Selective transcriptomic recovery by (2R,6R)-hydroxynorketamine in opioid-abstinent mice: Machine learning identifies predictive biomarkers

Abstract

BackgroundOpioid abstinence induces persistent emotional disturbances and widespread neuroplastic changes, in areas of the brain including the hippocampus. Although ketamine and its metabolite (2R,6R)-hydroxynorketamine (HNK) show potential in reversing opioid abstinence-related deficits in rodents, the molecular mechanisms underlying their efficacy remain poorly understood. MethodsMale C57BL/6J mice underwent a 3-week opioid abstinence paradigm, followed by a single (2R,6R)-HNK (10 mg/kg, i.p.) or saline injection on day 28. Sucrose and social preference tests were used to assess behavioral deficits. We conducted RNA sequencing of ventral hippocampal tissue from these mice, followed by differential gene expression and functional enrichment analyses. Additionally, Random Forest machine was applied to identify predictive differentially expressed genes (DEGs) associated with (2R,6R)-HNK treatment response. ResultsTranscriptomic analysis identified 206 DEGs in morphine-abstinent mice without treatment compared to opioid-naive controls (MOR-SAL vs. SAL-SAL), implicating altered immune signaling, synaptic function, and structural plasticity. Comparison of opioid-abstinent mice treated with (2R,6R)-HNK to opioid-naive controls (MOR-HNK vs. SAL-SAL) revealed 186 residual DEGs, enriched for Th17-mediated immune and fear regulation pathways, suggesting a persistent intermediate molecular phenotype despite normalized behavioural scores. DEGs overlap analysis between MOR-HNK vs. MOR-SAL and MOR-SAL vs. SAL-SAL indicated that (2R,6R)-HNK treatment reversed 55 DEGs in opioid-abstinent mice, including Transthyretin (Ttr) and T-cell surface glycoprotein (Cd5) expression levels. Machine learning identified interleukin 1 receptor accessory protein-like 1 (Il1rapl1) and cytotoxic T lymphocyte-associated protein 2 beta (Ctla2b) as top predictors of (2R,6R)-HNKs treatment response. Notably, while (2R,6R)-HNK induces transcriptional changes in opioid-naive mice (SAL-HNK), it does not affect behavior compared to untreated controls (SAL-SAL). In contrast, its therapeutic effects are evident in morphine-abstinent mice (MOR-HNK), highlighting its context-dependent efficacy. Conclusion(2R,6R)-HNK promotes both transcriptional and behavioral recovery in opioid-abstinent mice, reversing key gene expression changes. However, persistent dysregulation of neuroimmune and emotion-related pathways suggests an intermediate molecular state, reflecting ongoing recovery.