Cell binding by flow cytometry
Recombinant antigen binding tells you a molecule works against a protein. Cell binding tells you it works against the protein as the cell presents it: correctly folded, correctly modified, at native density, surrounded by everything else on that membrane. For a program that will act on cells, it is the measurement that counts.
The basic experiment
Live cells expressing the target are incubated with a titration of the antibody, washed, stained with a labeled secondary or with directly conjugated antibody, and read on a cytometer. Median fluorescence intensity against concentration gives a binding curve.
Three design choices carry most of the risk.
Cell line pair. Run a target-positive line and a matched negative line, ideally the parental line and a transfected derivative. Without the negative, non-specific binding is invisible.
Temperature. Bind on ice or at four degrees to measure surface binding. At thirty-seven degrees, receptors internalize, and what you measure becomes a mixture of binding and uptake.
Viability. Dead cells bind almost anything. A viability dye in every tube, and gating that excludes dead cells, removes the single largest source of false positives.
Controls that earn their place
- An isotype or irrelevant binder matched in format and concentration, because a bivalent IgG and a VHH have different non-specific behavior.
- Secondary-only wells, to catch the secondary binding the cells directly.
- Fc receptor blocking where the cells express Fc receptors, since Fc-mediated binding looks exactly like specific binding.
- A reference binder with known behavior on this line, run every time, so day-to-day drift is visible.
- Compensation and single stains whenever more than one fluorophore is in the panel.
Reading the result
The useful outputs are the curve and the window. The midpoint gives an apparent affinity under these conditions, which is not the intrinsic KD: receptor density and bivalency both shift it, and a bivalent molecule binding a dense receptor will look much tighter than it is.
The maximum signal matters too. Two antibodies can share a midpoint and differ severalfold in maximum binding, which usually means they see different fractions of the receptor population, for example a subset in a particular conformation or glycoform.
Report the fluorescence as median rather than mean, and show the distribution when it is not unimodal. A bimodal population says something the summary statistic hides.
Common traps
Aggregated antibody produces high, saturating, concentration-independent staining that looks like strong binding. Check the material before blaming the biology.
Overexpression lines flatter weak binders. A binder selected on a line with artificially high receptor density can fail on cells with physiological levels, so confirm on a line that expresses the target naturally before any decision.
Washing conditions change fast off-rate binders dramatically. Keep them constant across a panel and be aware that a rapid off-rate can be washed away entirely.