Potency Assays for Cell and Gene Therapy: Types, Strategy, and FDA Expectations

What is a potency assay?

A potency assay is a quantitative laboratory test used to measure the biological activity and therapeutic strength of a biopharmaceutical product. For a cell or gene therapy, that means a readout tied to the mechanism of action: a CAR-T product killing antigen-positive cells, or a gene therapy vector producing functional protein in the cells it transduces.

For clarity, this guide focuses specifically on potency assays as the validated, quantitative QC tests used for lot release in cell and gene therapy manufacturing.

The regulatory definition is broader. Under 21 CFR 600.3(s), potency is the specific ability or capacity of a product, shown by appropriate laboratory tests or adequately controlled clinical data, to effect a given result.^[1] Section 610.10 then requires potency testing, in vitro, in vivo or both, designed specifically for each product.^[1] In practice every licensed lot is released against a potency specification, and potency data also carry comparability after process changes, stability studies, and expiry dating.

Potency is also distinct from identity. A flow assay showing that 40% of cells express the CAR tells you what the product contains. It says nothing about whether those cells do anything when they meet their target, which is the question a potency assay answers.

Why potency assay strategy speeds up development

Potency assays are often treated as a late CMC deliverable. The programs that move fastest treat them as a development tool, because the same assay that eventually releases commercial lots is the one that tells you whether a process change helped or hurt.

FDA said as much in 2011, advising sponsors to start potency assay development through product characterization during preclinical and early clinical work, and to evaluate more than one assay while they do.^[2] The 2023 draft guidance goes further and recommends developing multiple assays against known or potential potency-related attributes and running them in parallel during early trials, dropping the redundant ones later.^[3]

Starting early pays off in four places:

• Process development. Culture conditions, gene modification steps and cell selection can be compared on the attribute that actually matters.
• Comparability. Moving from a manual process to a closed or automated one is far easier to defend with a quantitative potency readout already in place.
• Qualification. Assays built on research-grade target cell lines or instruments often cannot be qualified for GMP use. Target lines may lack the documentation a submission needs, and switching materials later means bridging studies and requalification.
• Release timing. Autologous products are often made fresh with a short shelf life. A potency strategy designed around that constraint avoids a release test that cannot finish before the product expires.

The cost of getting it wrong is well documented. Salmikangas and colleagues reported that, in their experience, major issues with potency tests were noted in almost 50% of all advanced therapy marketing authorization applications in the EU.^[4] In the US, Iovance's TIL therapy lifileucel was delayed for roughly three years while FDA asked for additional potency assay work, and the program moved forward on a multi-assay strategy that included a functional co-culture assay. It was eventually approved in February 2024. Clinical results alone did not carry it.

Types of potency assays

Potency assays fall into a handful of families. Most cell and gene therapy programs end up using more than one.

In vivo assays

Animal models measure activity in a whole organism, which is their strength and their problem. They are slow, variable and difficult to validate, and FDA's 2023 draft encourages replacing them with in vitro methods where possible.^[3] They still have a role in product characterization and early proof of concept.

Cell-based functional bioassays

The closest in vitro match to mechanism of action. Common formats include target cell killing, cytokine release after antigen stimulation, proliferation, and reporter gene assays in which an engineered line produces a luminescent or fluorescent signal when a pathway is activated. They are the most biologically relevant option and usually the most variable, because they depend on living target or reporter cells.

Binding and biochemical assays

Ligand binding, ELISA and enzyme activity assays are precise and easy to validate. On their own they rarely capture the full activity of a cell therapy, but they are useful components of a matrix and are often the main potency readout for secreted or expressed proteins.

Flow cytometry assays

Flow cytometry reads potency-related attributes on individual cells: CAR or transgene expression, activation markers such as CD69 and CD25, degranulation measured by CD107a, and intracellular cytokines. Because it resolves subpopulations, it often sits at the center of a cell therapy matrix. Its results are only as stable as the controls and gating behind them, which is covered in more detail in our guide to flow cytometry controls.

Molecular and analytical surrogate assays

qPCR or ddPCR for transgene mRNA, protein expression assays and other analytical methods can stand in for biological activity, but only once a sponsor has shown, with suitably qualified assays and data from multiple lots, that the surrogate correlates with a relevant biological activity.^[2] Vector copy number, for example, is a useful attribute, but it is not a potency measure on its own.

Two further distinctions matter for release. Quantitative assays report a value, qualitative assays report pass or fail, and FDA expects any qualitative assay used for potency to be paired with at least one quantitative one.^[2] And most bioassays report relative potency, the activity of a test lot compared with a reference material run in the same assay, which makes the reference itself a critical reagent.

The assay matrix approach

An assay matrix is a set of complementary assays that together give an adequate measure of potency when no single assay can. The matrix can combine biological assays, biological and analytical assays, or analytical assays alone, as long as the results correlate with a relevant biological activity.^[2]

FDA's 2011 guidance lists the situations where a matrix is likely to be needed:^[2]

• The mechanism of action is complex or not fully characterized.
• The product has multiple active ingredients or multiple biological activities.
• The product has limited stability.
• The available bioassay is not quantitative, not robust enough, or lacks precision.

Most cell therapies tick at least two of those boxes. A typical CAR-T matrix pairs a flow cytometry measure of CAR expression with a functional readout such as IFN-gamma release or cytotoxicity against antigen-positive targets. The flow assay is precise and fast; the functional assay ties the result to mechanism. Neither would be sufficient alone.

The 2023 draft reframes the matrix as one part of a wider potency assurance strategy that also covers process design, process control, material control and in-process testing.^[3] For products with a very short shelf life, it describes releasing on physicochemical potency assays and predictive in-process tests, then running the potency bioassay immediately after manufacture and reviewing it post-release.

CAR-T potency assays

A CAR-T product works through a chain of events: the CAR binds its antigen, the T cell activates, releases cytokines, kills the target cell and proliferates. A potency assay can sample any link in that chain, and a good matrix samples more than one.

• Recognition: CAR expression and antigen binding by flow cytometry.
• Activation: CD69, CD25 and CD107a after co-culture with antigen-positive targets.
• Effector function: IFN-gamma and other cytokines released on stimulation.
• Killing: cytotoxicity against target cells, by impedance, luminescence, release assays or flow.

The weak point in most of these assays is the target. Tumor cell lines drift with passage number, including antigen density, and antigen density drives activation strength directly. A line that expressed CD19 at one level in year one will not necessarily match it in year three, and because potency is measured against that target, the drift shows up as a change in the product.

Synthetic targets remove that variable. Engineered cell mimics such as StimCytes carry BCMA, CD19 or CD20 at a fixed antigen density and can stimulate target-specific CAR-T cells in vitro, giving a mechanism-of-action readout such as IFN-gamma release against a target that holds its specification across lots, sites and years. They also sidestep the chain-of-custody and culture maintenance work that comes with qualifying a cell line. For flow-based components of the matrix, the T Cell Activation/Exhaustion Mimic provides a consistent positive control for activation and exhaustion markers.

The same fixed-density principle underpins our functional CAR-T benchmarking controls, and in-house cryopreserved CAR-T controls remain an option for labs that can manage their short post-thaw working window.^[5]

Gene therapy potency assays

For a viral vector given directly to patients, potency has two parts: the vector has to get its payload into cells, and the payload has to produce a functional product. FDA's 2023 draft is specific here. Even at the earliest stage, release testing for a vector that expresses a transgene should generally include a potency assay that quantifies transgene mRNA or protein in transduced cells.^[3] As development continues, sponsors are expected to evaluate risks to the vector particles and nucleic acid more comprehensively and to add assays for the functional activity of the expressed product, such as enzyme activity or a downstream cellular response.

The practical challenges are different from cell therapy. Transduction depends heavily on the permissiveness of the cell line used, the multiplicity of infection and the assay window, so the cell substrate for a gene therapy potency assay needs the same control over drift as a CAR-T target line. Expression assays are precise but only indirectly functional, which is why gene therapy potency usually ends up as a matrix as well.

Ex vivo gene-modified cell products, CAR-T included, sit between the two categories. They are typically characterized by vector copy number and transgene expression, but release still needs a readout of what the modified cells do.

What FDA expects from IND to BLA

FDA's current potency guidance for cell and gene therapy is the 2011 final guidance.^[2] The December 2023 draft on potency assurance will supersede it when finalized, and it is already shaping review conversations.^[3] Together, they set out a progressive path.

Before the first IND

The 2023 draft asks sponsors to identify initial potency-related critical quality attributes, perform a risk assessment, and describe the potency assurance strategy in Module 3 of the initial IND, including the mechanism of action, the quality target product profile, and the potency assays and acceptance criteria used for release.^[3]

Early-phase trials

For preclinical, Phase 1 and early Phase 2 studies, limited quantitative information on relevant biological attributes may be sufficient. Acceptance criteria should be set as a numerical range and are expected to be wider than in later phases, tightening as manufacturing and clinical experience grow.^[2]

Pivotal trials

Lots used in pivotal studies need a potency assay or matrix with established limits. FDA's 2011 guidance is blunt about the consequence: without sufficient assurance of potency for those lots, the trial may be considered deficient in design and placed on clinical hold.^[2] Stability studies for expiry dating also need a potency assay that is stability indicating.

Licensure

A BLA must describe and justify a validated potency assay or assay matrix with defined acceptance criteria, based on data collected across all phases of development.^[2] Validation follows the key parameters: accuracy, precision, specificity, linearity, range and robustness, plus sample and system suitability criteria. ICH Q14 and Q2(R2) and the USP bioassay chapters <1032>, <1033> and <1034> are the working references for design, validation and analysis.^[6][7]

Reference materials and controls

Every potency number is relative to something. Most bioassays are calibrated against a reference material assigned an arbitrary potency value, and for cell and gene therapies there is usually no compendial standard to use. The 2023 draft expects sponsors to develop an in-house reference, often a well-characterized drug product lot, and to qualify replacement lots against the current one with multiple independent assays before supplies run out.^[3]

It also recommends a separate control material in every run for system suitability, alongside positive and negative controls with predefined acceptance criteria, and control charting of those results to catch drift over the life of the assay.^[3] This is where many programs struggle, as donor cells vary, cell lines drift, and a reference lot of drug product is finite.

Precision-engineered cell mimics address the control side of that problem. They are manufactured to a specification rather than harvested or cultured, so they give you:

• Earlier assay development, before clinical material or a final manufacturing process exists.
• Lot-consistent positive and negative controls across runs, operators and sites.
• Tunable antigen density, so activation strength and assay sensitivity can be set deliberately. See quantitative antigen density.
• Simpler reagent qualification, without the chain-of-custody, modification tracking and culture maintenance that cell lines need.

Where a program needs a target or control that no catalog product covers, custom TruCytes mimics can be built to a defined phenotype. The full range is collected under potency controls, and how cell mimics work explains the underlying technology.