Utility of Imaging-Based Biomarkers for Glutamate-Targeted Drug Development in Psychotic Disorders: A Randomized Clinical Trial.

Importance: Despite strong theoretical rationale and preclinical evidence, several glutamate-targeted treatments for schizophrenia have failed in recent pivotal trials, prompting questions as to target validity, compound inadequacy, or lack of target engagement. A key limitation for glutamate-based treatment development is the lack of functional target-engagement biomarkers for translation between preclinical and early-stage clinical studies. We evaluated the utility of 3 potential biomarkers-ketamine-evoked changes in the functional magnetic imaging (fMRI) blood oxygen level-dependent response (pharmacoBOLD), glutamate proton magnetic resonance spectroscopy (1H MRS), and task-based fMRI-for detecting ketamine-related alterations in brain glutamate. Objective: To identify measures with sufficient effect size and cross-site reliability to serve as glutamatergic target engagement biomarkers within early-phase clinical studies. Design, Setting, and Participants: This randomized clinical trial was conducted at an academic research institution between May 2014 and October 2015 as part of the National Institute of Mental Health-funded Fast-Fail Trial for Psychotic Spectrum Disorders project. All raters were blinded to study group. Healthy volunteers aged 18 to 55 years of either sex and free of significant medical or psychiatric history were recruited from 3 sites. Data were analyzed between November 2015 and December 2016. Interventions: Volunteers received either sequential ketamine (0.23 mg/kg infusion over 1 minute followed by 0.58 mg/kg/h infusion over 30 minutes and then 0.29 mg/kg/h infusion over 29 minutes) or placebo infusions. Main Outcomes and Measures: Ketamine-induced changes in pharmacoBOLD, 1H MRS, and task-based fMRI measures, along with symptom ratings. Measures were prespecified prior to data collection. Results: Of the 65 volunteers, 41 (63%) were male, and the mean (SD) age was 31.1 (9.6) years; 59 (91%) had at least 1 valid scan. A total of 53 volunteers (82%) completed both ketamine infusions. In pharmacoBOLD, a highly robust increase (Cohen d = 5.4; P < .001) in fMRI response was observed, with a consistent response across sites. A smaller but significant signal (Cohen d = 0.64; P = .04) was also observed in 1H MRS-determined levels of glutamate+glutamine immediately following ketamine infusion. By contrast, no significant differences in task-activated fMRI responses were found between groups. Conclusions and Relevance: These findings demonstrate robust effects of ketamine on pharmacoBOLD across sites, supporting its utility for definitive assessment of functional target engagement. Other measures, while sensitive to ketamine effects, were not sufficiently robust for use as cross-site target engagement measures. Trial Registration: clinicaltrials.gov Identifier: NCT02134951.