If you want to know whether a peptide survives contact with blood proteases, you do not need an animal. You need plasma, a heated block, a clock, and a way of measuring how much peptide is left. That experiment is the plasma stability assay, and it is one of the most informative in-vitro measurements available for a peptide.
It is also one of the easiest to run badly, and one of the easiest to over-read. This post covers the design, then the limits.
What the assay measures
A peptide is incubated in plasma or serum at 37 °C. Samples are removed at intervals, the reaction in each is stopped, and the remaining intact peptide is quantified, almost always by liquid chromatography with mass spectrometric detection. Plot the remaining peptide against time and you get a decay curve, from which an in-vitro half-life falls out.
That is the entire measurement, and its meaning is narrow and precise: this is how fast this molecule is destroyed by the enzymes present in this matrix, under these conditions. Peptides are vulnerable to proteases because a peptide bond is what a protease cuts, and plasma carries a substantial population of them. A short unmodified peptide can have an in-vitro half-life of minutes. A modified one can run for many hours.
Plasma or serum?
These are two different matrices and the distinction is not pedantry.
Plasma is blood with an anticoagulant added, centrifuged to remove cells. The coagulation cascade has been prevented, so its proteases have not fired and the pathway is intact. The choice of anticoagulant matters too: a chelating anticoagulant removes calcium and therefore suppresses the metalloproteases that need it, while a direct thrombin inhibitor leaves them active.
Serum is blood allowed to clot, then centrifuged. The cascade has run to completion, its enzymes have already acted, and fibrinogen and some other components are gone.
The same peptide can give a different half-life in each, and published work comparing blood, plasma and serum directly for a set of peptides found exactly that. Two consequences follow. Pick the matrix that matches what you are modelling and say which it was, because a stability figure without the matrix named is close to uninterpretable. And do not compare your number with a literature number from the other matrix as though they were the same measurement.
Species matters on the same principle. Human, rat and mouse plasma have different protease complements, and rodent plasma is often the more aggressive. A half-life measured in one species does not transfer to another, which is a specific instance of the argument in why trial data does not transfer.
The design
Concentration. Low enough that the peptide does not saturate the proteases, which would flatten the curve and make the molecule look more stable than it is. A peptide should be a trace substrate.
Time points. Spaced to bracket the expected half-life, which means guessing it first and widening if you guess wrong. A short unmodified peptide may need points at 0, 5, 15, 30, 60 and 120 minutes; an acylated one may need hours or a full day. Points clustered entirely before or entirely after the half-life give a curve with no usable slope.
The zero timepoint. Taken by adding peptide to plasma already quenched, or quenched immediately on mixing, so it defines the starting concentration under the same handling as every other sample. Using a nominal calculated concentration instead of a measured zero point imports every pipetting and recovery error into the result.
The buffer-only arm. The same peptide incubated in buffer at the same temperature for the same duration. Chemical degradation runs in the absence of any enzyme, and the routes it uses are the ones mapped in which residues degrade first. Without this arm you cannot say whether loss in the plasma tube was proteolysis or ordinary hydrolysis, and for a peptide carrying a labile bond that is a real ambiguity.
A positive control. A peptide known to be rapidly cleaved, run in the same plasma on the same day. If it survives, the plasma was not active, perhaps because it had been frozen and thawed too often or stored too long, and the stability you measured for your compound is an artefact. This single control catches the assay's most embarrassing failure mode.
Replicates. Three independent incubations at minimum, because the curve is being fitted and a fit through single points carries no uncertainty estimate.
Quenching, which is where it goes wrong
At each timepoint the reaction has to stop completely and immediately. If it does not, the sample continues to react while it queues for analysis, and the apparent time course is partly a record of how long each tube waited.
The standard approach is to add a volume of cold organic solvent, commonly acetonitrile, often with an acid. This denatures and precipitates the plasma proteins, which both stops the enzymes and removes the protein load that would otherwise foul the chromatography. The precipitate is spun out and the supernatant analysed.
Three details carry most of the risk. The solvent should be cold and in excess, because a marginal volume gives incomplete precipitation and partial quenching. An internal standard added with the quench solvent, ideally an isotopically labelled version of the peptide, corrects for recovery losses during precipitation, which are real and variable. And recovery itself should be checked rather than assumed: peptides adsorb to surfaces and co-precipitate with protein, and a peptide that is 60% recovered is still measurable but only if you know it is 60%. The post on why peptides stick to plastic covers the adsorption side of that.
Reading the result
Fit the remaining-peptide values against time. Decay in these assays is usually first-order over the useful range, so a plot of the logarithm of remaining peptide against time is approximately a straight line and the half-life is the natural logarithm of two divided by the slope.
Deviations are informative rather than inconvenient. A curve that drops fast and then flattens often means a fraction of the peptide is bound to plasma protein and protected while the free fraction is cleaved. A curve that starts slowly and accelerates can mean a cleavage product is itself a substrate. A curve that goes nowhere means either a genuinely stable peptide or inactive plasma, which is precisely the question the positive control answers.
If the analysis is by mass spectrometry you also get the fragments, and the fragments name the cleavage sites. That is the most actionable output of the whole experiment: knowing where a peptide is cut tells you which bond to modify, and it is how the standard stabilisation strategies were found in the first place. How a peptide sequence is verified covers the fragment reading.
What it cannot tell you
Four limits, and they are the reason this assay is a screen rather than a prediction.
It is not a circulating half-life. Clearance in an organism involves glomerular filtration, hepatic uptake, distribution into tissue and binding to plasma proteins, none of which exist in a tube. An in-vitro plasma half-life bounds one contribution to the real thing. The post on why half-life is a property of a body makes that argument at greater length, and the acylated compounds are the clearest demonstration: their long duration comes substantially from albumin binding and slow redistribution rather than from protease resistance alone.
It is not an activity measurement. The assay counts intact molecules. A peptide can be cleaved into a fragment that is still active, or can be intact and inactive because it has oxidised or isomerised somewhere that the mass spectrometer's quantitation does not distinguish.
It is matrix-specific and species-specific. The number belongs to the plasma it was measured in.
It says nothing about a dry vial. Stability in plasma and stability in storage are unrelated questions with unrelated mechanisms. Storage stability is the subject of how to run a peptide stability study, and this catalogue holds no plasma stability data for any of its compounds.
Frequently asked questions
What is the difference between plasma and serum for this assay?
Serum has already clotted, so the coagulation proteases have fired and been consumed; plasma is anticoagulated and keeps those pathways intact. The same peptide can give different half-lives in each, so the matrix must be reported.
How long should the incubation run?
Long enough to pass the half-life and short enough that the matrix is still behaving. Time points should bracket the expected half-life rather than cluster on one side of it, which usually means a pilot run to find the range.
Why does my peptide look stable in plasma when it should not be?
Check the positive control first. Plasma that has been stored too long or thawed repeatedly loses protease activity, and inactive matrix makes everything look stable. The second thing to check is whether the peptide concentration was high enough to saturate the enzymes.
Does a long in-vitro half-life mean a long half-life in vivo?
No. It means the molecule resists plasma proteases. Real clearance also involves the kidney, the liver, tissue distribution and protein binding, and for the acylated compounds albumin binding contributes more than protease resistance does.
Do I need a buffer-only control?
Yes, otherwise chemical degradation and enzymatic cleavage are indistinguishable. For a peptide carrying a labile bond, a meaningful share of the loss can be chemical.
Does Pepstral publish plasma stability data for these compounds?
No. We hold none, and this post describes the method rather than reporting results. What the certificates do report is identity, purity, content and endotoxin, which the certificate library sets out.
References
- Di L, Kerns EH. Drug-like properties: concepts, structure design and methods. The standard treatment of in-vitro stability assays, including plasma stability protocol design and the role of quenching. www.sciencedirect.com/book/9780128010761/drug-like-properties
- Bottger R, Hoffmann R, Knappe D. Differential stability of therapeutic peptides with different proteolytic cleavage sites in blood, plasma and serum. PLOS ONE 2017;12(6):e0178943. The direct comparison showing blood, plasma and serum give different results for the same peptide. doi.org/10.1371/journal.pone.0178943
- Jenssen H, Aspmo SI. Serum stability of peptides. Methods in Molecular Biology 2008;494:177-186. doi.org/10.1007/978-1-59745-419-3_10
- Knudsen LB, Lau J. The discovery and development of liraglutide and semaglutide. Frontiers in Endocrinology 2019;10:155. Background on albumin binding and acylation as a half-life strategy. doi.org/10.3389/fendo.2019.00155
Every product mentioned is sold for laboratory research use only and is not for human or animal use. Nothing on this page describes or recommends use of the material sold here in humans or animals.
