Measuring protein concentration by sandwich ELISA

An enzyme-linked immunosorbent assay (ELISA) measures how much of one substance a liquid contains by letting antibodies bind it on the surface of a plastic plate and then making the bound amount visible with an enzyme that produces colour. The method was introduced by Engvall and Perlmann to measure immunoglobulin G (Engvall & Perlmann, 1971). In a sandwich ELISA, a capture antibody fixed to the bottom of each well holds the protein, and a second, detection antibody binds a different part of the same protein. Only a protein held by both antibodies gives a signal, which makes the assay specific and sensitive.

The enzyme most often used is horseradish peroxidase (HRP). With hydrogen peroxide it oxidises the colourless substrate TMB to a blue product; adding acid stops the reaction and gives a yellow product that is stable at acid pH (Josephy et al., 1982); it is read at 450 nm, with a second reading at 570 or 630 nm subtracted to remove the absorbance of the plate itself. The colour of each sample is converted into a concentration with a standard curve, made from known amounts of the pure protein. Because the curve is S-shaped, it is fitted with a 4-parameter logistic function, the accepted reference model for ligand-binding assays (Findlay & Dillard, 2007).

The numbers are only as good as the checks behind them. Substances in plasma, CSF or a tissue lysate can change antibody binding (a matrix effect), antibodies can bind related proteins (cross-reactivity), and very high concentrations can suppress the signal in one-step assays (the high-dose hook effect); all three can make a result wrong without making it look wrong (Ward et al., 2017). Spike recovery, dilution linearity, duplicate precision and limits of quantification are the experiments that show whether a method is fit for its purpose (Andreasson et al., 2015).

The protein measured is brain-derived neurotrophic factor (BDNF), a growth factor important for neurons and synapses. BDNF can be measured in brain extracts and human blood, but in human blood most of it is stored in platelets and released when they are activated, so serum contains far more than plasma (Fujimura et al., 2002; Want et al., 2023). Commercial BDNF assays differ in their precision and in whether they also detect the precursor, proBDNF (Polacchini et al., 2015). Unlike in a real laboratory, the true concentration in each sample is known in the simulation, and the final station shows how each decision affected the result.

Learning objectives

• Design an ELISA study with the animal or the person as the unit of analysis, enough units per group and coded samples
• Prepare plasma, CSF or tissue lysate so that the protein is preserved, and choose a dilution that brings the samples into the range of the standards
• Explain why a sandwich needs two antibodies against different parts of the protein, from different species when an anti-species secondary is used, and why a blocking step and thorough washing are needed
• Calculate and pipette a serial dilution of standards, and lay out a plate in duplicate with blanks, spike recovery and dilution-linearity wells
• Fit a 4-parameter logistic curve, back-calculate concentrations with the dilution factor, and apply the limit of detection and the limits of quantification
• Judge duplicate CVs, spike recovery and dilution linearity, recognise matrix and hook effects, and compare groups with a test suited to concentration data

Ethics and safety

Mouse tissue comes from animals bred and killed under a protocol approved by an Institutional Animal Ethics Committee (in India, the CCSEA under the Prevention of Cruelty to Animals Act 1960; elsewhere, frameworks such as EU Directive 2010/63/EU, the US Animal Welfare Act and PHS Policy, and the UK Animals (Scientific Procedures) Act 1986); group sizes are justified in advance and tissue from one set of animals can serve several assays (Reduction, one of the 3Rs). Studies are reported following the ARRIVE guidelines, developed by the NC3Rs in the United Kingdom and endorsed by journals internationally (Percie du Sert et al., 2020). Human blood and CSF are collected only with written informed consent, under a protocol approved by an ethics committee (in India, an Institutional Ethics Committee following the ICMR National Ethical Guidelines of 2017; elsewhere, a research ethics committee or, in the United States, an institutional review board), by trained clinical staff (CSF by lumbar puncture, a medical procedure); samples are coded so that the laboratory never sees names, and data are handled under data-protection law (in India, the Digital Personal Data Protection Act 2023; in the European Union, the General Data Protection Regulation). Human groups are defined by diagnosis, so nothing is manipulated and confounders such as age, sex and medication must be recorded. All human samples are handled as potentially infectious, with gloves, eye protection and approved disposal; the stop solution is a strong acid.

Key terms

• ELISA: Enzyme-linked immunosorbent assay: antibodies on a plastic plate bind a protein, and an enzyme turns the bound amount into colour.
• Sandwich ELISA: An ELISA in which the protein is held between a capture antibody on the plate and a detection antibody that binds a different site (epitope) on the same protein.
• Epitope: The small part of a protein that one antibody recognises. Two antibodies can sandwich a protein only if their epitopes do not overlap.
• Blocking: Coating the remaining free plastic with an unrelated protein (here bovine serum albumin) so that antibodies and enzyme do not stick to it.
• HRP and TMB: Horseradish peroxidase, the enzyme on the detection reagent, oxidises the substrate TMB to a blue product; acid stops the reaction and turns it yellow.
• Optical density (OD): Absorbance: how much light of one wavelength a well absorbs. 450 nm measures the yellow product; a reference reading at 570 or 630 nm measures only the plate itself and is subtracted.
• Standard curve: The signal of standards of known concentration, used to convert the signal of each sample into a concentration.
• 4-parameter logistic (4PL): An S-shaped curve with a bottom, a top, a midpoint and a slope, used to fit ELISA standards.
• Dilution factor: How many times a sample was diluted before it went into the well; the concentration read from the curve is multiplied by it.
• LOD, LLOQ and ULOQ: The limit of detection is the lowest concentration distinguishable from a blank; the lower and upper limits of quantification bound the range in which values are accurate and precise.
• Coefficient of variation (CV): The standard deviation as a percentage of the mean; for duplicate wells it shows how well they agree.
• Matrix effect: A change in the measured signal caused by the other contents of the sample (plasma proteins, lipids, detergents), not by the protein itself.
• Spike recovery: A known amount of pure protein is added to a sample; the fraction of it that the assay finds shows whether the matrix interferes.
• Hook effect: In a one-step sandwich assay, a very high concentration saturates both antibodies separately, so fewer sandwiches form and the signal falls, giving a falsely low result.