Seltorexant

Investigational selective OX2-only orexin antagonist from Johnson & Johnson / Minerva Neurosciences, currently in Phase 3 for MDD-with-insomnia and Phase 2 for primary insomnia + Alzheimer's-related sleep. Pivotal MDD3001 trial (May 2024) hit all primary + secondary endpoints — significant MADRS improvement + sleep gains in SSRI/SNRI-refractory MDD with insomnia. NOT YET FDA-APPROVED; not commercially available in any market as of 2026-05-05. Mechanistically the cleanest next-gen orexin drug for cognitive preservation (OX2-selective spares OX1 arousal/reward circuits, half-life only 2-3 hours), but WATCH-LIST for Dylan: he has neither MDD nor insomnia — he has a delayed chronotype that orexin antagonism does not phase-shift. Daridorexant is the currently-available DORA option. Re-evaluate at FDA approval (most likely 2026-2027) and at first real-world user-rating cohort.

Seltorexant is a selective OX2 receptor antagonist (2-SORA) with >100× binding affinity for OX2R over OX1R (Wikipedia; Drug Hunter). Discovered at Janssen and partnered with Minerva Neurosciences for development. This is the mechanistic dividing line vs all three approved DORAs (suvorexant, lemborexant, daridorexant), which block both OX1 and OX2 non-selectively.

Why OX2-selective vs dual matters

The orexin system has two receptor subtypes with substantially different distributions and functions (covered in detail in the daridorexant file — summary here):

• OX1R (HCRTR1): preferential for orexin-A; concentrated in locus coeruleus (norepinephrine), VTA (dopamine reward), amygdala. Drives arousal, vigilance, sympathetic tone, reward-seeking, fear/stress responses. Knockout mice show altered reward and motivation phenotypes but only mild sleep changes.
• OX2R (HCRTR2): binds both orexin-A and orexin-B; concentrated in the tuberomammillary nucleus (histamine — the brain's primary wake driver) and posterior hypothalamus. Drives wakefulness via histamine cascade. OX2R knockout produces a clear sleep phenotype. OX2R is the receptor most directly responsible for translating orexin neuron firing into "stay awake."

Theoretical advantages of OX2-selective blockade:

1. Sleep promotion is preserved or enhanced because OX2 is the dominant wake-driving receptor. Block it, sleep happens.
2. OX1-driven arousal/reward/sympathetic circuits stay intact. This is where the cognitive-preservation argument lives: chronic DORAs theoretically blunt OX1 functions (motivation, reward, sympathetic tone, stress responses), while a 2-SORA leaves those untouched. For a brain-priority user, OX1-sparing is the main reason this class is mechanistically interesting.
3. Lower theoretical narcolepsy/cataplexy risk. Animal data indicate that complete loss of orexin signaling is required to produce the full narcolepsy phenotype (PNAS 2010). OX2 blockade alone causes sleepiness without cataplexy. DORAs have produced rare cataplexy-like leg weakness in trials (see daridorexant file); 2-SORAs theoretically should not.
4. Possible separate antidepressant signal beyond the sleep effect. The Phase 2/3 MDD trials showed MADRS improvement that could not be fully explained by sleep gains alone — there is an emerging hypothesis that OX2 blockade modulates affective circuitry (amygdala, BNST, reward) in ways relevant to depression. Not fully characterized; this is the most active controversy in the orexin literature.

The "100× selectivity" caveat

100× selectivity sounds clean but is not absolute. At very high doses, OX1R blockade still occurs. Janssen's clinical doses (10-40 mg) appear to stay within OX2-selective territory based on PK modeling, but at supra-therapeutic exposures the selective edge erodes.

Pharmacokinetic profile (the half-life argument)

• Tmax: 0.3-1.5 hours — fast absorption, sleep-onset friendly.
• Half-life: ~2-3 hours — substantially shorter than all three DORAs (suvorexant 12 hr, daridorexant 8 hr, lemborexant 17-19 hr). (Wikipedia/Clinical PK reviews; Trial.MedPath)
• Metabolism: primarily hepatic CYP3A4, with minor CYP2C19 contribution.
• Protein binding: >99% (HSA + α1-acid glycoprotein).
• No residual effect at 4 hours post-daytime dose in pharmacology studies.

The 2-3 hour half-life is the second engineering claim of the molecule: even cleaner morning clearance than daridorexant. Theoretically, this means dose at bedtime, gone before morning, no residual fog. The trade-off is that maintenance-of-sleep effect across a full 8-hour night is less dose-flexible than with longer-half-life DORAs — late-night WASO might re-emerge.

Janssen development context

• Discovered at Janssen Research & Development; first reported ~2014.
• Co-developed with Minerva Neurosciences (license deal 2014, expanded 2017-2020).
• Phase 1: completed.
• Phase 2 insomnia (NCT02464046, primary insomnia, 10/20/40 mg vs zolpidem vs placebo): 40 mg arm dropped at interim analysis for the depression study; 20 mg became the lead dose.
• Phase 2b adjunctive MDD (10/20/40 mg + SSRI/SNRI): 20 mg statistically significant on MADRS at week 6 in patients with insomnia symptoms (IDS-C 8+ score) (Oxford IJNP 2021).
• Phase 3 MDD3001 readout May 2024: pivotal study, achieved all primary and secondary endpoints, statistically significant + clinically meaningful MADRS improvement and sleep disturbance improvement at day 43 in patients with prior inadequate SSRI/SNRI response (J&J press release).
• Phase 3 vs quetiapine XR (adjunctive): 26-week head-to-head, seltorexant 57.4% response vs quetiapine 53.4% (numerically favorable, NOT statistically significant) — primary endpoint missed (academic.oup.com IJNP 2025; J&J press release).
• Phase 2 Alzheimer's-related sleep disturbance: ongoing.
• NDA submission: not yet filed as of 2026-05-05; originally projected for 2023, slipped to "later." Most credible analyst projection is 2026 NDA → 2027 PDUFA if J&J files cleanly off MDD3001.