORA])
`promoted sleep [39, 46–50], whereas selective OX1R antagonists
`(selective orexin-1 receptor antagonist [1-SORA]) had no obvious
`effect on sleep [39, 47]. This efficacy pattern is consistent with
`findings in orexin pathway knockout mice, which have shown
`that OX2R plays a pivotal role in the maintenance of wakefulness
`and suppression of non-REM and REM sleep [8, 14]. Conversely,
`OX1R is thought to play a role in REM sleep suppression,
`because, although OX1R knockout mice lack an overt sleep
`phenotype [18], OXA reduces REM sleep in OX2R knockout mice,
`but not in orexin receptor double knockout mice [14]. Further,
`OX2R knockout mice have milder REM sleep-related narcolepsy
`symptoms than preproorexin knockout mice, which are devoid of
`any functional input on both orexin receptors [8]. Unsurprisingly,
`several DORAs and 2-SORAs have been developed and evaluated
`as potential treatments for insomnia in humans [51]. To date,
`one DORA, suvorexant, has been approved for the treatment of
`insomnia [52].
`Lemborexant (also known as E2006) is a DORA currently
`being evaluated in phase 3 clinical trials for the treatment of
`insomnia disorder and in phase 2 trials for irregular sleep/
`wake rhythm disorder in patients with Alzheimer’s dementia.
`Previous lemborexant publications have detailed the discovery
`[53], the in vitro and in silico characterization [54], and results
`from a phase 2 study in patients with insomnia [55]. Here, we
`describe the results from key preclinical studies underlying the
`in vivo characterization of lemborexant.
`
`Materials and methods
`Animals
`
`Animal care and experimental procedures were performed in
`an animal facility accredited by the Health Science Center for
`Accreditation of Laboratory Animal Care and Use of the Japan
`Health Sciences Foundation. All protocols were approved by the
`Institutional Animal Care and Use Committee and carried out
`in accordance with the Animal Experimentation Regulations of
`Eisai Co., Ltd.
`Male C57BL/6NCrlCrlj mice (hereafter referred to as wild-
`type mice) and male F344/DuCrlCrlj (F344) and Sprague Dawley
`rats were supplied by Charles River Laboratories (Yokohama,
`Japan). Orexin neuron-deficient mice (orexin/ataxin-3 Tg/+;
`C57BL/6N background) [7] were supplied by Prof. Takeshi Sakurai
`(University of Tsukuba, Japan) and propagated by breeding orexin/
`ataxin-3 Tg/+ males with C57BL/6NCrlCrlj females (Charles River
`Laboratories, Yokohama, Japan). All rodents were maintained
`under a 12-h light-dark cycle with food and water available ad
`
`Page 2 of 14
`
`

`

`libitum. All experiments were conducted during the light phase
`(Zeitgeber time 0:00 = lights on; Zeitgeber time 12:00 = lights off).
`Experiments conducted only with wild-type mice were
`performed with breeder-supplied animals, while in experiments
`comparing orexin/ataxin-3 Tg/+ with wild-type mice, the
`corresponding wild-type (orexin/ataxin-3 +/+) littermates were
`used.
`
`Chemical compounds
`
`Lemborexant, almorexant, and zolpidem were synthesized
`in-house and suspended as free bases in the vehicle solution
`specified for each study. Doses for each compound in the mouse
`experiments were set based on the minimum necessary dose
`for sleep promotion in mice, and are therefore different for each
`compound.
`
`Effect of single-dose lemborexant on the orexin-
`induced increase in ACTH in rats
`
`Previous studies have demonstrated that OXA and OXB increase
`plasma corticosterone concentrations in rats [56], with the
`efficacy of OXB suggesting that OX2R is mediating this effect.
`Mediators located downstream of orexin signaling are thought
`to be corticotropin releasing factor [56] and neuropeptide Y
`[57]. As the release of corticosterone is triggered by ACTH, we
`developed an in vivo functional assay, in which plasma ACTH
`concentrations are increased by central application of OXB. In this
`assay (details ensue), we applied [Ala11, D-Leu15]-orexin B, which
`is far more selective for OX2R (>400-fold vs. OX1R) than natural
`OXB (>10-fold vs. OX1R) [58], to minimize potential cross-talk
`from OX1R. We carried out the study using rats because we have
`observed that vehicle-treated mice have intrinsically high ACTH
`levels that are not distinguishable from ACTH levels in [Ala11,
`D-Leu15]-orexin B-treated mice (data not shown). This is likely
`due to mice being more difficult to habituate to experimental
`conditions than rats. The ability of lemborexant to inhibit the
`[Ala11, D-Leu15]-orexin B-induced increase in plasma ACTH
`concentrations was determined. The optimum dose of [Ala11,
`D-Leu15]-orexin B was found to be 1 nmol/head (data not shown),
`which was associated with an approximate fourfold increase in
`plasma ACTH concentrations relative to vehicle treatment.
`Male F344 rats (age: 5 weeks; body weight: 85.8–103.5 g) were
`implanted with infusion cannulae into the left lateral ventricle
`for intracerebroventricular (i.c.v.) injection. Four to five days after
`surgery, rats were habituated for oral administration (p.o.) and
`handling once before the study. Six to seven days after cannula
`implantation, rats received p.o. vehicle (5% [v/v] dimethyl
`sulfoxide, 9.5% [v/v] cremophor in saline; n = 10) or lemborexant
`(5 mL/kg suspended in vehicle) 1, 3, 10, or 30 mg/kg (n = 5, 6, 6,
`and 5, respectively). One hour later, vehicle control rats received
`5 µL of phosphate buffered saline (PBS) or [Ala11, D-Leu15]-orexin
`B (1  nmol/head, 0.2  mmol/L in PBS, Tocris Bioscience, Japan)
`via i.c.v. injection (n = 5 each). All lemborexant pretreated rats
`received [Ala11, D-Leu15]-orexin B via i.c.v. injection. Fifteen
`minutes later, rats were decapitated and blood samples were
`collected with Na2EDTA (100  mg/mL, 100  µL). Blood samples
`were then centrifuged (1000 × g, 10 min at 4°C) and supernatant
`plasma was stored at −80°C for later measurement of ACTH and
`lemborexant concentrations. After decapitation, blue ink was
`injected i.c.v. to confirm the correct placement of cannulae.
`
` Beuckmann et al.
`
`| 3
`
`Coronal cross-sections were made near the cannulae; placement
`was judged (by an experienced observer) to be correct if blue ink
`was visible in the lateral ventricle(s). Data from 3 out of 32 rats
`with incorrect cannula placement were excluded from analysis.
`Plasma ACTH concentrations were measured using an ACTH
`radioimmunoassay kit (ACTH IRMA “MITSUBISHI,” Mitsubishi
`Chemical Medience, Tokyo, Japan). Measurement of radioactivity
`and calculation of ACTH concentrations were conducted using
`a scintillation counter (ARC-1000M, Hitachi Aloka Medical, Ltd.,
`Tokyo, Japan).
`Plasma lemborexant concentrations were measured by liquid
`chromatography-tandem mass spectrometry (LC-MS/MS). Plasma
`samples were precipitated with four volumes of acetonitrile
`containing an
`internal standard
`(10  ng/mL
`imipramine).
`Following vortex mixing and centrifugation, the supernatant was
`filtered, and the resultant filtrate was injected into the LC-MS/
`MS. Lemborexant and the internal standard were identified based
`on their respective retention times and the mass units of the
`monitoring ions on the mass chromatograms. The calibration
`curve was obtained at a concentration range from 1 to 1,000 ng/
`mL by least-squares linear regression, with a weighting factor
`of 1/X2 on the ratio of the peak area of lemborexant to that of
`the internal standard against the nominal concentrations in the
`calibration standards. The limit of quantitation for lemborexant
`was 1  ng/mL. The unbound fraction of lemborexant (0.167) in
`F344 rat plasma was determined by equilibrium dialysis and the
`unbound plasma concentration was calculated.
`Unbound plasma concentrations of
`lemborexant and
`corresponding inhibition of OXB-triggered increases in plasma
`ACTH were determined for each rat. Values were analyzed using
`Michaelis–Menten kinetic analysis, using the least-square fit
`method and fixing maximum inhibition to 100%.
`
`Effect of single-dose lemborexant on spontaneous
`locomotor activity in wild-type and orexin neuron-
`deficient mice
`
`Wild-type male mice (age: 9 weeks; body weight: 19.4–22.5  g)
`were dosed p.o. with vehicle (5% [v/v] dimethyl sulfoxide,
`10% [v/v] cremophor in 10  mmol/L HCl; 10  mL/kg; n  =  16) or
`lemborexant (30 [n = 8] or 100 mg/kg [n = 7]) at Zeitgeber time 3:40
`or 5:30. One hour after dosing, mice were placed in an open field
`arena (VersaMax, AccuScan Instruments, Columbus, OH) and
`locomotor activity was automatically recorded as infrared light
`beam breaks as previously described [49]. For activity values, all
`horizontal and vertical infrared light beam break counts were
`summed over 1 h after the start of locomotor activity recording.
`In a separate study, orexin neuron-deficient mice (age: 18–26
`weeks; body weight: 27.9–36.0 g) were dosed p.o. with vehicle (as
`above; n  =  8) or 100  mg/kg lemborexant (n  =  8), the maximum
`dose tested in wild-type mice, at Zeitgeber time 3:40 or 5:30.
`Thirty minutes later, locomotor activity was recorded and
`summed over 1 h as described above.
`
`Effect of single-dose lemborexant on vigilance
`state measures in wild-type and orexin neuron-
`deficient mice
`
`Under deep ketamine/xylazine anesthesia, wild-type male mice
`(age: 10–11 weeks; body weight: 21.7–26.1 g) were implanted with
`transcranial supradural electroencephalography (EEG) and nuchal
`
`Page 3 of 14
`
`

`

`4
`
`| SLEEPJ, 2019, Vol. 42, No. 6
`
`electromyography (EMG) electrodes as previously described [6, 49].
`Briefly, four holes were drilled into the skull (relative to bregma:
`1.1 mm rostral and 1.45 mm lateral on both sides; 3.5 mm caudal
`and 1.45  mm on both sides) and four gold-plated contacts of a
`six-contact board mount socket (#929975-01-36, 3M, Tokyo, Japan)
`were inserted to the dura mater and fixed with dental cement. The
`remaining two contacts were connected to teflon-coated stainless
`steel wires, gold-plated at the tips (EMG leads), which were placed
`intranuchally into pockets formed by blunt dissection of neck
`muscle left and right of midline. EEG signals were read from one
`side only (usually right) between rostral and caudal electrodes.
`EEG/EMG signals were collected via a connector attached to
`the board mount socket, which was connected to an amplifier
`via a rotating swivel, allowing the mouse to move freely while
`connected to the cable during the experiment. Mice were allowed
`to recover and to habituate to the recording room for 1 week while
`being housed in recording cages, and were then divided into
`body-weight matched groups and habituated to p.o. dosing with
`vehicle (10 mL/kg 0.5% [w/v] methylcellulose 400, WAKO, Osaka,
`Japan) at Zeitgeber time 3:30–4:10 for 2 consecutive days. The
`next day, mice were dosed p.o. with vehicle (n = 5), lemborexant
`(1 [n = 5] or 10 mg/kg [n = 4]), almorexant (10 [n = 5] or 100 mg/
`kg [n = 4]), or zolpidem (3 [n = 5] or 30 mg/kg [n = 4]) at Zeitgeber
`time 3:30–4:10. Note: almorexant, a DORA, and zolpidem, a widely
`used sleep drug, were used as active comparators for paradigm
`validation. EEG and EMG signals were continuously recorded for
`7  h after dosing (sampling frequency 128 Hz), divided into 10-s
`epochs, and analyzed as previously described [49] according to
`standard rodent sleep criteria [59] using SleepSign software (v3,
`Kissei Comtec, Matsumoto, Japan). An experienced observer,
`blinded to treatment, subsequently provided visual confirmation
`of vigilance state judgement. Data were accumulated into 1-h
`bins for time course graphs and accumulated for the first 3 h after
`dosing for statistical analysis.
`In a separate study, orexin neuron-deficient mice and
`wild-type control littermates (age: 12–13 weeks; body weight:
`23.4–33.8  g) were implanted with EEG and EMG electrodes,
`as described above for wild-type mice. After approximately
`2 months of recovery and housing in the recording room, mice
`were habituated to EEG/EMG-recording cages for 3 days before
`p.o. dosing with vehicle (10  mL/kg 0.5% [w/v] methylcellulose
`400, WAKO) at Zeitgeber time 3:30–4:00 for three consecutive
`days under recording conditions. Preliminary EEG/EMG signal
`recordings and sleep time analyses were performed for p.o.
`dosing days 2 and 3. Mice with sufficient EEG/EMG signal quality
`(n = 7 per treatment group and genotype) were then randomly
`divided into treatment groups with minimal differences in
`average total sleep time (within 3 h after dosing). The next day,
`mice were dosed p.o. with vehicle or lemborexant (30 mg/kg) at
`Zeitgeber time 3:30–4:00. EEG/EMG signal recording (sampling
`frequency 128 Hz) was started shortly before dosing and was
`continued for 3 h and 45 min after dosing. Data were analyzed
`as described for the study in the previous paragraph.
`included
`Vigilance states assessed
`in both studies
`wakefulness, non-REM sleep, and REM sleep. Representative
`traces for each of these vigilance states are shown in
`Supplementary Figure 1.
`
`Effect of single-dose lemborexant on vigilance state
`measures in rats
`
`Under deep sodium pentobarbital anesthesia, male Sprague
`Dawley rats (age: 7–10 weeks; body weight: 311–449  g) were
`
`implanted with battery-driven, wireless
`intraperitoneally
`telemetry devices (TL11M2-F40-EET, Data Sciences International,
`St Paul, MN) for EEG/EMG measurements. Two silver screws were
`fixed 2.0 mm left and right of lambda through the skull bone so
`as to touch the dura mater. EEG leads were knotted to the screws,
`while EMG leads were placed intranuchally into pockets formed
`by blunt dissection of neck muscle left and right of midline.
`After 9–11 days of recovery in home cages, rats were habituated
`to the recording room and p.o. dosing procedure for 2  days,
`while still being housed in the same home cages throughout the
`experiment. A total of 12 rats were then selected for the study
`and assigned to body-weight matched groups for five dosing and
`recording sessions. The sessions were conducted with 2–3 days
`of intermittent wash-out periods, with p.o. dosing taking place at
`Zeitgeber time 2:00–3:00 and subsequent recording of EEG/EMG
`signals for 4  h using a telemetry system (recording software:
`Dataquest A.R.T. Platinum v4.10; Data Sciences International,
`New Brighton, MN) for
`later off-line analysis
`(software:
`NeuroScore v1.1; Data Sciences International). Continuous
`recordings were divided into 10-s epochs and automatically
`analyzed via an algorithm that determined vigilance states.
`Automated analysis results were then verified and, if necessary,
`corrected by a trained observer blinded to treatment. No animal
`received the same test compound at the same dose twice. There
`were a total of 10 treatment groups in the study; vehicle (10 mL/
`kg 0.5% [w/v] methyl-cellulose 400 [n = 6]), lemborexant (3, 10, 30,
`100, or 300 mg/kg [all n = 6]), and zolpidem (3, 10, 30, or 100 mg/
`kg [all n = 6]).
`Vigilance states assessed included cumulative wakefulness,
`non-REM sleep, and REM sleep times for 2  h after dosing.
`Representative traces for each of these vigilance states are
`shown in Supplementary Figure 2.
`
`Effect of chronic-dose lemborexant on vigilance
`state measures in rats
`
`Male Sprague Dawley rats, fully habituated to experimental
`conditions from the previously described single-dose study,
`were used after a 5-day washout period. All rats were dosed
`p.o., once-daily at Zeitgeber time 2:00–3:00, with vehicle (0.5%
`[w/v] methylcellulose 400)  on day 1–3, then vehicle (n  =  2),
`lemborexant 30  mg/kg (n  =  5), or zolpidem 100  mg/kg (n  =  5)
`from day 4 to 24, and finally vehicle on days 25 and 26. EEG/EMG
`signals were recorded (as already described for the single-dose
`study) on days 1 and 2 (pretreatment), days 4, 7, 11, 14, 18, 21, and
`24 (treatment), and days 25 and 26 (posttreatment) for nearly 3
`hours. Lemborexant and zolpidem doses were chosen based on
`the maximum effects in the previous single-dose experiment.
`Vigilance states assessed included wakefulness, non-REM
`sleep, and REM sleep for 2 h after dosing, using the same analysis
`procedure as described for the single-dose study. Sleep latency,
`defined as the time between dosing and the first occurrence of
`1 min of uninterrupted sleep, was also assessed.
`
`Effect of single-dose lemborexant on ethanol-
`induced anesthesia in wild-type mice
`
`Wild-type male mice (age: 13 weeks; body weight: 21.8–28.1  g)
`were dosed p.o. with vehicle (10 mL/kg 0.5% [w/v] methylcellulose
`400), lemborexant (1, 3, or 10  mg/kg), almorexant (30, 100, or
`300  mg/kg), or zolpidem (3, 10, or 30  mg/kg) (all groups, n  =  6)
`during the light phase. After 5 min, mice received intraperitoneal
`injections of 3.0  g/kg ethanol (20% [w/v] in saline). The
`
`Page 4 of 14
`
`

`

`almorexant and zolpidem doses were based on those used in
`a previous rat study, where almorexant up to 300 mg/kg did not
`show interaction with ethanol, but zolpidem from 10  mg/kg
`upwards showed interaction with ethanol [60]. Note that the
`dose of ethanol was chosen because it caused comparably short
`anesthesia with little data variation (Supplementary Figure 3).
`The time from ethanol injection to regaining of righting reflex
`was measured and taken as anesthesia duration, with 240 min
`as the cutoff time.
`In a second ethanol study, wild-type male mice (age: 13
`weeks; body weight: 23.2–27.7  g) were dosed p.o. with vehicle
`(10  mL/kg 0.5% [w/v] methylcellulose 400)  or lemborexant (30,
`100, or 300 mg/kg) (all groups, n = 6) during the light phase. The
`time from ethanol injection to regaining of righting reflex was
`recorded as already described.
`
`Effect of single-dose lemborexant on motor
`coordination in wild-type mice
`
`Motor coordination was assessed using a rotarod treadmill
`MK-660C (Muromachi Kikai, Tokyo, Japan). With the device on
`hold, five mice were placed on the central axle, facing the same
`direction away from the experimenter, without interacting with
`or seeing each other. The axle was then accelerated to 40 rpm
`within 180 s, at which point the test was terminated. The time
`from the start of rotation until a mouse fell was automatically
`recorded and taken as the latency to fall, with a cutoff at 180 s.
`Before the first test of the day, a prerun with mice not used for
`testing was carried out to scent the device. Between each test,
`the axle was wiped to remove urine and feces.
`For this study, male wild-type mice (age: 14 weeks; body
`weight: 22.1–29.1  g), previously trained on the treadmill for 3
`consecutive days with subsequent 14 days rest, were allocated
`to body-weight equivalent groups to receive single p.o. dosing of
`vehicle (10 mL/kg 0.5% [w/v] methylcellulose 400), lemborexant
`(30, 100, or 300  mg/kg), or zolpidem (100  mg/kg) (all groups,
`n = 11). The starting lemborexant dose (30 mg/kg) was selected
`as this is approximately threefold the sleep-promoting dose,
`while zolpidem 100  mg/kg has previously been reported to
`impair motor coordination in rats [60]. On the day before the
`study, mice were given 0.25  mL p.o. water and placed on the
`treadmill for a training run. Mice that fell before 120  s were
`placed back on the device. On the day of the study, mice were
`first tested on the treadmill at Zeitgeber time 2:00 (predosing).
`Treatments were then administered at Zeitgeber time 4:00 and
`motor coordination was assessed at Zeitgeber times 4:30, 6:00,
`7:30, and 9:00. All five dosing groups were allocated through all
`five test compartments on the treadmill. Between tests, mice
`were returned to their home cages and allowed to rest, with free
`access to food and water.
`
`Plasma and CSF concentrations of lemborexant after
`single dosing in rats
`
`Male Sprague Dawley rats (age: 9 weeks; body weight: 337–
`355  g) were dosed p.o. with lemborexant 30  mg/kg in 0.5%
`(w/v) methylcellulose 400 (5 mL/kg) at Zeitgeber time 4:30–5:00.
`Two hours later, rats were anesthetized and plasma and CSF
`samples were obtained from the abdominal aorta and cisterna
`magna, respectively, for the measurement of lemborexant
`concentrations by LC-MS/MS.
`
` Beuckmann et al.
`
`| 5
`
`Plasma concentrations of lemborexant after single
`dosing in mice
`
`Male C57BL/6N mice (age: 14 weeks; body weight: 26.7–31.7  g)
`were dosed p.o. with lemborexant 10 or 300 mg/kg in 0.5% [w/v]
`methylcellulose 400 (10  mL/kg) at Zeitgeber time 3:00–9:15. At
`0.25, 0.5, 1, 3, 5, 6, 18, and 24 h after dosing, mice were anesthetized
`and plasma samples were obtained from the abdominal aorta
`for measurement of lemborexant concentrations by LC-MS/MS.
`
`Statistical analyses
`
`Data are presented as the mean ± standard error of the mean
`(SEM) and were generally compared by t test (for comparisons
`involving two groups) or one-way analysis of variance (ANOVA)
`followed by Dunnett multiple comparison test (for comparisons
`involving three or more groups). Exceptions to one-way ANOVA
`were: the single-dose vigilance study in rats (mixed-effects
`model with treatment group and period modeled as fixed effects
`and animal as a random effect); the chronic-dose vigilance study
`in rats (repeated measures ANOVA); and the motor coordination
`study in mice (repeated measures analysis of covariance with
`predosing values as a covariate). For all studies, p <0.05 (two-
`sided) was considered to indicate statistical significance.
`Statistical analyses were performed using the SAS software
`package version 8.2 (SAS Institute Japan. Tokyo, Japan).
`
`Results
`Effect of single-dose lemborexant on orexin-induced
`increases in ACTH in rats
`
`Rats centrally treated with [Ala11, D-Leu15]-orexin B had
`significantly higher plasma ACTH concentrations than rats
`treated with vehicle (Figure 1). Lemborexant exhibited dose-
`related inhibition of this increase in ACTH, with the level of
`inhibition being significant from 3 mg/kg p.o. and higher. ACTH
`concentrations that were equivalent to those in rats treated
`with vehicle occurred with lemborexant doses from 10  mg/kg
`p.o. and higher.
`The unbound plasma lemborexant concentration achieving
`50% inhibition (IC50) was (mean ± SEM) 5.9 ± 1.7 nmol/L (Figure 2).
`
`Effect of single-dose lemborexant on spontaneous
`locomotor activity in wild-type and orexin neuron-
`deficient mice
`
`reduced
`significantly
`lemborexant
`In wild-type mice,
`spontaneous locomotor activity compared with vehicle at both
`p.o. doses tested (30 and 100 mg/kg) (Figure 3A). Conversely, in
`orexin neuron-deficient mice, which exhibited less activity than
`vehicle-treated wild-type mice, lemborexant (100  mg/kg p.o.)
`did not significantly reduce spontaneous locomotor activity
`compared with vehicle (Figure 3B).
`
`Effect of single-dose of lemborexant on vigilance
`state measures in wild-type and orexin neuron-
`deficient mice
`
`In wild-type mice, lemborexant (10  mg/kg p.o.) significantly
`reduced wakefulness and increased non-REM sleep compared
`
`Page 5 of 14
`
`

`

`6
`
`| SLEEPJ, 2019, Vol. 42, No. 6
`
`Figure 1. Inhibitory effect of single-dose lemborexant on the plasma ACTH
`increase induced by centrally applied [Ala11, D-Leu15]-orexin B in rats. Rats
`received single-dose oral doses of vehicle or lemborexant (1, 3, 10, or 30 mg/kg)
`(all n = 5, except for 30 mg/kg n = 4) during the light phase. #p < 0.05 versus i.c.v.
`PBS control (t-test); *p  <  0.05 versus vehicle/[Ala11, D-Leu15]-orexin B (one-way
`analysis of variance followed by Dunnett multiple comparison test).
`
`Figure 2. Relationship between unbound plasma concentrations of lemborexant
`and corresponding inhibition of [Ala11, D-Leu15]-orexin B-triggered increases in
`plasma ACTH concentrations in rats. Data for individual animals are depicted
`as white open symbols; cohort means are depicted as gray open symbols (mean
`± s