ELISA for Peptide Quantification and Its Limits

An immunoassay returns a concentration, and the number looks like a measurement of how much peptide is present. It is a measurement of how much material in the sample was recognised by two antibodies, which is a different quantity whenever anything in the sample resembles the target closely enough to be mistaken for it.

What the format measures

In the common sandwich arrangement, one antibody immobilised on a plate captures the analyte, a second labelled antibody binds a different part of it, and the label generates a signal proportional to how much was caught between them. A standard curve made from known concentrations converts signal to concentration.

The requirement for two independent binding sites is what makes the format specific — and what makes it poorly suited to short peptides, which may not present two epitopes far enough apart for both antibodies to bind at once. For small analytes a competitive format is used instead, in which the sample competes with a labelled version for a limited quantity of antibody, and signal falls as concentration rises.

The cross-reactivity problem is severe for peptides

An antibody raised against a peptide recognises a stretch of that sequence. Anything sharing that stretch is a candidate for detection.

Three consequences follow, and all three are routine in this material. Degradation products retaining the epitope are detected as though they were intact peptide, so an assay may report full concentration in a sample that has substantially degraded. Analogs differing by a residue or a terminal modification may be detected equally well, which is exactly the situation for closely related sequences of the kind compared in three compounds one dalton apart. And a fragment shorter than the target can be fully reactive.

The assay therefore measures immunoreactive material, not intact peptide. Where the distinction matters — and for a stability question it always does — a separation-based method is required, because chromatography resolves what an antibody cannot.

The standard curve is an assumption

Concentrations are read against a curve built from a reference preparation. Everything about the reported number depends on that preparation being what it says it is.

If the standard is the same material being measured, the assay reports a ratio and inherits the standard’s own content uncertainty — the net content problem in why a 10mg vial is not 10mg of peptide. A standard weighed out without accounting for water and counter-ion content is systematically high, and every sample read against it is systematically low by the same proportion.

A curve is also only valid across the range it covers. Readings extrapolated beyond the highest standard are not measurements, and in a competitive format the curve is steepest in the middle and nearly flat at both ends, so samples falling at either extreme carry very large uncertainty.

Matrix effects

The sample is not buffer. Medium, serum, and anything else present can interfere with antibody binding, block the plate, or contribute background.

The standard control is a spike-and-recovery experiment: add a known quantity of analyte to the actual sample matrix and check that the assay returns it. Recovery far from complete means the matrix is interfering and the numbers need a correction or a different approach. A parallel check is linearity of dilution — a sample diluted twofold should read half. Where it does not, something in the matrix is affecting the result concentration-dependently.

Two failure modes worth naming

• The high-dose hook. In a sandwich format, a very high analyte concentration saturates both antibodies separately and prevents the sandwich forming, so signal falls. A grossly concentrated sample can read as low. Running two dilutions catches it; running one does not.
• Adsorptive loss before the plate. The sample was prepared, diluted and transferred before it arrived, and peptides are lost to surfaces along the way, per adsorptive loss. The assay faithfully measures what reached it.

Where it is the right tool

Immunoassays are sensitive, work at concentrations well below what chromatography reaches, tolerate complex samples, and scale to many samples at once. For detecting a known analyte at low concentration in a complex matrix, nothing else is as practical.

They are the wrong tool for establishing identity, for assessing purity, and for any question where degradation products must be distinguished from intact material — all of which belong to the methods in HPLC and mass spectrometry.

Why a peptide antibody is harder to make than it sounds

Short peptides are poorly immunogenic on their own and are conjugated to a carrier protein to raise antibodies at all. The conjugation chemistry attaches the peptide at one end, so the antibodies raised recognise the exposed portion and may be blind to the region used for attachment.

That has a direct consequence for what the assay sees. An antibody raised against a peptide coupled through its C-terminus tends to recognise N-terminal features, so a C-terminal modification — an amidation, or the loss of a residue from that end — may be invisible to it. Where the modification is the thing that distinguishes two compounds, as with the terminal chemistry in acetylation and amidation, the assay cannot tell them apart at all.

Reading a reported concentration

What the standard was and how its content was established. Whether the antibody’s cross-reactivity against likely degradation products and related analogs was characterised, and at what percentage. Whether spike-recovery and dilution linearity were checked in the actual matrix. And whether samples fell within the calibrated range.

A concentration reported without those is a signal converted by an uncharacterised curve, and it is worth treating as an order of magnitude rather than a figure.