Seltorexant

Seltorexant (JNJ-42847922) is Janssen's selective orexin-2 receptor antagonist (2-SORA) in Phase 3 for major depressive disorder (MDD) with insomnia symptoms, dosed at 20 mg as adjunctive therapy on top of standard antidepressants [1][2]. The bet is that blocking the brain's primary wake-promoting receptor improves sleep and lifts mood in a population where insomnia and depression amplify each other. The completed Phase 3 NCT04533529 already hit, with seltorexant 20 mg producing a 2.6 point MADRS separation versus placebo at Day 43 (statistically significant, all primary and secondary endpoints met) [3][8]. The companion head-to-head Phase 3 NCT04513912 (Pinter et al. 2026) showed seltorexant numerically beating quetiapine XR on response rate at 26 weeks (57.4% vs 53.4%, not statistically significant) with comparable MADRS improvement (-23.0 vs -22.7) and substantially better tolerability, including 1.6 kg less weight gain [4][6]. Two confirmatories are recruiting: NCT06559306 (adjunctive, n=752, estimated primary completion December 2026) and NCT07573176 (monotherapy, n=600) [1][2]. A clean monotherapy readout would establish the first orexin-based mechanism in psychiatry. Zuranolone (GABA-A PAM, 2023) and auvelity (NMDA antagonist combo, 2022) are newer MDD mechanisms approved after esketamine (2019), but neither is an adjunctive-only orexin product, so seltorexant would still open a new category.

Novel compound, never approved anywhere. Phase 3, with one positive Phase 3 already in hand. No publicly disclosed FDA designations such as breakthrough therapy, fast track, or priority review. The development arc started with Phase 2b NCT03227224 (Savitz et al. 2021, n=287, completed) [5][9]. The 20 mg dose showed a least-squares mean MADRS difference vs placebo of -4.5 at week 3 (p=0.003) and -3.1 at week 6 (p=0.083); in the prespecified insomnia subgroup (ISI >= 15), week-6 separation widened to -4.9 (p significant). That signal selected 20 mg for Phase 3 and validated the MDD-with-insomnia enrichment strategy. Two confirmatory Phase 3s followed. NCT04533529 (n=588, completed) tested adjunctive seltorexant 20 mg vs placebo on background antidepressants; J&J reported in May 2024 that the study met all primary and secondary endpoints, with a 2.6 point MADRS separation vs placebo at Day 43 and similar adverse event rates between arms [3][8]. NCT04513912 (n=757, completed) compared seltorexant 20 mg to quetiapine XR as adjunctive therapy over 26 weeks. Pinter et al. (Int J Neuropsychopharmacol 2026) reported response at 26 weeks of 57.4% (seltorexant) vs 53.4% (quetiapine XR), not statistically significant; MADRS change -23.0 vs -22.7; mean body weight change +0.5 kg vs +2.1 kg; fewer dropouts on seltorexant [4][6]. Janssen is now running two further Phase 3 confirmatories. NCT06559306 (n=752, recruiting) tests adjunctive seltorexant in MDD with insomnia symptoms, primary endpoint MADRS at Day 43, estimated primary completion December 2026 [1]. NCT07573176 (n=600, recruiting) tests monotherapy in MDD with moderate-to-severe insomnia symptoms, same primary endpoint [2]. NCT07573176 is verified as registered on ClinicalTrials.gov and listed as recruiting under Janssen Research & Development sponsorship. A specific FDA submission timeline has not been disclosed publicly, but completion of the December 2026 readout would support a 2027 NDA filing.

Orexin (also called hypocretin) is a neuropeptide your brain releases to keep you awake. It binds two G-protein coupled receptors, OX1 and OX2, with OX2 doing most of the wakefulness work. Dual orexin receptor antagonists (DORAs) like suvorexant, lemborexant, and daridorexant block both receptors and are approved for insomnia. Seltorexant blocks only OX2 (the 2-SORA strategy, selective OX2 receptor antagonism), on the theory that selective blockade preserves more REM sleep architecture and reduces next-morning grogginess relative to DORAs. The depression hypothesis layers on top: people with MDD often show dysregulated orexin signaling that drives insomnia, hyperarousal, and possibly mood symptoms directly. Quiet OX2 at night, the patient sleeps better, and an antidepressant signal emerges that goes beyond what improved sleep alone would predict. The genetic validation is thin. Open Targets gives HCRTR2 a 0.37 evidence score for MDD (modest) and 0.60 for insomnia (well-established, but not what Phase 3 primary endpoints measure). The strongest precedent the program now leans on is Janssen's own Phase 2 signal in MDD-with-insomnia (NCT03227224) [5][9] and the positive Phase 3 NCT04533529 [3][8], not external mechanistic validation. The mechanism is genuinely new in psychiatry. The historical failure mode for sleep-targeting antidepressants has been a clean sleep effect that doesn't carry mood with it; the NCT04533529 readout is the first credible evidence that this particular mechanism does carry mood.

NCT06559306 is the lead adjunctive Phase 3, dosing seltorexant 20 mg on top of background antidepressants in adult and elderly MDD patients with insomnia symptoms, n=752, recruiting [1]. Primary endpoint is change from baseline in MADRS (Montgomery-Asberg Depression Rating Scale, a 0-60 clinician-rated depression scale where 0 is no symptoms and 60 is most severe) at Day 43. Phase 3 adjunctive MDD drugs typically need a 2 to 3 point separation from placebo at this timepoint to support approval, which sounds tiny but is the FDA-accepted regulatory threshold for this endpoint; it represents roughly 4 to 5 percent of the scale and is on the order of the placebo-corrected effect sizes seen for approved adjunctive antipsychotics. NCT04533529 cleared this bar at 2.6 points [3][8]. The companion Phase 3 NCT07573176 tests seltorexant as monotherapy in MDD with moderate-to-severe insomnia symptoms (MDDIS), n=600, same primary endpoint [2]. Selecting for MDD-with-insomnia is a smart enrichment strategy: it focuses on the population where a dual-acting mechanism should produce the largest signal, lowers expected placebo response, and creates a label that differentiates from generic SSRIs and antipsychotic adjuncts. The risk on the new confirmatories is replication. The completed Phase 3 NCT04533529 separated at 2.6 points, comfortably above the typical bar but not by a wide margin; a replication trial that lands at 1.5 to 2.0 points would create regulatory friction. The monotherapy bet is more aggressive than adjunctive; Janssen would not run NCT07573176 without seeing meaningful monotherapy signals in earlier work and confidence in OX2 antagonism as a standalone antidepressant mechanism. Day 43 is also a short readout window for an antidepressant. Either the effect is fast, driven by sleep normalization, or it is unlikely to emerge at all.

Our model puts this drug's chances of eventual approval at 20%. That starts from the historical approval rate for Phase 3 drugs in this area, which is about 51%, then adjusts based on ten facts about the trial and sponsor. The main factors pulling the estimate down are the sponsor's weak approval record, limited earlier-phase results, and a randomized trial design; the main factor pushing it up is an above-average number of secondary endpoints. The remaining factors fell close to average, so they did not move the number much from where the base rate left it.

Efficacy risk is moderate, not large, given NCT04533529 has already cleared the 2.6 point bar [3][8]. MDD Phase 3 trials run placebo MADRS improvements of 30 to 40 percent of baseline, and replication failure of a borderline 2 to 3 point separation in NCT06559306 is the most likely path to a stumble. Janssen mitigated this by enriching for MDD-with-insomnia. Safety risk is on-mechanism: approved orexin antagonists carry FDA labeling for next-day somnolence, complex sleep behaviors (e.g., sleep-driving, sleep-eating, performed without conscious awareness), sleep paralysis, and hypnagogic hallucinations (dream-like visions at the edge of falling asleep) [7]. Suvorexant is Schedule IV. None of these are deal-breakers in insomnia, where the population is older and risk-tolerant for a sleep aid, but layering an OX2 antagonist onto psychiatric polypharmacy (taking multiple psychiatric drugs simultaneously, common in inadequately responsive MDD) raises the tolerability bar. Execution risk is moderate. Janssen has been at this since 2017 (NCT03227224 start) [5], and Idorsia's daridorexant (Quviviq) has been chipping at the insomnia market while the depression hypothesis takes its time to confirm. Commercial risk is the underrated one. Adjunctive MDD is dense with options: generic SSRIs and SNRIs, atypical antipsychotics (quetiapine XR, aripiprazole, brexpiprazole), lithium, esketamine (Janssen's own), zuranolone, auvelity, and TMS. The US addressable population is large (roughly 21 million US adults with MDD per year, and 60 to 70 percent have clinically significant insomnia, with adjunctive use indicated in the 30 to 50 percent who do not adequately respond to first-line treatment, giving a 4 to 7 million patient ballpark). Payers will compare seltorexant's MADRS delta against quetiapine XR's, and Pinter et al. 2026 made that comparison: numerically similar efficacy, materially better tolerability [4][6]. A pricing premium requires Janssen to convert that tolerability advantage into formulary positioning, and the bar is set high by generic competition.

Worth watching. The signal to track is a clean MADRS separation in NCT06559306 at Day 43 (estimated primary completion December 2026), ideally accompanied by an analysis showing the mood effect exceeds what total-sleep-time improvement alone would predict [1]. That is the test that distinguishes a new MDD mechanism from a sedative dressed up as an antidepressant. The monotherapy Phase 3 NCT07573176 is the higher-stakes bet [2]: if both confirmatories land, this becomes a new adjunctive category and likely a new monotherapy entry, the first orexin-based mechanism in psychiatry. If only the adjunctive Phase 3 hits, the franchise is smaller and the payer pitch is harder. Expect a topline readout on NCT06559306 in Q1 2027 (December 2026 primary completion plus standard data lock and unblinding), and a likely 2027 NDA filing if the data clears. Watch for J&J data previews at APA 2027 and ECNP 2027. Janssen's neuroscience portfolio has been thin since esketamine. Seltorexant is their next test of whether they can build a second adjunctive MDD franchise, and given that internal incentive, expect aggressive disclosure timing if the data clears the bar.