Dynamic Alterations in the Blood Transcriptome Characterize Drug Use Behavior and Co‐Morbidities in Cocaine Use Disorder: A Preliminary Study

Individuals with cocaine use disorder (CUD) who attempt abstinence experience craving and relapse that can benefit from multimodal treatment monitoring. Longitudinal studies linking behavioral manifestations in CUD to the blood transcriptome are not only limited but also computationally complex. Therefore, we developed an analytical pipeline to investigate the connection between drug use behaviors during abstinence and change in the blood transcriptome. We conducted a longitudinal study with CUD (n = 12 subjects) and collected behavioral metrics and blood RNA-seq at baseline, 3, 6, and 9 months. Our analytical pipeline of the high-dimensional data encompasses hierarchical k-means clustering to classify subjects to responder groups based on behavioral scores and abstinence duration, in silico cell deconvolution, differential analysis with correlated multivariate testing over time, gene set enrichment analysis, and gene co-expression with time splines and RNA-seq data. The pipeline captured dynamic changes in behavioral scores and abstinence duration in responder groups. Genes showing differential transcript-level expression were enriched in substance use and cardiovascular disease-associated genetic risk loci in responder groups. Lastly, time-dependent gene co-expression revealed dynamic changes related to immune processes, cell cycle, RNA-protein synthesis, and second messenger signaling for days of abstinence. This is a preliminary investigation, providing an innovative and scalable pipeline for blood-based longitudinal RNA-seq studies in CUD, potentially applicable to other substance use disorders. It outlines a data-driven approach for analyzing composite longitudinal drug use behavioral phenotypes with blood-based transcriptomics. We also demonstrate changes in drug use behaviors and the blood transcriptome during drug abstinence.