NBI-1117568

NBI-1117568 is Neurocrine Biosciences' oral selective muscarinic M4 receptor orthosteric agonist, licensed from Nxera Pharma (formerly Sosei Heptares), now in Phase 2 for acute mania in bipolar I disorder (NCT07288320, n=150, Young Mania Rating Scale endpoint at Day 21) and parallel Phase 3 testing in schizophrenia across four trials [1][2][3]. Neurocrine describes it as the first and only M4-selective orthosteric agonist in clinical development, differentiating it from Bristol Myers Squibb's Cobenfy (xanomeline-trospium), which activates M1, M4, and other muscarinic subtypes and requires peripheral trospium to blunt M2/M3 side effects [4]. The compound is Neurocrine's bet that M4 selectivity can treat serious psychiatric illness without the dopamine receptor blockade that drives weight gain and motor side effects of standard antipsychotics, and without the broader muscarinic footprint that forced the trospium add-on in Cobenfy. The commercial context changed in September 2024, when Cobenfy became the first muscarinic antipsychotic approved by the FDA, validating the mechanism class and setting a competitive bar [4]. Nxera is eligible for up to approximately $1.5B in development milestones plus commercial milestones (total up to $2.6B) and tiered royalties on net sales under the 2021 license, and received a $35M milestone on the positive Phase 2 schizophrenia readout [11][12]; that royalty burden is a material drag on Neurocrine's net economics. NBI-1117568 must now show either differentiated efficacy, better tolerability, or an indication Cobenfy does not yet own. Bipolar mania is that kind of white-space indication. Readouts over the next 18 to 24 months decide whether Neurocrine builds a franchise or ships a follower.

Novel small molecule, never approved anywhere. In bipolar I mania, Phase 2 recruiting (NCT07288320, n=150) with YMRS at Day 21 as primary [1]. In schizophrenia, Neurocrine is running four Phase 3 trials in parallel: two acute inpatient studies (NCT07105098 and NCT06963034, each targeting roughly n=280, PANSS at Week 6) [2][3], a long-term safety study (NCT07114874, n=600) [5], and a relapse-prevention trial (NCT07227818, n=560) [6]. No breakthrough therapy, fast track, orphan, or priority review designations have been publicly disclosed. The decision to move forward in Phase 3 schizophrenia followed August 2024 Phase 2 topline (SCZ2028), a 4-arm dose-finding study (20 mg QD, 40 mg QD, 60 mg QD, 30 mg BID vs placebo) in which the 20 mg once-daily arm met the primary endpoint with a 7.5-point placebo-adjusted reduction in PANSS total at Week 6 (p=0.011, effect size 0.61) and an 18.2-point reduction from baseline; the higher-dose arms showed numerical improvement versus placebo but did not reach statistical significance [7]. First Phase 3 schizophrenia readouts are expected in 2026 to 2027. The bipolar I mania Phase 2 topline is likely late 2026 or 2027 given the 21-day treatment phase and typical Phase 2 timelines, though Neurocrine has not provided specific guidance in recent SEC filings [8].

The brain uses acetylcholine as a signaling molecule, and muscarinic receptors are the switches acetylcholine flips. M4 receptors sit on medium spiny neurons in the striatum, where their activation dampens dopamine release through autoreceptor-like feedback. In schizophrenia and acute mania, dopamine is thought to be hyperactive in subcortical regions, driving psychosis and manic symptoms. Classical antipsychotics block the D2 dopamine receptor directly. That works but causes motor side effects, weight gain, and elevated prolactin. The muscarinic approach is indirect: activating M4 turns down dopamine release upstream, so you get the antipsychotic effect without the postsynaptic D2 blockade. The mechanism class was de-risked clinically by Cobenfy (xanomeline-trospium), a pan-muscarinic agonist that hit PANSS in three Phase 3 schizophrenia trials and received FDA approval in September 2024 [4]. Cobenfy's cognitive benefit signal is attributable to M1 activation in cortex and hippocampus [9]; because NBI-1117568 is M4-selective, that same procognitive read-across should not be assumed and must be demonstrated empirically. The open question that NBI-1117568 is built to answer is whether an M4-selective orthosteric agonist, without the peripheral anticholinergic trospium add-on used to blunt M2/M3-driven GI side effects in Cobenfy, can match efficacy with cleaner native tolerability. Neurocrine and Nxera describe the compound as having very low affinity for M1, M2, M3, and M5; specific numerical fold-selectivity ratios have not been publicly disclosed in the Phase 2 publications. AbbVie's emraclidine, a selective M4 positive allosteric modulator (a different mode of action than orthosteric agonism), failed Phase 2 schizophrenia in November 2024, so M4-targeting is not a sure thing and the orthosteric-vs-allosteric distinction may matter [10].

The bipolar I mania study (NCT07288320) is Phase 2, placebo-controlled, 150 adults in a current manic episode, primary endpoint change from baseline in Young Mania Rating Scale (YMRS) at Day 21 [1]. YMRS is the standard mania scale, 11 items, higher scores worse. Three weeks is appropriate for acute mania given that quetiapine and olanzapine typically separate from placebo by week 2 to 3. No active comparator, which is defensible for a proof-of-concept Phase 2 but means efficacy will be judged against the placebo delta, not against existing drugs. The parallel schizophrenia Phase 3 program uses PANSS total score at Week 6 as primary across both acute inpatient trials (NCT07105098, NCT06963034), matching the Phase 2 endpoint timepoint and allowing cross-study effect-size comparison [2][3][7]. A central concern across the program is tolerability: muscarinic agonists have historically been held back by nausea, vomiting, sweating, hypersalivation, and bradycardia from peripheral M2/M3 activation. Neurocrine's thesis is that M4 selectivity avoids this liability natively, without the trospium crutch Cobenfy requires. The Phase 2 safety database supports that thesis directionally: the most common adverse events in SCZ2028 were somnolence, dizziness, and headache; GI and cardiovascular events were infrequent and treatment discontinuation rates due to adverse events were similar between NBI-1117568 and placebo [7]. Neurocrine has not published a direct head-to-head GI AE comparison versus Cobenfy, so the differentiation claim rests on cross-trial comparison, which is suggestive rather than definitive. Enrollment appears on pace across all five active registered studies as of late 2026 ClinicalTrials.gov entries.

Our model estimates a 4% chance this drug is eventually approved. It starts from the historical base rate for Phase 2 drugs in this area (about 24%), then adjusts using ten facts about the trial and sponsor. What moves the number most: it is held back by the sponsor's thin or weak approval record, weak or limited earlier-phase results, a randomized design, and a comparator/control arm. The other facts land near average for this stage, so they leave the estimate roughly where the base rate put it.

Efficacy risk is first-order. In Phase 2 schizophrenia, only the 20 mg once-daily arm reached statistical significance; the 40 mg, 60 mg, and 30 mg BID arms missed [7]. If the underlying dose-response is biphasic rather than simply monotonic within the therapeutic range, Phase 3 reproducibility at 20 mg is less certain than a clean monotonic dose-response would predict. If Neurocrine picked the wrong Phase 3 dose, the schizophrenia program fails and the bipolar readout gets reinterpreted retrospectively against a weakened mechanism thesis. Safety risk is the classic muscarinic liability: nausea, vomiting, sweating, hypersalivation, and bradycardia from residual peripheral M2/M3 activation. The entire premise of M4 selectivity is sidestepping these effects natively; if real-world Phase 3 tolerability resembles Cobenfy, the differentiation story collapses. Execution risk is low. Neurocrine has the CNS trial infrastructure and roughly $2.86B in Ingrezza-driven revenue to fund the program through readout per the most recent 10-K [8]. Economic risk from the Nxera license is material: up to approximately $1.5B in development milestones, commercial milestones summing to a $2.6B total, and tiered royalties on net sales compress Neurocrine's net margin on the asset relative to a wholly-owned franchise [11]. Commercial risk is the live question. Cobenfy launched late 2024 at approximately $22,500 per year [4]. A muscarinic challenger needs either superior efficacy, meaningfully better tolerability, or an indication Cobenfy does not yet own. Bipolar mania is exactly that white-space indication. If NBI-1117568 hits YMRS, Neurocrine reaches the mood-stabilizer market ahead of BMS. If it misses, the bipolar read-across for muscarinic antipsychotics gets damaged more broadly, hurting the schizophrenia program's option value.

Worth watching, carefully. The signal to key off is not the bipolar Phase 2 readout itself but the first schizophrenia Phase 3 acute inpatient topline (NCT07105098 or NCT06963034), expected 2026 to 2027 [2][3]. If those hit PANSS at the 20 mg Phase 3 dose, the muscarinic franchise reprices and the bipolar program becomes high-value optionality on a second indication. If they miss, the bipolar study is a sunk cost and the mechanism takes a second black eye after emraclidine [10]. Neurocrine's valuation already prices in some probability of a Cobenfy-adjacent franchise, so the asymmetry tilts toward downside surprise, and the Nxera royalty structure caps the upside even in a success scenario [11]. Check back when the first Phase 3 schizophrenia topline hits. The near-term tell is any 8-K from Neurocrine on enrollment pace or interim safety reads, both of which have historically preceded Phase 3 pivots in this therapeutic area. For bipolar mania specifically, watch the Day 21 YMRS point estimate and the placebo response rate: Phase 2 mania trials routinely fail on inflated placebo arms rather than inactive drug.