Biotech Explained · Regenerative Medicine

How labs check a batch of cells: counts, viability, purity and potency

By Helen Venturi · · 8 min read

How labs check a batch of cells: counts, viability, purity and potency

A vial of an ordinary drug can be filtered, heated or irradiated at the end of manufacturing to make sure nothing harmful survives. A bag of living cells cannot. Anything that would kill a bacterium would also kill the medicine. So the safety and quality of a cell product depend on two things: a clean process from start to finish, and a set of tests run on every batch before anyone is allowed to give it to a patient.

Those tests are summarized on a document called a certificate of analysis. Each line pairs a question with a method and a pass mark, known as the specification. The exact list differs from product to product, but most cell therapy release panels cover the same ground.

What is checked The question it answers Common methods Typical time to result
Cell count Is there enough for the planned dose? Counting chamber, automated counter, flow cytometry Minutes
Viability What share of the cells are alive? Dye exclusion (trypan blue), fluorescent stains Minutes
Identity Are these the cells we meant to make? Flow cytometry for surface markers, sometimes genetic tests Hours
Purity What else is in the bag? Flow cytometry, residual reagent tests Hours to days
Sterility Are bacteria or fungi growing? Culture in growth media, rapid automated culture Days to 2 weeks
Mycoplasma Are these hard-to-see bacteria present? DNA-based tests, culture Hours (DNA test) to about 4 weeks (culture)
Endotoxin Is there bacterial debris that causes fever? Limulus amebocyte lysate test or a recombinant version Under an hour
Potency Will the cells do their job? Product-specific functional or marker assays Hours to weeks

The rest of this piece takes those lines in order, with most of the space going to the last one, because that is where the real difficulty sits.

Counting the cells

Dose in cell therapy is usually expressed as a number of cells, often per kilogram of the patient's body weight. That makes the count the foundation of everything else. The old method is a glass counting chamber under a microscope, with a technician tallying cells in a grid of known volume. Automated counters now do most of this work, photographing a stained sample and counting by software.

Counts depend on which cells you count. Cord blood banks report total nucleated cells, meaning everything with a nucleus, and separately count the CD34 positive cells that include the blood-forming stem cells. Our page on cord blood banking explains why those two numbers decide whether a public bank keeps a donated unit or turns it away.

Is it alive?

Viability is the share of cells that are alive at the moment of testing. The oldest test is trypan blue: a living cell with an intact membrane keeps the dye out, a dead cell lets it in and turns blue. Fluorescent stains do the same job with more precision, and some can be combined with the identity tests described below so that dead cells are excluded from the analysis.

Viability matters most around freezing. Cells are frozen with a protective additive, stored in liquid nitrogen vapor and thawed shortly before use, and each step costs some cells. Many products are therefore tested after a trial thaw as well as before freezing. Our page on stem cell technology covers the freezing process in more detail.

There is a catch built into the method. A dye test tells you the membrane is intact. It does not tell you the cell is healthy. A cell in the early stages of programmed death can still keep the dye out, which is why viability is never treated as proof that a product works.

Is it the right cell, and what else is in there?

Identity

Most identity testing relies on flow cytometry. Cells are tagged with antibodies that carry fluorescent labels, each antibody sticking to one surface protein, and then streamed one at a time past lasers. The instrument records which labels each cell carries. A T cell product should be rich in cells carrying CD3. A blood stem cell product is checked for CD34. A CAR-T product is also tested for the new receptor itself, to confirm how many cells took up the gene.

Some cell types have no single defining marker, so identity becomes a pattern. For mesenchymal stromal cells, a 2006 position statement from the International Society for Cellular Therapy, published in the journal Cytotherapy, proposed minimum criteria that are still widely quoted: the cells should stick to plastic in culture, carry the markers CD73, CD90 and CD105, lack blood cell markers such as CD45 and CD34, and be able to turn into bone, fat and cartilage cells in a dish. Even so, two products that both meet those criteria can behave very differently.

Purity

Purity looks at everything that should not be there. That includes unwanted cell types, such as stray tumor cells in a product made from a cancer patient's blood, or, for products made from stem cells, any cells that remain in an immature state and could in theory form a growth. It also includes leftovers from manufacturing: magnetic beads used to sort cells, growth factors, antibodies, residual virus used to deliver a gene, and the freezing additive itself. Each one needs its own test or a validated reason why it is not a concern.

Is it clean?

Sterility

The standard sterility test places samples in growth media and watches for bacterial or fungal growth for 14 days. Many labs now use automated culture systems that flag growth sooner, and a quick Gram stain can catch heavy contamination within an hour.

Here a practical problem appears. Some cell products are given fresh, or have a shelf life of hours to days once thawed, so the patient may receive the cells before the full culture period is over. In those cases the release decision rests on in-process tests and the faster methods, and there is a written plan to notify the treating team at once if a culture later turns positive.

Mycoplasma

Mycoplasma are tiny bacteria without a cell wall. They pass through filters that stop other bacteria, do not cloud culture media the way other contaminants do, and can quietly change how cells behave. A classic culture test takes about four weeks, which is too slow for most cell products, so DNA-based tests that return results within a day are widely used for release.

Endotoxin

Endotoxin is a component of the outer wall of many bacteria. It survives even when the bacteria are dead and can cause fever and a drop in blood pressure if infused. The traditional test uses a reagent made from horseshoe crab blood cells, which clots in its presence. A recombinant version that does not require crabs has been adopted by a growing number of labs.

Potency: the hard part

US regulations define potency as the specific ability or capacity of a product to produce a given result. For a pill, that usually means confirming the amount of active ingredient. For living cells it is much harder, for several reasons at once.

  1. The mechanism may be only partly known. Mesenchymal stromal cells are thought to act through dozens of secreted signals and contacts with immune cells. Which one should a test measure?
  2. The test has to be quantitative and reproducible, with a pass mark, run on every batch.
  3. It has to finish in time. A functional test that takes two weeks is of limited use for a product that must be infused in two days.
  4. Cells change with handling. A product can pass in the manufacturing lab and lose function during shipping or thawing.

Labs respond by building what the FDA calls a potency assurance strategy. The agency's December 2023 draft guidance, Potency Assurance for Cellular and Gene Therapy Products, describes this as a combination of process design, control of materials, in-process testing and release assays, rather than a single test carrying all the weight.

What that looks like in practice depends on the product:

  • CAR-T cells can be mixed with target cells carrying the matching marker, and the lab measures the signaling molecules the T cells release, or counts how many targets they kill.
  • Cord blood units are often tested in colony-forming assays, in which a few cells are grown in a semi-solid gel and the colonies of blood cells they form are counted after about two weeks. Because that is slow, the CD34 positive cell count serves as a faster stand-in for release.
  • Mesenchymal stromal cell products may be tested for their ability to slow the growth of activated immune cells in a dish, or for the production of a specific immune-modulating enzyme after stimulation. Neither captures everything the cells might do in a patient.

Several of the slower or more complex assays end up as characterization tests: run during development to learn about the product, but not used to release each batch.

What this means when you read about a cell product

Release testing is invisible to patients, but it shapes what reaches them. When you read about a new cell therapy, a few questions are worth keeping in mind. Does the company or lab describe what its potency assay measures, and how that links to the effect seen in patients? Are the cells given fresh or frozen, and was viability reported after thawing? For products sold outside approved trials, was any of this done at all?

That last question is the sharpest one. A clinic offering an unapproved stem cell injection may truthfully say its cells were "tested". Ask which of the eight lines in the table above were run, and what the pass marks were. Our visit to a cell therapy manufacturing lab shows what it takes to produce that paperwork for a single batch, and it is a useful yardstick for anyone comparing a licensed product with a cash-only offer.

Working this out at the bench? The Hemocytometer calculator shows the formula with your numbers.

This explainer is general information, not medical advice. Decisions about treatment, donation or trials belong with a clinician who knows your history.

More on this topic