What does it take to develop and manufacture a Gene Therapy (AAV / Viral Vector) drug?
Gene therapy stretches a development team across two very different worlds. The early science looks like any biologics program: pick a transgene, choose or engineer a capsid serotype that reaches the target tissue, design the expression cassette and promoter, then prove in preclinical models that the construct does what you claim without lighting up the immune system. The hard, expensive reality lands later, in CMC. The vector you make is not a small molecule with a clean structure; it is a population of viral particles, some full of your genome and some empty, and your process and your analytics define the product as much as the sequence does. That is why so many gene therapy programs that read well on biology stall at manufacturing.
A typical program needs a CRO for the discovery and preclinical work (capsid screening, biodistribution and shedding studies, immunogenicity and GLP toxicology in a relevant species), then a CDMO for process development and GMP supply. The manufacturing side is where the specialist gap is widest. Upstream, you are running either a transient transfection process (three or four plasmids into HEK293 cells, which means you also need GMP plasmid DNA) or a producer or packaging cell line. Downstream, you are clarifying, concentrating, and polishing with affinity chromatography and tangential flow filtration, then trying to separate full capsids from empty ones, often by anion exchange or gradient ultracentrifugation. Each step bleeds yield, and AAV yields are low to begin with, so process choices made early ripple straight into cost of goods and how many patients a batch can dose.
The analytics are the part generalists underestimate. You need vector genome titer (qPCR or ddPCR), capsid titer, the full-to-empty ratio, infectivity and potency assays that tie back to mechanism, residual host cell DNA and protein, residual plasmid and helper, replication-competent virus testing, and aggregation. A CDMO that has built and qualified these assays for AAV, and ideally has a platform process across serotypes, is in a different category from a contract biologics shop that lists gene therapy on a capabilities page. Specialist viral vector CDMOs differ from generalists in three concrete ways: a real platform process they can adapt rather than invent from scratch, a qualified AAV-specific analytical panel, and people who have actually carried a vector through a regulatory filing and the FDA potency and comparability questions that come with it.