Binding Affinity

Affinity is the ratio of how fast a ligand leaves its target to how fast it arrives: the dissociation constant equals the off-rate divided by the on-rate. Because that ratio has units of concentration, a smaller number means tighter binding, and the value itself is the free concentration at which half the binding sites are filled at equilibrium. It is measured by saturation or competition experiments with a labelled ligand, or directly by surface plasmon resonance, which returns the on- and off-rates separately. A competition result is reported as an inhibition constant after correcting for the tracer used.

The usable range spans orders of magnitude. Typical small molecules sit in the micromolar to nanomolar band, therapeutic peptides in the nanomolar to sub-nanomolar band, and monoclonal antibodies often reach picomolar. Affinity does not rank clinical usefulness, though: tirzepatide binds the GIP receptor with affinity close to that of native GIP but the GLP-1 receptor appreciably more weakly than native GLP-1, and is nonetheless among the most effective agents in its class.

Affinity earns its keep when paired with a concentration. Combined with the free, unbound plasma level, it tells you what fraction of target is engaged at a given dose, which converts a pharmacokinetic curve into a statement about pharmacology. Compared across two targets, the ratio of affinities is the starting point for any selectivity claim.

The recurring mistake is comparing affinities that were never comparable, across different labs, buffers, temperatures, membrane preparations or species orthologs, where two-fold to ten-fold discrepancies for the same pair are routine. A second is treating affinity as activity: a molecule can bind superbly and do nothing, or block. A listing that says high affinity without naming a receptor, a number and an assay has said nothing at all.