Programmable genetic medicines: delivery, manufacturability and clinical translation

Programmable genetic medicines can enable transient protein expression, transcript silencing, splice correction, durable gene addition, sequence rewriting or cellular reprogramming. Their translation nevertheless depends on whether an engineered instruction reaches the intended cells at an interpretable dose and duration in a product that can be characterized, manufactured and clinically managed. We synthesize RNA medicines, viral vectors, genome editors and engineered cells within a multidimensional translational architecture that distinguishes therapeutic operation, payload, delivery architecture and final product format. Clinical experience in liver, eye, blood and immune-cell settings shows how receptor-mediated uptake, local administration or ex vivo control can reduce delivery uncertainty, whereas CNS, muscle, lung and solid-tumour applications remain constrained by distribution, dose and heterogeneity. A ClinicalTrials.gov search performed on 26 June 2026 retrieved 2,551 unique records after NCT-level deduplication; rule-based screening retained 1,667 records in the core descriptive set. These are registered-study records rather than unique products, participants, efficacy estimates or measures of platform maturity. We also examine how CMC, potency, comparability, immunogenicity and long-term follow-up shape the interpretation of platform knowledge. The relevant unit of translation is a payload-delivery-manufacturing-regulatory architecture rather than a payload alone. The proposed domains are a conceptual discussion tool, not a validated scoring system, and they clarify why modular payload changes may require renewed biological, CMC and clinical interpretation.