One Dose. Permanent Effect. No Way to Take It Back.
Regulatory strategy for gene therapy — vectors, durability, and fifteen years of follow-up, in a category where precedent is still being written.
Every Other Drug Can Be Stopped. Yours Is Still Working in Ten Years.
Discontinuation is the safety valve underneath all of drug regulation — if a therapy harms, you stop it. Gene therapy removes that valve. A single administration can produce a permanent genetic change, which means the benefit and the risk are both open-ended, and the regulatory system compensates with the longest follow-up obligations in medicine: up to fifteen years for integrating vectors, because insertional mutagenesis is a real and delayed hazard.
It also compresses the evidence problem. These products treat rare diseases, so the trials are small, often single-arm, frequently in children, with natural history as the comparator. You are asked to characterize an irreversible intervention on thin data — which is why CMC and nonclinical carry more weight here than in any other category, and why the vector, not the transgene, is usually what the reviewer wants to talk about.
Empty capsids, residual DNA, and titre — the analytics that decide whether your dose is a dose.
Four Questions a Gene Therapy File Has to Answer Before the Clinic Does.
In a one-shot, irreversible product, these are the questions that decide the program — and none of them get easier later.
What is actually in the vial?
Full versus empty capsid ratio, residual host cell DNA and plasmid, aggregates, and a titre method whose numbers you would defend under oath. Empty capsids deliver immunogenicity without therapy — they are dose in every sense except the useful one.
Does the potency assay mean anything?
Transduction, expression, and function are three different measurements, and the agency wants the one that reflects mechanism. Building it late is the single most common reason a gene therapy BLA slips.
Who does the immune system exclude?
Pre-existing anti-AAV antibodies rule out a share of your intended population, and re-dosing is generally not available. Your eligible population is smaller than your prevalence data suggests — and that belongs in the trial design, not the commercial forecast.
How long does it last?
Durability is the value proposition and the open question. Expression that wanes at year four turns a curative claim into a conversation about re-dosing you cannot have.
The follow-up commitment for integrating vectors. Made at IND, honoured long past launch, and operationally real: patients tracked, data collected, and a registry that has to still exist when the original team does not. We design that machinery at the start, because retrofitting a fifteen-year obligation is not possible.
FDA is building routes to reuse platform data instead of repeating it — if you documented it as a platform.
FDA Is Actively Building Ways to Stop Making You Repeat Yourself.
The most important recent shift in this category is structural, not scientific. The platform technology designation program — whose first grant went to an AAV vector platform in June 2025 — plus CMC flexibilities guidance for cell and gene therapy products and a prior-knowledge framework, together move the agency away from a strictly de novo model. The explicit acknowledgement: requiring a full PPQ program for every minor payload variation on the same platform is scientifically redundant.
The strategic consequence is sharp. Companies whose vector platform is characterized, documented, and consistent can leverage it across products; companies whose “platform” is a marketing word cannot. The flexibility is real, it is earned with evidence, and the decision that determines whether you qualify is made years before you apply for it.
What a Gene Therapy Program Plans Around.
Three facts that shape a category still writing its own precedent.
Long-term follow-up for integrating vectors — the longest routine obligation in medicine, committed at IND.
The capsid ratio. Empty capsids deliver immunogenicity without therapy — a quality attribute with clinical consequences.
Designation and prior-knowledge routes now exist. They reward platforms that were documented as platforms.
Six Failure Modes We Are Brought In to Prevent.
The science is hard. The programs usually fail on the parts that were foreseeable.
Potency assay built at BLA
A functional assay developed after the pivotal data exists — and a specification that cannot be justified.
Research-grade vector into the clinic
Process developed for a lab, scaled under trial pressure, with comparability to the clinical material that nobody can demonstrate.
Seropositivity ignored in the forecast
Pre-existing immunity excluding a large share of patients — a commercial model built on prevalence rather than eligibility.
Empty capsids uncontrolled
A ratio drifting between lots, dosing immunogenicity that the clinical data then has to explain.
Long-term follow-up as an afterthought
A fifteen-year commitment with no registry design, no funding line, and no plan for patient retention.
“Platform” without the evidence
Prior-knowledge leverage assumed, and a vector platform that was never characterized consistently enough to support it.
Gene Therapy Leadership for Products the Rulebook Is Catching Up To.
Our gene therapy leads have built vector CMC packages, defended potency strategies, and designed long-term follow-up that survives the decade.
“You get one administration and fifteen years of consequences. Everything the file cannot answer before dosing, the registry answers slowly and in public.”
The discipline we bring to AAV, lentiviral, and genome-editing programs.
Developing a Gene Therapy? Build the Potency Assay and the Platform Case Now.
Bring senior gene therapy leadership in early — the decisions that decide your BLA are made in process development.
Senior-led. Embedded in your team. No junior hand-offs.