Cytotoxicity Assays for IC50, Selectivity and Oncology by PhDs at 20+ Year, FDA-Inspected Lab

The mechanism decides. Metabolic readouts like MTT and resazurin report dying metabolism, ATP luminescence reports energy collapse, LDH release reports membrane rupture, and imaging shows morphology and kinetics. A compound that halts proliferation without killing reads very differently across those, which is exactly how misreads happen. We match the readout to your expected mechanism and confirm compound interference with the detection chemistry before any panel runs.

Either. More than 200 cancer cell lines are held in house, authenticated and banked, so most oncology panels start without a sourcing delay. If your model is proprietary or engineered, send it with its history and we qualify it into the assay. Matched normal lines run in the same experiment where selectivity is the question.

It is the ratio between the concentration that kills the tumor line and the concentration that harms a matched normal line, measured in the same experiment, same passage, same plate design. A potent compound with no selectivity window is a toxin, not a lead. Reporting both numbers side by side is how the killing claim survives its first skeptical reader.

Both exist and they answer different questions. IC50 reports the concentration at half-maximal effect against the measured readout; GI50 corrects for the starting cell count, so it separates growth inhibition from net killing. We report the parameter that fits your mechanism and say which convention was used, because a number without its definition invites the wrong comparison.

Co-culture at defined effector-to-target ratios: your CAR-T, TCR-T or NK product against target-positive and target-negative lines, with killing read by the format that fits, and granzyme or cytokine readouts alongside where the program wants mechanism next to potency. The same design carries from characterization into potency contexts.

Yes. Checkerboard or fixed-ratio designs with Bliss or Loewe analysis distinguish additivity from synergy from antagonism, which is the question partners and reviewers actually ask of a combination claim. Design support is part of the work, since a badly designed checkerboard proves nothing at twice the cost.

Usually permeability, efflux or serum binding, and occasionally the target biology itself. The disagreement is information: it says the molecule engages the purified target but struggles to reach or stay at it in a living system. We run the cellular and biochemical numbers side by side and, when the gap appears, say which ADME question to ask next rather than leaving two numbers to argue.

Yes. Rat and human primary hepatocyte cytotoxicity gives an early liver-liability read at the stage where it changes decisions, before a tox study is bought. It is a screen, not a substitute for GLP toxicology, and we frame the data accordingly.