BH-200 (NELIVAPTAN), A VASOPRESSIN V1B RECEPTOR ANTAGONIST, IN GENETICALLY DEFINED SUBGROUPS OF MDD: PRECISION TARGETING OF HPA-AXIS DYSREGULATION IN THE OLIVE PHASE 2 PROGRAM

Disturbances of hypothalamic–pituitary–adrenal (HPA) axis function are implicated in a biologically defined subset of patients with major depressive disorder (MDD). Pharmacologic modulation of the stress hormone system, including CRHR1 and vasopressin V1b receptor (V1bR) antagonism, has been explored for decades; however, trials conducted in unselected MDD populations have not led to regulatory approval. We hypothesized that biological heterogeneity dilutes treatment effects and that stress-axis–driven subgroups can be prospectively identified. In a prior positive Phase 2 trial with the selective V1bR antagonist BH-200 (nelivaptan), a bimodal distribution of response suggested distinct responder populations. Because the dexamethasone–CRH (Dex-CRH) test is not practical for routine use, a genetic classifier (V1b polygenic score; V1bPGS) was developed based on vasopressin signaling biology and Dex-CRH–characterized cohorts. The 14-SNP tool assigns patients to three subgroups (A, B, C), reflecting differential ACTH suppression profiles and putative vasopressin pathway activity. The OLIVE trial was a randomized, double-blind, placebo-controlled Phase 2 trial conducted in eight European countries. A total of 338 patients with MDD were randomized 2:1 to BH-200 250 mg BID or placebo for eight weeks. The prespecified primary endpoint was change from baseline to Week 8 on the 17-item Hamilton Depression Rating Scale (HAMD-17) in patients classified as Subgroup C. In the modified intention-to-treat population (N=331), BH-200 demonstrated clinically meaningful improvement versus placebo (Δ = −2.98, p = 0.0003). Treatment separation emerged early and was sustained through a 4-week follow-up. Response rates (≥50% reduction in HAMD17) were 50.0% with BH-200 versus 25.7% with placebo, odds ratio (OR), 2.98 (95% CI, 1.74, 5.10), and remission rates (HAMD-17 ≤7) were 25.8% versus 16.8%, OR 1.62 (95% CI, 0.86, 3.08) respectively. In the prespecified primary analysis population (Subgroup C; n=123), the treatment difference was Δ = −1.94 (p = 0.153; Cohen’s d 0.28). However, genetic stratification revealed marked heterogeneity of effect. The largest benefit was observed in Subgroup A (n=89), with Δ = −4.47 (p = 0.005; Cohen’s d 0.55), response rate 58.9% versus 25.9%, OR 4.22 (95% CI, 1.51, 11.83), and remission rate 33.9% versus 11.1%, OR 3.82 (95% CI, 0.98, 14.98), with separation evident from Week 1 onward. Subgroup B (n=119) showed intermediate benefit with Δ = −2.85 (p = 0.037, Cohen’s d 0.49). These findings suggest biologically informative refinement of the classifier, and support optimization of patient enrichment strategies for subsequent development. BH-200 was generally well tolerated. The most frequent adverse event was headache (8.9%). Elevations of ALT or AST > 3× ULN occurred in 5.8% of treated patients. These changes were asymptomatic and reversible during treatment or after discontinuation. No serious adverse events occurred in the active arm, and the hepatic signal appears monitorable within standard safety frameworks. The OLIVE program demonstrates that vasopressin V1b receptor antagonism is a viable antidepressant mechanism, and that integration of a biologically grounded genetic companion diagnostic can enhance signal detection and enrich for larger clinical benefit. This precision psychiatry strategy informs the next phase of clinical development for BH-200.