FDA vs. EMA Manufacturing Controls And GMP For Cell And Gene Therapy At The IND/CTA Stage

FDA vs EMA manufacturing controls and GMP for cell and gene therapy at the IND / clinical-trial-application stage

For cell and gene therapy (CGT / ATMP) developers, the first regulatory divergence is not efficacy or non-clinical data. It is manufacturing. FDA and the EU authorities reach broadly similar destinations on product control, yet they get there through structurally different legal machinery at the point a trial begins. FDA runs on an IND with phase-appropriate cGMP and no manufacturing licence; the EU runs on a clinical-trial application (CTA) that sits on top of a manufacturing/import authorisation and Qualified Person (QP) batch release. Understanding that split, and the many places where the two systems quietly converge, is what determines how a sponsor builds its early-phase quality strategy for a bilateral program.

The headline difference: a manufacturing licence and QP release vs. a Part 211 exemption

The single most consequential difference is procedural. In the EU, manufacture and import of an investigational medicinal product (IMP), including an investigational ATMP, is generally subject to a manufacturing/import authorisation under Article 61 of Regulation (EU) 536/2014, the MIA-IMP. To hold it, the manufacturer must have suitable premises, equipment and control facilities, and must permanently have at least one appropriately qualified QP ¹⁰⁷. Before any batch is used in a trial, the QP must certify that each batch was manufactured and checked in accordance with the clinical-trial authorisation, EU GMP, and other applicable requirements; for imported product the QP must confirm manufacture to a standard at least equivalent to EU GMP ^1045956. The product stays under sponsor control until both QP certification and sponsor release have occurred ⁵⁹.

FDA has no analogous manufacturing licence for investigational products. There is no MIA-IMP and no QP concept. Instead, most Phase 1 investigational drugs, including biological products, are exempt under 21 CFR 210.2(c) from the finished-pharmaceutical cGMP regulations of 21 CFR Part 211 ²⁸⁴¹. That is not an exemption from cGMP as a statutory matter: the FD&C Act section 501(a)(2)(B) standard still applies, and FDA expects controls appropriate to a first-in-human, small-scale, safety-focused study, described in "Current Good Manufacturing Practice for Phase 1 Investigational Drugs" (2008) ²⁸³². The practical effect is that an EU sponsor must stand up a GMP-licensed, QP-covered manufacturing operation to dose its first patient, whereas a US sponsor can begin under a lighter, phase-appropriate control set without a facility licence.

Phase-appropriate cGMP vs. the risk-based approach: same philosophy, different labels

Both systems reject a one-size-fits-all GMP standard for investigational CGT, but they codify the flexibility differently.

FDA calls it phase-appropriate cGMP. The amount and nature of CMC information, and the rigor of controls, is commensurate with clinical phase ³⁹⁴⁰. The 21 CFR 210.2(c) exemption from Part 211 applies to Phase 1 studies as defined in 21 CFR 312.21(a), typically small studies of safety and pharmacology that may also seek early evidence of effectiveness, and FDA is explicit that CGT products manufactured for Phase 2 or Phase 3 must comply with 21 CFR Parts 210 and 211 ⁴¹³⁴³⁰. In other words, the regulatory floor steps up at the Phase 1 / Phase 2 boundary. FDA does not expect process validation for investigational products, but expects studies throughout development to build the knowledge needed for consistent commercial manufacture, with release acceptance criteria required from the Phase 1 IND under 21 CFR 312.23(a)(7) and tightened as product knowledge grows before Phase 2/3 ²⁹⁴¹.

The EU calls it the risk-based approach (RBA), set out in the "Guidelines on Good Manufacturing Practice specific to Advanced Therapy Medicinal Products" (EudraLex Volume 4, Part IV). Critically, Part IV expressly applies to investigational ATMPs, not only authorised ones ⁵²⁶⁷. The manufacturer designs organisational, technical and structural GMP measures according to the specific risks of the product and process, with effort and documentation proportionate to risk, and with adaptations particularly justified in Phase I and Phase I/II where product knowledge is still developing ⁴⁵⁴⁷. But the RBA is not a derogation: it does not override the clinical-trial authorisation, does not remove any regulatory obligation, and the CTA itself is expected to explain the quality strategy where an RBA is used ⁵²⁴⁷. A functional pharmaceutical quality system is required from the first stages of investigational development ⁴⁷.

The philosophical convergence is real: both agencies let controls mature with knowledge, both require predefined controls to tighten before pivotal trials, and neither demands full process validation for investigational material. The mechanical difference is that the EU threshold is a documented, authority-visible risk rationale embedded in the CTA, while the US threshold is a bright-line regulatory-phase transition into full Part 211.

The CMC / quality dossier

FDA's IND-stage expectation, in "Chemistry, Manufacturing, and Control (CMC) Information for Human Gene Therapy Investigational New Drug Applications (INDs)," is that the IND contain enough CMC information to assure safety, identity, quality, purity and strength (including potency) under 21 CFR 312.23(a)(7)(i) ⁴. Sponsors are expected to develop a controlled process capable of producing consistent critical quality attributes (CQAs), identify CQAs through characterization, and establish critical process parameters (CPPs) through process qualification, with this discipline emphasized for complex multistep products such as CAR T cells ⁶.

The EU equivalent is the IMPD. EMA's current guideline, "Guideline on quality, non-clinical and clinical requirements for investigational advanced therapy medicinal products in clinical trials" (EMA/CAT/22473/2025, effective 1 July 2025), governs CTAs for investigational gene- and cell-therapy products, with particular focus on early/exploratory trials ¹²³¹³¹. The IMPD should follow CTD Module 3 structure, split into active-substance and finished-product sections, and document process parameters, in-process controls, release specifications and acceptance criteria even in exploratory trials, reviewed as development advances ¹²⁷. The extent of quality data may be adapted to identified risks (cell origin, vector type, genetic-modification method, non-cellular components, intended use), with the risk analysis updated over the lifecycle and summarised in the IMPD ¹³¹. Notably, the EMA guideline explicitly cross-refers manufacture back to the Volume 4 Part IV ATMP GMP guideline, tying the quality dossier and the GMP regime together in a way the FDA IND structure does not formally mirror ¹²⁷¹³⁵.

The general-drug analogue on the EU side, the CTR quality guideline "Guideline on the requirements to the chemical and pharmaceutical quality documentation concerning investigational medicinal products in clinical trials (Rev. 2, January 2022)," reinforces that IMPD quality documentation is risk- and development-stage-proportionate, not the full marketing-authorisation standard, and can be simplified where the active-substance source is already in an EU-authorised medicine ⁹³⁹¹.

Potency

FDA has moved potency expectations noticeably earlier than a casual reading of "phase-appropriate" would suggest. The draft "Potency Assurance for Cellular and Gene Therapy Products" recommends that even at the earliest development stage, release testing for a directly administered viral vector expressing a transgene generally include a potency assay quantifying transgene mRNA or protein in transduced cells ⁵. For very short shelf-life products, FDA's draft approach combines physicochemical lot-release potency assays with in-process testing predictive of potency, plus post-release confirmatory bioassays ⁵. A final-product biological-activity potency assay with lot-release specifications should be established before a trial intended to provide substantial evidence of effectiveness ¹⁵.

The EMA investigational-ATMP guideline, as captured in the retrieved text, does not set a standalone potency-assay design or acceptance-criteria requirement for early trials; instead potency sits inside the general obligation to document release specifications and acceptance criteria (reviewed as development advances) and to use only analytical methods suitable for their intended purpose ¹²⁷. In practice this makes FDA's early-potency signalling more prescriptive on paper, while the EU folds potency into the risk-based specification/control strategy.

Comparability and manufacturing changes

Both systems treat manufacturing change as an expected, managed feature of CGT development, and both anchor it in risk assessment. FDA expects sponsors to understand a change's effect on product quality using CQAs identified through characterization, treats a change that could affect quality as essential information to be submitted as an IND information amendment under 21 CFR 312.31(a)(1), and points to the draft "Manufacturing Changes and Comparability for Human Cellular and Gene Therapy Products" ⁶².

EMA's emphasis is on timing. It does not recommend substantial manufacturing changes during pivotal studies, because they create ATMP comparability problems at marketing-authorisation stage and can undermine the validity of data generated with pre-change material; confirmatory studies should use a process that is as mature as feasible ¹²⁷. EMA directs applicants to "Questions and answers on Comparability considerations for Advanced Therapy Medicinal Products" (EMA/CAT/499821/2019) ¹³⁴. The shared message is to front-load process maturity; the EU states the "no change during pivotal" expectation more bluntly.

Starting materials, vectors and cell banks

This is an area of strong technical convergence, expressed in each system's own vocabulary.

FDA's vector-specific expectations include replication-competent retrovirus testing at multiple production points; qualification of the master viral bank for replication-competent adenovirus (RCA) with per-lot testing and a recommended maximum of 1 RCA per 3x10^10 viral particles; and, for plasmids, complete sequencing and contaminant specifications ²³¹⁸. For CAR T and other cellular products, the IND should describe leukapheresis handling from collection through manufacturing (washing, cryopreservation, shipment, temperature monitoring), with incoming-material acceptance criteria (e.g., minimum cell number, viability, CD3+ percentage) developed as experience accumulates ⁷⁷.

EMA's investigational-ATMP guideline maps the same ground onto formal starting-material definitions: critical starting materials should receive the same level of information as the active substance, potentially in a dedicated S.2.3 section with stability data ¹³⁰. For ex vivo gene therapy, the starting materials include the unmodified cells, the vector, any nucleic acid/protein used for modification, and the components used to produce them ¹²⁰. A master cell bank (and master seed bank) should be established, where possible, before exploratory trials, characterised per ICH Q5D, though a working cell bank may not be available early ¹²⁰¹²⁸. For replication-deficient viral vectors, demonstrated absence of replication-competent virus at the vector starting-material level (validated method) can remove the need for further downstream RCV testing only if a risk assessment rules out RCV generation during manufacture ¹²⁰. Raw materials of biological origin are pointed to Ph. Eur. 5.2.12 ¹³⁰. The technical bar is very similar; the EU simply codifies it through starting-material classification and pharmacopoeial cross-reference.

Sterility, aseptic processing and short shelf-life logistics

Both agencies foreground the defining manufacturing problem of living-cell products: they generally cannot be terminally sterilized, may have very short dating, and often must be released before all test results return.

FDA identifies aseptic processing, contamination control, stability and release planning as central IND CMC issues for somatic cell therapy ⁸⁶. For Phase 1, FDA points to ISO 5 / Class A laminar-airflow conditions for sterile manipulations and recommends process simulation (media fill) to show the process/environment can produce sterile product; in multi-product facilities it expects controls preventing cross-contamination and demonstrating removal of prior product from shared equipment, especially with live viral/vector processing ²⁷. For fresh CAR T, the maximum formulation-to-infusion time should be defined and supported by stability data, with logistics for sampling, QC reporting, QA lot-release review and clinical-site handling; cryopreservation is noted as a way to buy time for full release testing ⁷⁸. Chain of identity is a specific FDA control: at least two unique identifiers on autologous product, batch-record label checks before each step, and two independent patient-and-label checks at bedside ⁷⁷.

EMA's Part IV requires validation of aseptic processing (with sterilisation-process validation meeting the marketed-product standard even for investigational material), verification of air-quality-system suitability and the premises' contamination control for investigational ATMPs, and a development-stage approach to method validation in which sterility, microbial and other safety-critical assays are validated already in first-in-human/exploratory trials, while batch-release and stability method validation is expected for pivotal trials ^5862132. The EMA quality guideline separately requires batch definition and traceability from cell sourcing to final-container labelling ¹²². The substance is close; the EU expresses it as explicit validation obligations tied to trial phase, with the QP verification step (autologous starting-material/recipient matching, in-process controls, environmental monitoring, deviation impact, storage/transport) acting as a formal gate that has no direct FDA counterpart ⁶⁴.

What this means for a bilateral program

For a sponsor planning first-in-human CGT trials on both sides of the Atlantic, the practical planning conclusions are:

• Budget for an EU manufacturing/import authorisation (MIA-IMP) and a named QP before your first EU dose; there is no equivalent gating step for a US IND ¹⁰⁷²⁸.
• Expect the EU to require your risk-based GMP rationale to be written into the CTA and consistent with the clinical-trial authorisation, whereas FDA's flexibility is a regulatory-phase feature that flips to full Part 211 at Phase 2 ⁴⁷³⁴.
• Treat FDA's early-potency expectations as the more explicit near-term signal, and the EU's "no substantial change during pivotal" position as the more explicit comparability-timing constraint; a program that satisfies both will front-load both a quantitative potency assay and process maturity ⁵¹²⁷.
• On vectors, cell banks, sterility and chain of identity, the technical requirements are close enough that a single well-designed control strategy can usually satisfy both, provided it is documented in each system's expected format (IND CMC vs. CTD Module 3 IMPD with QP-verifiable records) ^2312058.