Cell Viability Assay Types: What Each Measures and When to Use It

A cell viability assay reports a number between 0 and 100%, and none of them measures whether a cell is alive. Each cell viability assay measures a proxy — reductive capacity, energy charge, membrane integrity, or DNA synthesis — and the proxy you pick determines what your dose–response curve can and cannot tell you. This guide covers the principles, the chemistry behind each format, the failure modes, and the selection logic.

What a cell viability assay actually measures

The phrase cell viability suggests a binary state: a cell is alive, or it is not. In practice, no plate-reader assay measures that state directly. Each one measures a physiological signal that healthy cells produce and dying cells stop producing, then converts that signal into a percentage relative to an untreated control.

This distinction is not pedantry. It is the single most common source of misinterpretation in cytotoxicity work. A tetrazolium assay reports how much reducing capacity a well contains. An ATP assay reports how much adenosine triphosphate a well contains. A lactate dehydrogenase (LDH) release assay reports how much cytoplasmic enzyme has leaked into the medium. Under ordinary conditions all three track viable cell number closely enough that the distinction never surfaces. Under the non-ordinary conditions that make up most interesting experiments — metabolic inhibitors, redox-active compounds, cell-cycle arrest, differentiation, hypoxia, 3D culture — they diverge, sometimes dramatically.

A well-documented example: when a panel of anti-cancer agents was profiled by direct image-based cell counting alongside ATP-bioluminescence and MTS tetrazolium readouts, the metabolism-based assays frequently underestimated compound potency and efficacy, and in some cases produced non-monotonic dose–response curves — signal rising with dose before falling — because of concentration-dependent phenotypic switching rather than any change in cell number.^[1] The assay was working correctly. It was reporting the metabolic state of the well, which was not the same thing as the number of cells in it.

Cell viability assay principles: four measurable proxies

Almost every plate-based cell viability assay in routine use reads one of four signals. Knowing which one your kit reads tells you immediately what it will miss.

1. Reductive / metabolic capacity

Living cells maintain a pool of reduced pyridine nucleotides (NADH, NADPH) and a set of dehydrogenases that will transfer electrons to an artificial acceptor. Add a tetrazolium salt or a redox dye, and the amount converted per unit time scales with how much active reducing machinery the well contains. This family covers MTT, MTS, XTT, WST-1, WST-8 — the chemistry behind CCK-8 and resazurin.

2. Energy charge (ATP)

ATP is consumed continuously and cannot be stockpiled, so intracellular ATP collapses within minutes of the loss of membrane integrity and metabolic function. Quantifying it by luciferase-catalyzed oxidation of luciferin gives the most sensitive of the common formats, resolving cell numbers in the tens. The same chemistry is also sold as a standalone ATP quantitation assay for when the endpoint you want is ATP concentration itself rather than a viability percentage.

3. Membrane integrity — the inverse marker

Rather than counting the living, measure the dead. LDH is a stable cytosolic enzyme retained by intact cells and released the moment the plasma membrane fails, so extracellular LDH activity is a widely accepted surrogate for membrane damage.^[2] Because it is measured in the supernatant, it can be sampled without sacrificing the plate.

4. DNA synthesis — proliferation, strictly speaking

A thymidine analogue such as BrdU is incorporated only into DNA being newly replicated, so the signal counts cells that divided during the labelling window rather than cells that are merely present and alive.

A fifth family sits alongside these: mechanism markers. Caspase-3/7 activity and mitochondrial membrane potential do not estimate viable cell number at all — they tell you how cells are dying. They are the natural follow-on once a cell viability assay has established that they are.

Types of cell viability assays: the chemistry behind each readout

The mechanisms differ in ways that matter operationally — whether the dye enters the cell, whether the product needs solubilizing, whether the assay can be read repeatedly, and which compound classes will corrupt it.

MTT — the original tetrazolium

MTT carries a net positive charge, which drives its uptake across the plasma membrane along the membrane potential. Once inside, it is reduced — principally by NADH — to a purple formazan.^[3] Reduction is not confined to mitochondria: it also occurs in the cytoplasm and at non-mitochondrial membranes including the endosomal/lysosomal compartment and the plasma membrane itself.^[3] The formazan is poorly water-soluble and precipitates as crystals, so the protocol requires a solubilization step before reading absorbance at ~570 nm.

WST-8 / CCK-8 — water-soluble tetrazolium

Second-generation tetrazoliums (XTT, WST-1, WST-8, and to some extent MTS) were engineered to yield a water-soluble formazan, removing the solubilization step entirely — you add reagent, incubate, and read. WST-8 reduction produces an orange-yellow formazan read at 450 nm, and the reagent’s low cytotoxicity means the same plate can be read repeatedly at different incubation times to find the optimal window.

Worth understanding: these dyes carry a net negative charge and are therefore largely cell-impermeable. The evidence indicates their reduction occurs at the cell surface or across the plasma membrane via trans-plasma-membrane electron transport, mediated by an intermediate electron acceptor, rather than inside mitochondria.^[3] The practical consequence is that WST-8 signal reflects cell-surface and trans-membrane electron flux as much as mitochondrial respiration — relevant if your treatment specifically targets oxidative phosphorylation.

Resazurin — fluorescent redox indicator

Resazurin is a non-fluorescent blue dye that metabolically active cells reduce to resorufin, which is strongly fluorescent (excitation ~530–570 nm, emission ~590–620 nm). Non-viable cells have no reductive capacity and leave the dye unconverted, so fluorescence intensity tracks the viable population. The reagent is non-toxic and homogeneous — add and read, no wash or transfer — and the fluorescent readout buys roughly an order of magnitude in sensitivity over MTT, resolving as few as 100 cells.

ATP bioluminescence — the sensitivity ceiling

A single working reagent lyses the cells to release ATP, which in the presence of luciferase drives oxidation of D-luciferin to emit light. Because ATP turns over in seconds and is not stored, the signal falls essentially immediately on loss of viability — and because bioluminescence has almost no background, detection reaches the tens of cells. Signal can be acquired within minutes of reagent addition, and the format suits both small experiments and high-throughput screens in 96- and 384-well plates.

The trade-off is that lysis makes the assay terminal: no time courses on the same well, and no multiplexing with a downstream live-cell readout. Because ATP per cell is itself a regulated quantity, treatments that alter energy metabolism or cell size will move the signal independently of cell number.^[1]

ADP/ATP ratio — normalizing the energy readout

This is the answer to the main weakness of a plain ATP cell viability assay. Because a ratio divides one nucleotide pool by another, it is far less sensitive to how many cells are in the well — which means it reports on cell state rather than cell number, and does so without needing an accurate seeding count.

The measurement is two-step in one well. The working reagent lyses the cells, and released ATP drives the luciferin reaction for a first luminescence reading. An enzymatic step then converts ADP to ATP, and a second reading captures the combined pool — ADP is recovered by difference. The interpretive logic is directional: raised ATP with lowered ADP indicates proliferating cells, while lowered ATP with raised ADP indicates cells that are dying. Per the supplier, the shift is considerably more pronounced in necrosis than in apoptosis, which gives the assay some power to separate the two modes.

ATP in 3D — spheroids and organoids

Three-dimensional cultures break the assumptions behind standard cell viability assay chemistry. Dye penetration into a spheroid is diffusion-limited, cells at the core are metabolically distinct from those at the periphery, and — critically — lysis reagents formulated for monolayers do not fully disrupt a dense aggregate, so ATP release is incomplete and viability is underestimated. Reagents intended for 3D work address this with formulations optimized for aggregate disruption.

One such reagent has been characterized by the supplier in HeLa spheroids grown 4–6 days in low-attachment plates, and applied across ovarian, lung, and thyroid cancer organoid models, retaining over 90% activity after 60 days at 4 °C.

LDH release — measuring the dead

LDH interconverts lactate and pyruvate, and stays inside the cell as long as the plasma membrane holds. Cytotoxic compounds that induce necrosis or late apoptosis compromise membrane integrity and release it, so LDH activity in the culture supernatant reports the extent of lytic damage. Detection is coupled to reduction of a tetrazolium salt to a formazan dye, read colorimetrically. Because the sample is medium rather than cells, an LDH plate can be read alongside a metabolic assay on the same wells.

This orthogonality is the main reason to run it. When a tetrazolium curve and an LDH curve disagree, the disagreement itself is informative: metabolic signal down with no LDH release points to cytostasis or metabolic suppression, not killing.

BrdU incorporation — a cell proliferation assay, not a viability assay

Bromodeoxyuridine is a thymidine analogue incorporated into newly synthesized DNA only in cells actively replicating, so colour development is proportional to incorporated BrdU and reports proliferation directly — independent of how many cells were seeded. Detection sensitivity extends to as few as 50–100 proliferating cells, in adherent or suspension culture.

Use it when the biological claim is about growth rate rather than survival. Pairing a BrdU cell proliferation assay with a metabolic cell viability assay is the cleanest way to distinguish an antiproliferative compound from a cytotoxic one.

Mechanism markers — caspase activity and membrane potential

Once a cell viability assay establishes that cells are dying, these establish how. Caspase-3 is the principal executioner protease of apoptosis, and its activity — measured with peptide substrates generating a fluorescent or colorimetric signal — distinguishes programmed death from necrosis. Mitochondrial membrane potential probes such as JC-1 report depolarization, an early apoptotic event that precedes membrane rupture and therefore precedes any LDH signal. Browse the full apoptosis & cell death range.

Cell viability assay selection: situation → recommended readout

A shortcut from the experiment you actually have to the chemistry that answers it.

Cell viability assay applications

Compound cytotoxicity and IC50 determination

The dominant application. Cells are plated, allowed to attach, treated across a concentration series, and read after a fixed exposure. The output is a sigmoidal curve and a potency value. Two design choices dominate quality: the exposure duration (a 24 h and a 72 h read on the same compound can differ by an order of magnitude in apparent potency, because slow-acting mechanisms need time to manifest) and the seeding density (too dense and controls saturate the assay’s linear range before treatment even begins).

Culture and media optimization

Cell viability assays are the standard readout for optimizing serum concentration, supplement combinations, coating substrates, passage conditions, and cryopreservation recovery. Here the metabolic proxies are well suited, because the question genuinely is about metabolic health rather than a body count.

Vector and transfection tolerance

When titrating AAV or lentiviral MOI, testing a transfection reagent, or optimizing electroporation, a viability plate run in parallel with the efficiency readout prevents the classic error of reporting high transduction efficiency in a population that has been substantially killed. Membrane-integrity readouts are particularly appropriate for physical delivery methods, where transient permeabilization is the mechanism.

Toxicology and environmental screening

Nanoparticles, industrial chemicals, and environmental samples are routinely screened for cytotoxic potential in cell culture. This is the application area where interference is most severe: particulates scatter light and adsorb dye, and complex extracts commonly contain reducing species. Orthogonal confirmation is not optional here.

Antimicrobial and drug-sensitivity testing

Resazurin reduction is a long-established readout for microbial susceptibility as well as mammalian cytotoxicity, and its fluorescent output supports the plate densities that susceptibility panels require.

3D models and organoid drug response

As spheroid and organoid models displace monolayers for compound profiling, viability measurement becomes the rate-limiting step, because the geometry breaks assumptions built into monolayer chemistry. Reagents formulated for aggregate lysis and validated in the specific model type are worth the premium; a monolayer protocol applied to a spheroid systematically underestimates viable cells.

Related BioHippo collections: Cell Cycle & Proliferation · Apoptosis & Cell Death · Drug Development Assay Kits · Oxidative Stress & Antioxidants

Where cell viability assays go wrong: six failure modes

These are the errors that produce publishable-looking curves carrying the wrong conclusion. Each is well documented, and each has a specific control that catches it.

1. The compound reduces the dye directly

Tetrazolium and resazurin chemistries detect reducing capacity. Any test article that is itself a reducing agent will generate signal without any cellular involvement at all. This was demonstrated cleanly when kaempferol was tested on breast cancer cells: cell number fell by crystal violet staining while MTT formazan formation increased. Direct reductive potential was confirmed in a cell-free system for kaempferol and resveratrol and for extracts of Hypericum perforatum and Cimicifuga racemosa, all producing instantaneous formazan in the absence of cells — and ascorbic acid, vitamin E, and N-acetylcysteine interfered similarly.^[5] The failure is directional and dangerous: it makes a cytotoxic compound look protective.

2. Metabolic state changes without cell number changing

The proxy assumes a fixed signal-per-cell. Treatments that alter metabolism break that assumption. Compared against direct image-based counting, ATP-bioluminescence and MTS readouts frequently underestimated potency and efficacy, and generated non-monotonic dose–response curves through concentration-dependent phenotypic switching.^[1] Discrepancies between metabolic activity and DNA content have been documented specifically as a limitation when using metabolic assays as proliferation proxies.^[6]

3. Cytostasis read as cytotoxicity

A compound that halts division produces fewer cells at endpoint than the growing control, so a metabolic assay reports reduced “viability” — but nothing died. The distinction matters enormously for mechanism and for therapeutic interpretation.

4. Seeding density and incubation time drive the answer

Every one of these chemistries has a finite linear range. Seed too densely and untreated controls plateau, compressing the apparent effect; seed too sparsely and treated wells fall below detection. Incubation time compounds this — the CCK-8 window is typically 1–4 h and needs empirical determination per cell type, with leukocytes generally requiring longer.

5. Viable but not proliferating — and the reverse

Quiescent, senescent, and terminally differentiated cells are alive and metabolically active but do not divide, so proliferation assays score them low. Conversely, early apoptotic cells retain membrane integrity and metabolic activity for hours and will register as fully viable by both metabolic and membrane-integrity readouts.

6. Plate artefacts

Edge-well evaporation concentrates medium and skews both metabolic and absorbance readings across a plate. Bubbles introduced during reagent addition corrupt optical density readings directly. Coloured or fluorescent test compounds contribute background at the detection wavelength. Particulates scatter light.

Controls that make a cell viability assay number mean something

A viability percentage is only interpretable relative to defined anchors. The minimum set:

• Reagent blank — medium plus detection reagent, no cells. Establishes the zero and the reagent’s own background.
• Untreated (vehicle) control — cells plus the exact vehicle concentration used in treated wells. This is the 100% anchor. DMSO alone is cytotoxic above roughly 0.5–1% in many lines, so vehicle must be matched, not omitted.
• Maximum-kill control — a defined lytic treatment establishing the floor. A detergent-type cytotoxic control such as saponin serves this role and provides the 100%-lysis reference that an LDH assay requires to express release as a percentage.
• Cell-free compound series — the interference control from failure mode 1. Cheap, and the only way to catch direct dye reduction.
• Cell-number standard curve — run once per cell line to establish the linear range and convert signal to cell number rather than relative percentage.
• Replicate structure — technical replicates within a plate quantify dispensing variability; independent biological replicates on separate days quantify what actually matters. Do not substitute the former for the latter.

The standard convention is to express treated-well signal as a percentage of vehicle control after blank subtraction: viability (%) = (A_treated − A_blank) / (A_control − A_blank) × 100, with inhibition as the complement. Reporting raw absorbance without blank subtraction is a common and avoidable error. For a broader treatment of assay controls and formats, the NIH Assay Guidance Manual chapters on cell viability^[7] and cytotoxicity^[2] are the standard open reference.

Cell viability assay FAQ

Is CCK-8 simply a better MTT?

For workflow, largely yes — WST-8 yields a water-soluble formazan, so there is no solubilization step, the reagent’s low cytotoxicity allows the same plate to be re-read, and phenol red and serum do not interfere. But the chemistries are not mechanistically interchangeable. MTT is cell-permeable and reduced intracellularly; WST-8 carries a net negative charge, is largely cell-impermeable, and is reduced at or near the plasma membrane via trans-membrane electron transport.^[3] If your treatment specifically targets mitochondrial function, that difference can change what the signal means. Both remain vulnerable to direct reduction by redox-active compounds.

Why do my MTT and ATP cell viability assay results disagree?

Because they measure different things, and disagreement is diagnostic rather than a failure. The two proxies uncouple when average metabolic activity or ATP content per cell shifts — under metabolic inhibitors, cell-cycle arrest, hypertrophy, or a switch between glycolysis and oxidative phosphorylation. Direct counting, DNA content, or LDH release will tell you which proxy has drifted. Documented discrepancies between metabolic proxies and actual cell number are well characterized in the literature.^[1,6]

What is the difference between a cell viability assay and a cytotoxicity assay?

A viability assay estimates how much living material remains, usually via a metabolic proxy, and reports it relative to an untreated control. A cytotoxicity assay measures damage directly — most commonly LDH released from cells whose plasma membrane has failed. The practical difference is that a metabolic viability assay cannot separate growth inhibition from cell killing, because both lower the endpoint signal, whereas LDH rises only when cells lyse. If the distinction between cytostatic and cytotoxic matters to your conclusion, you need both.

Which cell viability assay should I use for a natural product or plant extract?

Not a tetrazolium or resazurin assay alone. Polyphenols, flavonoids, and common antioxidants reduce these dyes directly, producing signal with no cells present and making cytotoxic material appear protective.^[5] Use ATP bioluminescence or LDH release as the primary readout, and always run the concentration series in cell-free wells to quantify chemical background.

Can I run a time course on the same plate?

With non-lytic reagents, yes. WST-8 and resazurin leave cells intact, and re-reading the same plate at intervals is standard practice for finding the optimal incubation window. LDH is sampled from the supernatant, so serial sampling also works. ATP assays lyse the cells and are strictly endpoint — for a luminescent time course you need one plate per time point.

How many cells should I plate for a cell viability assay?

Enough that untreated controls sit in the assay’s linear range without saturating it. For a WST-8 assay in a standard 96-well plate, at least 1,000 cells/well in 100 µL is a reasonable starting point, with leukocytes needing roughly 2,500 cells/well because their detection sensitivity is lower. Resazurin resolves down to about 100 cells and ATP luminescence to about 50, so both tolerate sparser plating. Determine it empirically with a standard curve for your line rather than importing a number from another paper.

Do I need a separate cytotoxicity assay if I already have a viability assay?

If the distinction between growth inhibition and cell killing matters to your conclusion, yes. A viability assay measuring metabolic activity cannot separate the two: both reduce endpoint signal. Adding LDH release resolves it, because arrest reduces metabolic signal without releasing LDH. Since LDH is measured in the supernatant, it can usually be added to existing wells rather than requiring a parallel plate.

What changes when I move to spheroids or organoids?

Three things. Dye penetration becomes diffusion-limited, so peripheral cells contribute disproportionately. Cells at the aggregate core are metabolically distinct from those at the surface. And lysis reagents formulated for monolayers do not fully disrupt a dense aggregate, so ATP release is incomplete and viability reads low. Use a reagent explicitly formulated and validated for 3D lysis, and treat absolute values across 2D and 3D as non-comparable.

Cell viability and cytotoxicity assay kits at BioHippo

The kits referenced throughout this guide, organized by the proxy each one measures. Browse the full assay kits range.

The short version: pick the proxy your treatment cannot corrupt

Cell viability assays are among the most reliable measurements in cell biology and among the most frequently over-interpreted. The number they produce is real; the assumption that it equals “percentage of cells alive” is the part that fails. Choose the chemistry whose proxy stays coupled to cell number under your specific treatment, anchor it with a cell-free interference control and a defined maximum-lysis reference, and confirm anything load-bearing with a mechanistically unrelated readout. Two assays that agree are worth far more than one assay with a tight error bar.

References

1. Chan GKY, Kleinheinz TL, Peterson D, Moffat JG. A simple high-content cell cycle assay reveals frequent discrepancies between cell number and ATP and MTS proliferation assays. PLoS One. 2013;8(5):e63583. PMID 23691072 · doi:10.1371/journal.pone.0063583
2. Riss TL, Niles AL, Moravec RA, Karassina N, Vidugiriene J. Cytotoxicity assays: in vitro methods to measure dead cells. In: Assay Guidance Manual. Bethesda (MD): Eli Lilly & Company and NCATS; 2019. NCBI Bookshelf NBK540958
3. Berridge MV, Herst PM, Tan AS. Tetrazolium dyes as tools in cell biology: new insights into their cellular reduction. Biotechnol Annu Rev. 2005;11:127–152. PMID 16216776 · doi:10.1016/S1387-2656(05)11004-7
4. Mosmann T. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J Immunol Methods. 1983;65(1–2):55–63. PMID 6606682 · doi:10.1016/0022-1759(83)90303-4
5. Bruggisser R, von Daeniken K, Jundt G, Schaffner W, Tullberg-Reinert H. Interference of plant extracts, phytoestrogens and antioxidants with the MTT tetrazolium assay. Planta Med. 2002;68(5):445–448. PMID 12058323 · doi:10.1055/s-2002-32073
6. Quent VMC, Loessner D, Friis T, Reichert JC, Hutmacher DW. Discrepancies between metabolic activity and DNA content as tool to assess cell proliferation in cancer research. J Cell Mol Med. 2010;14(4):1003–1013. PMID 20082656 · doi:10.1111/j.1582-4934.2010.01013.x
7. Riss TL, Moravec RA, Niles AL, et al. Cell viability assays. In: Assay Guidance Manual. Bethesda (MD): Eli Lilly & Company and NCATS; 2013, updated 2016. NCBI Bookshelf NBK144065

Product specifications, pack sizes, and availability are stated on each product page and supersede any figure quoted here. All products are for research use only; not for use in diagnostic or therapeutic procedures. Questions on assay selection: talk to a specialist.