NBI-1117568

Neurocrine Biosciences is testing NBI-1117568, an oral muscarinic M1/M4 receptor agonist, in a Phase 2 trial for acute mania in bipolar I disorder, with a 150-patient study reading out at Day 21 on the Young Mania Rating Scale [1]. The same drug is the centerpiece of a four-trial Phase 3 program in schizophrenia covering acute inpatient treatment, long-term safety, and relapse prevention [2][3][4][5]. The mechanism was validated commercially in September 2024 when Bristol Myers Squibb's Cobenfy (xanomeline plus trospium) became the first non-dopaminergic antipsychotic approved by FDA in decades, after BMS paid roughly $14 billion for Karuna Therapeutics to acquire it [8]. Neurocrine licensed NBI-1117568 from Sosei Heptares (now Nxera Pharma) under a November 2021 collaboration covering up to $2.6 billion in milestones plus royalties [11]. The bipolar mania trial is a faster, smaller bet on the same biology that powers the larger schizophrenia program.

NBI-1117568 is a novel small molecule, never approved for any indication. The bipolar I mania study (NCT07288320) is Phase 2, placebo-controlled, enrolling 150 adults hospitalized with an active manic episode, with a 21-day primary endpoint on the Young Mania Rating Scale (YMRS), an 11-item clinician-rated scale of manic symptoms [1]. The trial is recruiting and no public readout date has been guided. No FDA breakthrough therapy or fast track designation has been disclosed for the bipolar indication. The schizophrenia program is further along. Neurocrine reported topline Phase 2 schizophrenia results in August 2024: the 20 mg once-daily arm (the lowest dose tested) produced a statistically significant 7.5-point separation from placebo on the Positive and Negative Syndrome Scale (PANSS) at Week 6 with an effect size of 0.61, while the 40 mg QD, 60 mg QD, and 30 mg BID arms each missed significance [6]. This is a non-monotonic, inverted-U-shaped dose response, not a clean dose-ranging result. Tolerability was better than feared: discontinuation rates were similar to placebo, with somnolence (10.7% vs 2.9%) and dizziness (9.3% vs 1.4%) the only treatment-emergent events meaningfully more common than placebo [6]. Neurocrine moved straight into a four-trial Phase 3 program: two acute inpatient trials of identical design (NCT07105098 and NCT06963034, each n=284, 5-week PANSS readout), a relapse-prevention trial (NCT07227818, n=560), and a long-term open-label safety trial (NCT07114874, n=600) [2][3][4][5]. Acute Phase 3 readouts are likely in 2027 based on enrollment timelines. The bipolar Phase 2 is a smaller, faster option inside that broader development program.

Muscarinic receptors are switches for acetylcholine, one of the brain's main signaling molecules. There are five subtypes (M1 through M5). M1 sits mostly on cortical neurons that handle attention and memory; M4 is concentrated in the striatum, a deep brain region that controls movement and reward. Activating either one can dampen dopamine release indirectly through cholinergic circuits, without ever touching dopamine receptors themselves [7]. Classical antipsychotics work by blocking dopamine D2 receptors directly, which controls hallucinations and mania but causes movement disorders, weight gain, sedation, and elevated prolactin. The muscarinic approach takes a different route: tune down dopamine upstream through cholinergic circuits and skip the D2 blockade entirely. Why this matters for bipolar mania specifically: the dominant neurobiological model of acute mania is hyperdopaminergic activity in striatal and limbic circuits, the same anatomical regions where M4 receptors are densely expressed [7]. Lithium, valproate, and second-generation antipsychotics all converge on dampening this striatal dopamine signal, just from different angles. M4 agonism reaches the same circuit through cholinergic inhibition of dopamine release, so the mechanistic rationale for treating mania is direct, not a stretch by analogy to schizophrenia. The biology shares a target circuit; the clinical syndromes differ. The original muscarinic agonist (xanomeline) activated muscarinic receptors throughout the gut, causing nausea, vomiting, and sweating severe enough to derail the program in the 1990s. Karuna Therapeutics solved that by pairing xanomeline with trospium, an anticholinergic that stays outside the brain and blocks the peripheral side effects. The combination (Cobenfy) was approved by FDA in September 2024 for schizophrenia, giving the first commercial proof that muscarinic activation treats psychosis in humans [8]. NBI-1117568 attempts the same biology with a single molecule designed for better receptor selectivity, betting that selectivity removes the need for a chaperone anticholinergic.

NCT07288320 is a Phase 2, placebo-controlled study in 150 adults hospitalized with bipolar I disorder during an active manic episode. The primary endpoint is change from baseline in YMRS total score at Day 21 [1]. YMRS is the standard manic symptom scale used in every bipolar mania registration trial since the 1990s, so the endpoint is regulatorily clean and reads out fast. Acute mania trials are typically inpatient, which keeps adherence and assessment quality high. The 21-day duration is short for a Phase 2 efficacy study but matches the natural history of acute mania, which is treated for weeks rather than months. The 150-patient size is adequate to detect a clinically meaningful YMRS separation if the drug works at the dose chosen. Neurocrine has not publicly disclosed which dose arms are being tested in the bipolar trial, which is notable given that the Phase 2 schizophrenia readout showed only the lowest dose (20 mg QD) separated from placebo while three higher doses missed [6]. Dose selection is therefore the central scientific question coming into this trial. The trial is sponsored entirely by Neurocrine with no co-development partner [1]. Acute mania studies have a long history of high placebo response rates: meta-analyses of registration trials report mean placebo YMRS improvements of roughly 20 to 30 percent, driven by hospitalization, structured care, and regression to the mean from peak-episode enrollment [10]. The bigger the placebo response, the bigger the drug effect needs to be to separate.

The model gives this drug a 4% chance of eventually being approved. That figure starts from a 24% historical approval rate for Phase 2 drugs in this area, then adjusts using ten facts about the trial and sponsor. The estimate drops mainly because the sponsor has a thin or weak approval record, earlier-phase results were weak or limited, and the trial uses a randomized design with a comparator arm. The remaining facts were close to average for this stage, so they did not shift the number much.

The largest efficacy risk is dose-response fragility. Neurocrine's Phase 2 schizophrenia program tested four dose regimens and only the lowest (20 mg once daily) hit on PANSS, with 40 mg QD, 60 mg QD, and 30 mg BID all missing [6]. This inverted-U pattern is harder to interpret than a clean dose-response. It could reflect a true narrow therapeutic window, statistical noise in a small Phase 2, or both. It also leaves little headroom: if the 20 mg signal does not replicate, there is no obvious higher dose to fall back on. The bipolar mania trial inherits this question directly. Safety risk is lower than the historical xanomeline class would suggest. The Phase 2 schizophrenia readout showed discontinuation rates similar to placebo, with somnolence (10.7%) and dizziness (9.3%) the only adverse events meaningfully more common than placebo and no signal of the severe nausea, vomiting, and sweating that derailed first-generation xanomeline [6]. That said, longer-term tolerability in chronic dosing has not been reported, and detailed Phase 1 safety data have not been publicly disclosed in peer-reviewed form. Commercial risk is real even with a clean efficacy signal. Acute mania has dozens of generic options that work, and a novel branded agent at specialty pricing needs differentiation that justifies the cost. The natural argument is avoidance of the metabolic and movement side effects that come with second-generation antipsychotics, but that argument competes directly with Cobenfy, which by mid-2026 had reached roughly $43 million in quarterly sales (Q3 2025) and a growing prescription base that will only widen its head start in the muscarinic class [8][12]. Licensing economics also weigh on the asset: Nxera Pharma is entitled to tiered royalties and up to $2.6 billion in total milestones on the muscarinic portfolio Neurocrine licensed in 2021 [11], which compresses Neurocrine's margin on any approved indication.

Worth watching, with a high bar. Neurocrine's commercial business (anchored by Ingrezza for tardive dyskinesia) generated $2.51 billion in 2025 Ingrezza sales out of $2.81 billion total product revenue, with Crenessity adding $301 million in its launch year [9]. That cash flow funds the rest of the pipeline and means the company can absorb a Phase 2 miss without existential damage. The schizophrenia Phase 3 program is the main event; the bipolar mania trial is a fast option on the same molecule that could derisk it. Commercial frame for bipolar: roughly 3 to 4 million U.S. adults have bipolar I disorder, the population eligible for acute mania treatment in any given year is a fraction of that (most patients have one to two manic episodes per year requiring active treatment), and a novel branded agent capturing 5 to 10 percent of acute episodes would translate to roughly $300 to $700 million in peak U.S. sales, contingent on tolerability differentiation from generic standard of care. The specific signal to look for is a YMRS separation versus placebo of 4 points or more by Day 21, with a discontinuation rate from cholinergic side effects under 10 percent. That combination would establish NBI-1117568 as a credible single-molecule alternative to the xanomeline-trospium combination and would meaningfully derisk the larger schizophrenia bet. Anything less than a 3-point YMRS separation, or a discontinuation rate above 15 percent, makes the bipolar program hard to take into Phase 3 and adds pressure on the schizophrenia readouts in 2027. Items not in the public record that would sharpen this assessment: the bipolar dose selection, detailed Phase 1 safety data, and composition-of-matter patent expiry for NBI-1117568 (novel small molecules first disclosed in the mid-2010s typically run to the mid-2030s, but Neurocrine has not published a specific exclusivity window). Check back when Neurocrine guides a specific bipolar readout window on a quarterly earnings call, and when the Phase 3 schizophrenia acute trials enroll past 75 percent. The competitive comparator to track is Cobenfy's commercial launch trajectory [12].