Binding Affinity Explained: Kd, Ki, IC50 and EC50

Pharmacology papers report a dense set of constants, and K_d, K_i, IC₅₀ and EC₅₀ are easy to confuse. They are not interchangeable. Binding affinity describes how tightly a ligand holds its receptor; potency describes how much ligand is needed to produce an effect in a particular assay. Mixing the two is one of the most common errors in reading peptide literature, and it leads to comparisons that the underlying data cannot support.

This guide defines each term, explains how the measurements are made, and sets out the pitfalls that matter most when peptides are the ligands. Multi-receptor peptides, such as TIRZ-2, a tirzepatide-class research compound that acts at both the GIP and GLP-1 receptors, are used as an example of why careful definitions are necessary.

Affinity versus potency

Affinity is a property of the ligand-receptor pair. At equilibrium, a ligand binds and dissociates at rates set by its association rate constant (k_on) and dissociation rate constant (k_off). The ratio k_off/k_on is the equilibrium dissociation constant, K_d: the free ligand concentration at which half the receptors are occupied. A lower K_d means higher affinity.

Potency is a property of the ligand in a system. It depends on affinity, but also on efficacy (how well the bound ligand activates the receptor), on the number of receptors present, and on how strongly the downstream pathway amplifies the signal. Two peptides with identical affinity can differ widely in potency, and the same peptide can show different potencies in different cell lines. A related term, efficacy, refers to the maximal response a ligand can produce, independent of how much is needed to reach it.

Kd, Ki, IC50 and EC50 defined

Term	What it measures	Typical source	Depends on assay conditions?
Kd	Equilibrium dissociation constant of a ligand	Saturation binding, kinetic binding, SPR, ITC	In principle a constant, but sensitive to buffer, temperature and receptor construct
Ki	Affinity of an unlabelled competitor, derived from IC50	Competition binding	Corrected for tracer concentration and affinity
IC50	Concentration giving 50% inhibition of a signal or of tracer binding	Competition binding or functional inhibition	Yes, strongly
EC50	Concentration giving 50% of the ligand’s own maximal response	Functional assays (cAMP, calcium, reporter)	Yes, strongly

K_i deserves particular attention. In a competition experiment, an unlabelled peptide displaces a labelled tracer, giving an IC₅₀. That IC₅₀ rises if more tracer is used or if the tracer binds more tightly, so it is not a fixed property of the competitor. The Cheng-Prusoff equation, K_i = IC₅₀ / (1 + [L]/K_d), where [L] is the tracer concentration and K_d is the tracer’s affinity, corrects for this under the assumptions of simple competitive binding at equilibrium. Values are often reported on a log scale as pK_i or pEC₅₀ (the negative log of the molar value), which makes averaging and comparison more statistically sound.

Comparing potency across receptors

Multi-receptor peptides show why these distinctions matter. A tirzepatide-class dual GIP/GLP-1 receptor agonist is designed to activate two related class B1 receptors, and its pharmacology is described by a separate potency and efficacy at each. A statement that such a compound is “balanced” or “biased toward” one receptor is a comparison of two EC₅₀ values, usually each expressed relative to that receptor’s native ligand. That relative framing is important: the GLP-1 receptor and the GIP receptor are different proteins in different cell lines, so a raw EC₅₀ at one is not directly comparable with a raw EC₅₀ at the other.

The problem grows with a third receptor. For a GIP/GLP-1/glucagon triple agonist such as the retatrutide peptide, a potency profile needs three assays, ideally run in matched cell backgrounds with similar receptor expression, the same readout, and the same incubation time. Differences in receptor density between cell lines can shift EC₅₀ values by more than the differences between ligands. Published profiles also vary with species: human and rodent receptor orthologues can respond differently to the same peptide, so a ratio measured at human receptors in vitro may not hold in an animal model.

Measurement methods

Radioligand and fluorescent binding

Radioligand binding remains a reference method for receptors in membranes or whole cells. Saturation experiments, using increasing concentrations of labelled ligand, yield K_d and receptor density (B_max). Nonspecific binding is measured in parallel with an excess of unlabelled ligand and subtracted. Competition experiments then give IC₅₀ values for unlabelled compounds. Fluorescent tracers read by fluorescence polarisation or time-resolved FRET avoid radioactivity and allow homogeneous formats, at the cost of a bulkier label that can alter peptide binding.

Label-free biophysical methods

Surface plasmon resonance (SPR) and bio-layer interferometry measure binding in real time on an immobilised partner, reporting k_on and k_off as well as K_d. Isothermal titration calorimetry (ITC) measures heat released or absorbed on binding, yielding K_d, stoichiometry and enthalpy in solution without labels. These methods generally require purified protein, which is straightforward for soluble domains but difficult for intact GPCRs, so peptide studies sometimes use an isolated extracellular domain as a surrogate, with the caveat that it does not capture the full receptor.

Functional assays

EC₅₀ values come from concentration-response curves in functional assays such as cAMP accumulation, calcium flux, β-arrestin recruitment or reporter-gene expression. Each readout sits at a different point in the signaling cascade, and more distal readouts usually show greater amplification and therefore lower EC₅₀ values. Our article on peptide receptor binding describes the receptor biology behind these assays.

Common pitfalls

Several errors recur in peptide binding work. Ligand depletion occurs when receptor concentration approaches or exceeds K_d, so that bound ligand significantly reduces free ligand and the apparent affinity is underestimated. Non-equilibrium conditions are another: high-affinity peptides can dissociate very slowly, and short incubations then report IC₅₀ values that reflect kinetics rather than affinity.

Peptide-specific issues add to these. Adsorption to plastic lowers effective concentration at low nanomolar levels unless carrier protein or low-binding plasticware is used. For lipidated peptides, however, albumin in the buffer binds the fatty acid and reduces free peptide, so EC₅₀ values measured with and without serum albumin can differ substantially. Any comparison involving acylated dual agonists such as TIRZ-2 or other lipidated analogues should state the carrier protein and its concentration. Finally, purity and identity of the material set an upper limit on the quality of the data; a truncated impurity can behave as an antagonist and distort a curve. Guidance on reading reported values critically is in our guide on how to read a research paper.

Using these constants well

K_d and K_i describe binding; IC₅₀ and EC₅₀ describe an experiment. Treat the latter as system-dependent, report assay conditions alongside every value, and compare ligands only within matched assays and against shared reference agonists. These constants are in-vitro measures of molecular behaviour. They say nothing on their own about outcomes in whole organisms, which depend on pharmacokinetics, tissue distribution and physiology that a binding assay does not capture.

References

1. Cheng, Y. and Prusoff, W. H. Relationship between the inhibition constant (K_I) and the concentration of inhibitor which causes 50 per cent inhibition (I₅₀) of an enzymatic reaction. Biochemical Pharmacology, 1973.
2. Neubig, R. R. et al. International Union of Pharmacology Committee on Receptor Nomenclature and Drug Classification. XXXVIII. Update on terms and symbols in quantitative pharmacology. Pharmacological Reviews, 2003.
3. Motulsky, H. and Christopoulos, A. Fitting Models to Biological Data Using Linear and Nonlinear Regression. Oxford University Press, 2004.
4. Rang, H. P., Ritter, J. M., Flower, R. J. and Henderson, G. Rang and Dale’s Pharmacology. Elsevier.
5. IUPHAR/BPS Guide to Pharmacology. International Union of Basic and Clinical Pharmacology and British Pharmacological Society. Online database.