TAK-994, a novel orally available brain-penetrant orexin 2 receptor-selective agonist, suppresses fragmentation of wakefulness and cataplexy-like episodes in mouse models of narcolepsy

daytime


Introduction
Orexin, also called hypocretin, is a neuropeptide which regulates sleep/wakefulness (de Lecea et al., 1998;Sakurai et al., 1998;Tsujino and Sakurai, 2009). Orexin also has biological functions in energy homeostasis, stress control, and reward system (Tsujino and Sakurai, 2009). There are two types of orexins, orexin-A (OX-A) and orexin-B (OX-B); OX-A and OX-B originate from the proteolytic cleavage of a common precursor molecule, prepro-orexin, in orexin neurons of the lateral hypothalamus (de Lecea et al., 1998;Sakurai et al., 1998). Loss of orexin neurons is strongly linked to narcolepsy type 1 (NT1). The number of orexin neurons in individuals with NT1 was reduced by between 85% and 95% compared with normal subjects (Thannickal et al., 2000). Moreover, very low or undetectable levels of OX-A in the cerebrospinal fluid are observed in individuals with NT1 (Ripley et al., 2001).
Individuals with NT1 suffer from severe neurological conditions characterized by excessive daytime sleepiness (EDS), cataplexy, hypnagogic/hypnopompic hallucinations, sleep paralysis, and disrupted nocturnal sleep (Scammell, 2015;Abad and Guilleminault, 2017). However, no drugs that address the underlying disease pathophysiology of NT1 are available (Abad and Guilleminault, treatment of both EDS and cataplexy. However, despite these medications, individuals with NT1 report persistent daytime sleepiness and/or fatigue (Maski et al., 2017). Hence, a novel therapeutic approach could be useful for individuals with NT1.
Two postsynaptic G protein-coupled receptors (GPCRs) are specific for orexins, orexin 1 receptor (OX1R) and orexin 2 receptor (OX2R) (Sakurai et al., 1998). OX1R has one order of magnitude higher binding affinity for OX-A over OX-B, whereas OX2R has a similar affinity for both OX-A and OX-B (Sakurai et al., 1998). OX2R knockout (KO) mice show narcolepsy-like symptoms including fragmentation of sleep/wakefulness and cataplexy-like episodes, while OX1R KO mice do not show apparent behavioral abnormalities Sakurai, 2007).
Moreover, canines with null mutations in the OX2R gene (Hcrtr2) show marked narcolepsy-like symptoms (Lin et al., 1999). Thus, OX2R activation is anticipated to be a promising therapeutic strategy for NT1 (Irukayama-Tomobe et al., 2017). However, orexins cannot cross the blood-brain barrier (Fujiki et al., 2003). As such, non-peptide and brain-penetrant OX2R agonists are of interest as potential treatments for NT1.
This article has not been copyedited and formatted. The final version may differ from this version.

Ethics statement
The care and use of the animals and the experimental protocols in this study were approved by the Institutional Animal Care and Use Committee (IACUC) of Takeda Pharmaceutical Company Limited.

Animals
In our previous study, the wake-promoting effects of an OX2R agonist danavorexton assessed by using male mice could predict the clinical efficacy in individuals with NT1 (Evans et al., 2022).
Thus, male mice were used in this study. Male C57BL/6J mice obtained from CLEA Japan Inc.
(Tokyo, Japan) were used for slice electrophysiology study (3-to 4-week-old), electroencephalogram/electromyogram (EEG/EMG) study (13-to 14-week-old) and pharmacokinetic study (9-to 10-week-old). Orexin/ataxin-3 mice with a C57BL/6J genetic background were obtained from University of Tsukuba (Hara et al., 2001). Male orexin/ataxin-3 mice and their wild-type littermate mice (WT mice) were bred in our laboratory, and were used for EEG/EMG study and pharmacokinetic study at 25-to 31-week-old. Male OX2R KO mice with a C57BL/6J genetic background were generated in our laboratory (Yukitake et al., 2019), and were used for EEG/EMG study at 14-week-old. Orexin-tTA mice with a C57BL/6J genetic background This article has not been copyedited and formatted. The final version may differ from this version. administered orally to mice in a volume of 10 mL/kg of body weight.

Radioligand binding assay
The saturation binding and association/dissociation kinetics analysis of [ 3 H]TAK-994 to hOX2R expressing Expi293F cell membranes were conducted as described previously (Ishikawa et al., 2022). Data analysis was carried out using GraphPad Prism 6 (GraphPad software, Inc., CA, USA). To determine the receptor number (B max ) and the equilibrium dissociation constant (pK D ) in saturation binding studies, data were globally fitted with a non-linear regression of one site-total and non-specific binding. To determine the association rate constant (k on ), the dissociation rate constant (k off ), and pK D values, data were globally fitted to single-phase exponential association and decay equations. Dissociation t 1/2 was calculated as ln(2)/ k off .

β-arrestin recruitment assay
The binding of β-arrestin to hOX2R in hOX2R/CHO-EA cells was assessed using a PathHunter assay kit (DiscoveRx, CA, USA) as described previously (Yukitake et al., 2019). The PathHunter signals were measured using an EnVision plate reader (PerkinElmer Inc.). The responses to 0.3% DMSO in the absence and presence of 1 μM OX-A were used as 0% and 100% responses, respectively.

In vitro off-target profiling of TAK-994
Activity of TAK-994 (10 μM) on various receptors, ion channels and enzymes (106 targets in total) was evaluated at Eurofins Panlabs Discovery Services Taiwan, Ltd. (Taipei, Taiwan).

Slice electrophysiology study
Experiments were performed using slices of the posterior hypothalamus containing TMN, prepared from C57BL/6J mice as described previously (Yukitake et al., 2019). Histaminergic neurons in the TMN were identified by the presence of the transient outward current and the inwardly rectifying current activated by hyperpolarization (Kamondi and Reiner, 1991).
Pharmacological effects of test compounds on the membrane potential were assessed in the presence of 1 μM tetrodotoxin to block neuron firing. Membrane potentials during the last 30 s of each treatment condition were averaged, and the delta values between the membrane potential and the baseline were analyzed. Recordings with large access resistance (>30 MΩ) were abandoned. Signals were recorded using an Axopatch 700B amplifier and a Digidata 1322A digitizer board, bandpass This article has not been copyedited and formatted. The final version may differ from this version. were also calculated using SleepSign software.

Evaluation of sleep/wakefulness states in mice
This article has not been copyedited and formatted. The final version may differ from this version.

Evaluation of cataplexy-like episodes in mice
Murine cataplexy-like episodes were scored according to the following four criteria: (1)  Orexin-tTA;TetO DTA mice are known to have more severe cataplexy-like episodes compared to orexin/ataxin-3 mice even in the absence of chocolate (Tabuchi et al., 2014). Therefore, the effect of TAK-994 on spontaneous cataplexy-like episodes in orexin-tTA;TetO DTA mice was studied in the absence of chocolate. Vehicle or TAK-994 (1 or 3 mg/kg, p.o.) was administered to orexin-tTA;TetO DTA mice at ZT12, and then EEG, EMG and locomotor activity were recorded.
This article has not been copyedited and formatted. The final version may differ from this version.

Effects of repeated administration of TAK-994 on sleep/wakefulness states in mice
Orexin/ataxin-3 mice were grouped into two cohorts; a control group and a 14-day treatment group. In the control group, the vehicle (i.e., 0.5% MC in distilled water) was administered orally to mice three times a day (t.i.d.: ZT12, ZT15 and ZT18) for 14 days. In the 14-day treatment group, TAK-994 (30 mg/kg, p.o.) was administered to mice three times a day (ZT12, ZT15 and ZT18) for 14 days. EEG/EMG were recorded on day 1 and day 14 (active phase: ZT12-ZT21, subsequent sleep phase: ZT0-ZT10).

Pharmacokinetic study in mice
In C57BL/6J mice, blood samples were collected at various time points after administration of high-performance liquid chromatography-tandem mass spectrometry. The lower limit of quantitation was 3 ng/mL for plasma. The maximum concentration (C max ), the time taken to reach C max (T max ), area under the concentration-time curve, and mean residence time (MRT) were calculated.

Statistics
All data were presented as the mean ± SEM. The EC 50 values were calculated by XLfit (IDBS, MA, USA) or GraphPad Prism 6 from the data expressed as % of control. For in vivo studies, statistical analysis was performed using EXSUS (CAC EXICARE Corporation, Tokyo, Japan).
Pairwise differences between groups were evaluated using a two-tailed paired t-test. In experiments with multiple doses of test compounds, the statistical differences were analyzed using a two-tailed Williams test (for parametric data) or a two-tailed Shirley-Williams test (for nonparametric data).
Comparisons for time elapsed data were analyzed using repeated measures analysis of variance (ANOVA) followed by a post hoc two-tailed Student's t-test with Bonferroni correction. For all analyses, a P value of ≤ 0.05 was considered significant.
This article has not been copyedited and formatted. The final version may differ from this version.

TAK-994 selectively binds and activates recombinant and endogenous OX2R
Chemical structure of TAK-994 is shown in Fig. 1A to hOX2R in a monophasic manner, with pK D of 7.07 and B max of 4.03 pmol/mg protein (Fig. 1B).
Kinetic binding studies showed that [ 3 H]TAK-994 associated rapidly to hOX2R and dissociated rapidly from hOX2R within 5 min (Fig. 1C). The k on , k off and dissociation t 1/2 were 4.20 × 10 7 M -1 min -1 , 1.28 min -1 and 0.89 min, respectively (Supplemental Table S1). The dissociation constant (k off /k on ) calculated from the kinetics analysis (pK D = 7.47) was comparable with the value determined from the saturation binding assay (pK D = 7.07), suggesting good accuracy of these parameters.
Next, we characterized OX2R-selective agonistic activity of TAK-994 using hOX2R/CHO-K1 cells and hOX1R/CHO-K1 cells. TAK-994 increased calcium mobilization in hOX2R/CHO-K1 cells in a dose-dependent manner with EC 50 value of 19 nM (Fig. 1D). On the contrary, EC 50 value of TAK-994 in calcium mobilization using hOX1R/CHO-K1 cells was 14,000 nM (Fig. 1D). Therefore, allosteric modulation activity. The inhibitory or stimulatory activities of TAK-994 against various enzymes, receptors and ion channels (106 targets in total) were assessed. TAK-994 at 10 μM did not induce more than 50% inhibition or stimulation of any enzymes, receptors and ion channels, except progesterone receptor B (55% inhibition), indicating high selectivity of TAK-994 for OX2R with minimal off-target activity in vitro (Supplemental Table S3 and S4).
Histaminergic neurons in the TMN receive projections from orexin neurons in the hypothalamus (Marcus et al., 2001). OX2R, but not OX1R, is highly expressed in histaminergic neurons (Marcus et al., 2001;Mieda et al., 2011); therefore, activation of OX2R can be assessed by measuring the membrane potential of these cells. [Ala 11 , D-Leu 15 ]-OX-B, an OX-B peptide analog with 400-fold selectivity for OX2R over OX1R, was used as a control (Asahi et al., 2003). Oral administration of TAK-994 produces wake-promoting effects via OX2R activation at

3-fold lower dosage in NT1 model mice compared with WT mice during the sleep phase
We previously reported that s.c. administration of danavorexton at 1 and 3 mg/kg significantly promoted wakefulness in C57BL/6J mice during the sleep phase (Yukitake et al., 2019). To This article has not been copyedited and formatted. The final version may differ from this version.  Fig. S3, A and B). On the contrary, TAK-994 (30 mg/kg, p.o.) did not affect total wakefulness time in OX2R KO mice (Fig. 2C). Thus, TAK-994 has good oral bioavailability and brain penetration, and exerts arousal effects through activation of This article has not been copyedited and formatted. The final version may differ from this version. After a long period of orexin loss, NT1 model mice may cause changes in sensitivity to TAK-994 as was seen in danavorexton (Evans et al., 2022). To explore this possibility, we evaluated the wake-promoting effects of TAK-994 in orexin/ataxin-3 mice and their littermate WT mice.
Although its precise mechanism of action is unknown, dopaminergic and noradrenergic systems are prime targets for the action of modafinil (Gerrard and Malcolm, 2007). We compared the sensitivity of WT mice and orexin/ataxin-3 mice to wake-promoting effects of modafinil. Modafinil was This article has not been copyedited and formatted. The final version may differ from this version.  Fig. S6C). These results indicate that modafinil increased wakefulness in both WT mice and orexin/ataxin-3 mice at the same minimum effective dose (i.e., 30 mg/kg). Taken together, higher sensitivity in orexin/ataxin-3 mice compared with WT mice may be specific to OX2R agonists. OX2R system may be sensitized to OX2R ligand after the long-term loss of orexin in orexin/ataxin-3 mice.

TAK-994 ameliorates narcolepsy-like symptoms in orexin/ataxin-3 mice during the active phase
Orexin/ataxin-3 mice show narcolepsy-like symptoms such as fragmentation of sleep/wakefulness and cataplexy-like episodes (Hara et al., 2001). We examined the effects of TAK-994 on these symptoms in orexin/ataxin-3 mice after administration at the onset of the active period (ZT12) (Fig. 3A).  Fig. S7, A and B). The wake-promoting effects of TAK-994 (10 mg/kg, p.o.) were maintained for about 2 h after administration (Supplemental Fig. S7C). TAK-994 also significantly ameliorated fragmentation of wakefulness with fewer number of wakefulness episodes (Fig. 3D) and longer mean duration of wakefulness episodes (Fig. 3E). EEG power spectral analysis is widely used to investigate the drug properties associated with functional modulation of brain in clinical and preclinical studies (Lin et al., 2008;Conrado et al., 2013). Evaluation of the effect of TAK-994 on EEG power density during wakefulness in orexin/ataxin-3 mice revealed that TAK-994 (10 mg/kg, p.o.) significantly increased gamma power during wakefulness (Fig. 3F).
Next, we examined the effect of TAK-994 on cataplexy-like episodes in orexin/ataxin-3 mice during the active phase. We performed this study in the presence of chocolate as a rewarding stimulus to increase the frequency of cataplexy (Fig. 3G) (Clark et al., 2009;Oishi et al., 2013). In the presence of chocolate, cataplexy-like episodes were approximately 1 episode per hour, therefore, we decided to evaluate the number of cataplexy-like episodes during 3 h after TAK-994 administration (at ZT12). TAK-994 (3 mg/kg, p.o.) significantly suppressed cataplexy-like episodes in orexin/ataxin-3 mice (Fig. 3H). These results suggest that TAK-994 ameliorates multiple narcolepsy-like symptoms such as fragmentation of wakefulness and cataplexy-like episodes in orexin/ataxin-3 mice during the active phase.

TAK-994 maintains wake-promoting effects following chronic dosing in orexin/ataxin-3 mice during the active phase
Long-term treatment with GPCR agonists often lead to receptor desensitization and subsequent loss of efficacy (Rajagopal and Shenoy, 2018 intervals) was administered to orexin/ataxin-3 mice during the active phase for 14 days (Fig. 5A).
The plasma concentrations of TAK-994 after repeated administration and acute administration were comparable (Supplemental Fig. S10). Under these conditions, EEG/EMG were recorded on day 1 and day 14 (Fig. 5A). On day 1, TAK-994 significantly increased total wakefulness time during 9 h after the first administration (Fig. 5B), accompanied by decreases in total NREM sleep time and

Discussion
In early proof-of-concept studies, danavorexton (i.v. infusion) was well tolerated, and increased wakefulness in individuals with NT1 compared to placebo (Evans et al., 2022). However, danavorexton has limited oral availability, thus we explored orally available and brain-penetrant OX2R agonists. Profile of each GPCR activators can be different, like partial agonists, biased agonists, and allosteric modulators, that can affect the pharmacological profiles of each drug (Clark et al., 1999;Violin et al., 2014;Sako et al., 2019). Therefore, comprehensive preclinical characterization was conducted to discover a novel, orally available, and brain-penetrant OX2R agonist with comparable efficacy to that of danavorexton in individuals with NT1 and healthy volunteers.
We first conducted detailed in vitro characterization of TAK-994. Kinetic binding study revealed that TAK-994 showed a rapid association and dissociation kinetics to OX2R. The fast association and dissociation kinetic properties of TAK-994 suggest its low risks of residual activity after elimination from the plasma. In other words, its pharmacodynamic effect could be mainly driven by pharmacokinetic factors. In calcium mobilization assays, TAK-994 showed potent agonistic activity for OX2R (EC 50 value of 19 nM), more than 700-fold selectivity against OX1R, signal transduction pattern similar to that by orexin peptides (Yukitake et al., 2019). In other words, OX2R activation by TAK-994 can mimic physiological OX2R activation by orexin peptides.
TAK-994 (p.o.) produced wake-promoting effects in C57BL/6J mice, but not in OX2R KO mice, during the sleep phase, demonstrating that TAK-994 is an orally available brain-penetrant OX2R-selective agonist. Beside oral availability, overall in vivo pharmacological characteristics of TAK-994 are similar to that of danavorexton (Yukitake et al., 2019).
The effect of TAK-994 on narcolepsy-like symptoms was assessed using two narcolepsy mouse models: orexin/ataxin-3 mice and orexin-tTA;TetO DTA mice. In orexin/ataxin-3 mice, a slow, progressive loss of orexin neurons has been reported, with losses of 56%, 75%, 90% and 99% by 2, 4, 8 and 12 weeks, respectively, and almost complete loss by 15 weeks of age (Hara et al., 2001). In contrast, loss of orexin neurons is rapid in orexin-tTA;TetO DTA mice: 80% loss after only 1 week of DOX removal, and 95% loss by 2 weeks (Tabuchi et al., 2014). Although underlying mechanisms of action is not fully understood, orexin-tTA;TetO DTA mice are known to have more severe cataplexy-like episodes compared to orexin/ataxin-3 mice (Tabuchi et al., 2014). In both models,

TAK-994 ameliorated fragmentation of wakefulness and suppressed cataplexy-like episodes during
This article has not been copyedited and formatted. The final version may differ from this version. the active phase. This is in line with our previous observation using danavorexton: danavorexton suppressed fragmentation of wakefulness and cataplexy-like episodes during the active phase in orexin/ataxin-3 mice (Evans et al., 2022;Ishikawa et al., 2022). Moreover, similar to danavorexton (Evans et al., 2022;Ishikawa et al., 2022), wake-promoting effects of TAK-994 were observed following chronic dosing for up to 14 days in orexin/ataxin-3 mice. Overall, preclinical efficacy of TAK-994 in NT1 model mice was comparable to that of danavorexton. We also conducted EEG power spectral analysis during wakefulness in orexin/ataxin-3 mice and orexin-tTA;TetO DTA mice to investigate the drug properties on EEG and brain functions. TAK-994 significantly increased gamma power during wakefulness in both NT1 model mice. It has been reported that high-frequency oscillations including gamma activity are associated with attention and memory (Jensen et al., 2007).
Evaluation of efficacy of TAK-994 on attentional and cognitive behavior warrants further exploration.
TAK-994, but not modafinil, induced arousal effect at a 3-fold lower dosage in NT1 model mice compared with WT mice during the sleep phase. Similar sensitization was reported by other GPCRs such as dopamine receptor (Metz et al., 2005).
In summary, we found that TAK-994 is an orally available, and brain-penetrant OX2R-selective agonist with similar overall preclinical properties to that of danavorexton, except oral availability (Yukitake et al., 2019;Evans et al., 2022;Ishikawa et al., 2022). Preclinical characteristics of This article has not been copyedited and formatted. The final version may differ from this version.