Epigenetic and Genetic Profiling of Comorbidity Patterns among Substance Dependence Diagnoses.

Objective: This study investigated the genetic and epigenetic mechanisms underlying the comorbidity patterns of five substance dependence diagnoses (SDs; alcohol, AD; cannabis, CaD; cocaine, CoD; opioid, OD; tobacco, TD).

Methods: A latent class analysis (LCA) was performed on 31,197 individuals (average age 42±11 years; 49% females) from six cohorts to identify comorbid DSM-IV SD patterns. In subsets of this sample, we tested SD-latent classes with respect to polygenic burden of psychiatric and behavioral traits and epigenome-wide changes in three population groups.

Results: An LCA identified four latent classes related to SD comorbidities: AD+TD, CoD+TD, AD+CoD+OD+TD (i.e., polysubstance use, PSU), and TD. In the epigenome-wide association analysis, cg02833127 was associated with CoD+TD, AD+TD, and PSU latent classes. AD+TD latent class was also associated with CpG sites located on , , and , while additional epigenome-wide significant associations with CoD+TD latent class were observed in and genes. PSU-latent class was also associated with a differentially methylated region in . We also observed shared polygenic score (PGS) associations for PSU, AD+TD, and CoD+TD latent classes (i.e., attention-deficit hyperactivity disorder, anxiety, educational attainment, and schizophrenia PGS). In contrast, TD-latent class was exclusively associated with posttraumatic stress disorder-PGS. Other specific associations were observed for PSU-latent class (subjective wellbeing-PGS and neuroticism-PGS) and AD+TD-latent class (bipolar disorder-PGS).

Conclusions: We identified shared and unique genetic and epigenetic mechanisms underlying SD comorbidity patterns. These findings highlight the importance of modeling the co-occurrence of SD diagnoses when investigating the molecular basis of addiction-related traits.