Potential biomarker of brain response to opioid antagonism in adolescents with eating disorders: a pilot study

Abstract

Background: Eating Disorders (ED) affect up to 5% of youth and are associated with reward system alterations and compulsive behaviors. Naltrexone, an opioid antagonist, is used to treat ED behaviors such as binge eating and/or purging. The presumed mechanism of action is blockade of reward activation; however, not all patients respond, and the optimal dose is unknown. Developing a tool to detect objective drug response in the brain will facilitate drug development and therapeutic optimization. This pilot study evaluated neuroimaging as a pharmacodynamic biomarker of opioid antagonism in adolescents with ED.

Methods: Youth aged 13-21 with binge/purge ED completed functional magnetic resonance imaging (fMRI) pre- and post-oral naltrexone. fMRI detected blood oxygenation-level dependent (BOLD) signal at rest and during two reward probes (monetary incentive delay, MID, and passive food view, PFV) in predefined regions of interest associated with reward and inhibitory control. Effect sizes for Δ%BOLD (post-naltrexone vs. baseline) were estimated using linear mixed effects modeling.

Results: In 12 youth (16-21 years, 92% female), BOLD signal changes were detected following naltrexone in the nucleus accumbens during PFV (Δ%BOLD -0.08 ± 0.03; Cohen's d -1.06, p = 0.048) and anterior cingulate cortex during MID (Δ%BOLD 0.06 ± 0.03; Cohen's d 1.25, p = 0.086).

Conclusion: fMRI detected acute reward pathway modulation in this small sample of adolescents with binge/purge ED. If validated in future, larger trials, task-based Δ%BOLD detected by fMRI may serve as a pharmacodynamic biomarker of opioid antagonism to facilitate the development of novel therapeutics targeting the reward pathway, enable quantitative pharmacology trials, and inform drug dosing.

Clinical trial registration: https://clinicaltrials.gov/ct2/show/NCT04935931, NCT#04935931.

Keywords: adolescents; eating disorders; fMRI; naltrexone; opioid antagonism; pharmacodynamic biomarker.

Figure 1

A priori-defined Regions of Interest (ROIs). ROIs displayed on 3D merged image from all study participant brains. MNI coordinates expressed as x, y, z for regions of interest: ACC (L) -4, 32, 18 (R) 4, 32, 18; dlPFC -22, 52, 30; NAc (L) -12, 8, −8 (R) 14, 10, −8; vmPFC 9, 46, −13.

Figure 2

Group level reward pathway modulation by opioid antagonism. The forest plots display the linear random mixed effects model-derived mean and 95% confidence interval for %BOLD signal change following naltrexone. A priori regions of interest were nucleus accumbens, NAc, ventromedial prefrontal cortex, vmPFC, anterior cingulate cortex, ACC, and dorsolateral prefrontal cortex, dlPFC. Contrasts selected demonstrated Cohen’s d ≥ 0.8. *p < 0.05.

Figure 3

Individual-level Δ%BOLD following opioid antagonism. The boxplots display individual-level data points that represent the raw values of Δ%BOLD. The size of the data point corresponds with the amount of uncensored data for that individual (e., 1.0 means 100% of the data for that individual were uncensored; 0.8 means 80% of that individual’s data were uncensored and 20% censored due to motion). Following naltrexone, increased activity was seen in the anterior cingulate cortex, ACC, during reward anticipation (Cohen’s d 1.25, p = 0.086). Following naltrexone, reduced activity was seen in the nucleus accumbens, NAc, during passive food view task (Cohen’s d − 1.06, p = 0.048).