Substance use disorders and anxiety are clinically heterogeneous, and broad case-control genetic designs can obscure stage- and subtype-specific biology. A recent series of five transcriptome-wide association study (TWAS) preprints by Cheung examined opioid exposure versus dependence progression, alcohol misuse latent classes, cannabis use disorder, anxiety, and aging-related gene sets. Together, these studies suggest a cross-disorder architecture organized around a Pruning-Plasticity-Aging axis. This synthesis is narrative and hypothesis-generating, because the primary cross-disorder evidence comes from preprints and genetically predicted expression rather than measured patient expression. In this model, liability is not distributed along a single psychiatric-risk continuum. Instead, opioid exposure and some lower-risk or internalizing alcohol profiles appear to involve altered neuroimmune pruning and glial-synaptic refinement, whereas opioid dependence progression, heavier alcohol classes, broad-risk alcohol profiles, and cannabis use disorder show stronger involvement of glutamatergic plasticity, presynaptic adaptation, reward-circuit remodeling, AMP-activated protein kinase-mechanistic target of rapamycin (AMPK-mTOR) nutrient sensing, mitochondrial bioenergetics, and selected nicotinamide adenine dinucleotide (NAD)/sirtuin stress-response branches. Anxiety diverges by loading more strongly on inflammatory-apoptotic signaling, complement/sterile alpha and toll/interleukin receptor motif-containing protein 1 (SARM1)-linked axonal stress, and mitochondrial NAD-stress programmes. The most promising findings include stage-specific opposition in opioid TWAS profiles, latent-class plasticity inversion in alcohol misuse, a triggering receptor expressed on myeloid cells 2 (TREM2)-positive cannabis use disorder profile contrasted with a SARM1/complement C1q C chain (C1QC)-positive anxiety profile, and branch-specific rather than global aging biology. This review synthesizes these findings; places them in established addiction, neuroimmune, and aging biology; and outlines validation priorities for colocalization, fine-mapping, cell-type resolution, and functional testing.
