Endotoxin and Bioburden in Peptide Preparations: What Testing Establishes

No — a chemically pure peptide is not necessarily free of endotoxin. Chemical purity, sterility and endotoxin content are three independent properties, each established by a different test. A preparation can be high in purity by chromatography, correct by mass spectrometry, and still carry substantial endotoxin: it is not removed by the purification that establishes chemical purity, nor destroyed by the conditions that establish sterility 12. A chemistry certificate does not report it.

Four properties, four determinations

In casual use these collapse into one impression — clean material — and in fact they do not. Purity states the proportion of chromatographable material that is the target sequence. Sterility states that no viable organism was recovered under defined conditions. Bioburden counts viable organisms. Endotoxin content states the quantity of one bacterial cell-wall component present, alive or not.

The separation is easy to miss because the analysis ordinarily supplied addresses only the first. Reversed-phase chromatography resolves the target from related chemical species — deletion sequences, oxidised forms, residual reagents — and reports it as an area percentage. Lipopolysaccharide is not a related chemical species and is not detected by that method 12. Bioburden relates to endotoxin in one direction only: material that carried a Gram-negative population probably carries its residue, while a low count says nothing about residue present.

What endotoxin is

Endotoxin is lipopolysaccharide, a structural component of the outer membrane of Gram-negative bacteria: a variable polysaccharide chain, a core oligosaccharide, and lipid A, the anchoring lipid carrying most of the biological activity. It is a piece of the organism, not a secreted product — and so it is shed during normal growth and released in quantity when the cell envelope breaks up on death 1.

Killing the organism therefore does not remove the endotoxin. It liberates it. A heat or filtration step that takes a viable count to zero converts a population of intact cells into free lipopolysaccharide in the same volume. A preparation can be sterile precisely because a Gram-negative population was killed in it.

Lipopolysaccharide is also robust for a biological molecule. It is amphipathic, aggregates in aqueous media, adsorbs readily to surfaces, and survives conditions that destroy proteins outright 2. Those properties make it persistent in water systems and equipment, and are why removal from an aqueous preparation is a specialist operation.

Why it matters at the bench

The reason a bench scientist cares is experimental validity. Endotoxin is biologically active at very low concentrations in cell-based and immunological systems. Innate immune receptors recognise it, and the signalling that follows produces cytokine release, changes in gene expression, altered proliferation and shifts in viability — in monocytes and macrophages particularly, but also in endothelial and primary cells, and in any co-culture containing an immune-competent cell type 1.

What that produces is a confound with the shape of a genuine finding. Contaminating endotoxin travels with the preparation, so the response tracks the concentration applied. Every feature read as evidence of compound activity is present: a concentration-response relationship, reproducibility across replicates, and loss of effect when the preparation is omitted. Nothing in the form of the result separates peptide from contaminant 1.

State it as a rule of practice. An unexplained inflammatory or immune readout in a cell assay — cytokine induction, activation of an innate signalling pathway, an unexpected effect on viability — warrants endotoxin testing of the preparation before the effect is attributed to the peptide. It is the cheapest available explanation for a class of result that is easy to report and hard for anyone else to reproduce. Replication within the same contaminated preparation settles nothing.

Thermal stability and depyrogenation

The point most often missed is thermal. Autoclaving — saturated steam at 121 °C for a standard cycle — is a sterilisation process. It kills organisms. It does not destroy lipopolysaccharide, which survives such conditions largely intact and remains active afterwards 2. Autoclaved water in an autoclaved flask can be sterile and still carry the endotoxin of whatever grew in it beforehand.

Destroying endotoxin by heat requires dry heat at a substantially higher temperature, held longer than a sterilisation cycle. That is depyrogenation, and it uses different equipment. Glassware can be treated this way; solutions and heat-labile material cannot, and are handled by removal instead — adsorption onto affinity or ion-exchange media, two-phase extraction, ultrafiltration where the size difference permits, and membranes chosen to retain endotoxin rather than merely cells 2. Each route costs recovery, and separating a small peptide from an aggregating amphiphile is rarely clean.

Sources in the laboratory

Endotoxin enters a preparation from the environment in which it was made and handled. The list of sources is short and entirely ordinary, which is the difficulty: none of them looks like contamination.

Test methods and what each detects

Detection rests on one biological reaction and a recombinant reconstruction of it. The Limulus amoebocyte lysate reaction is a clotting cascade from horseshoe crab haemolymph triggered by lipopolysaccharide; the three lysate formats differ in how the cascade is read out, not in what starts it. Recombinant factor C substitutes an expressed version of its first enzyme.

Interference is the failure mode that matters, and peptides are among the materials that cause it. A preparation may inhibit the cascade, depressing the reading so a contaminated sample tests clean, or enhance it, so a clean sample tests contaminated. Which occurs depends on the sequence, the counterion, the buffer and the concentration presented, and cannot be predicted from the certificate 2.

The control is a spike recovery: a known quantity of standard endotoxin is added to a portion of the sample and assayed alongside the unspiked one. Recovery within the accepted window shows the sample is not interfering at that dilution, and the unspiked result can be read. If recovery fails, dilute further and repeat until it passes — at the cost of raising the lowest concentration the assay can resolve. Report the dilution used; a figure without it cannot be interpreted.

What a certificate does and does not state

A chemistry certificate of analysis states chemical attributes: sequence, molecular mass, purity by chromatography, identity by mass spectrometry, content, counterion and water. Microbiological attributes fall outside its scope unless separately named. The absence of an endotoxin entry is not a negative result — it means the test was not performed, or not reported, and the document cannot distinguish the two. The companion reference on reading a certificate of analysis covers the chemistry fields.

Where an entry is present, it is interpretable only with the detail accompanying it: the method, the dilution read, whether spike recovery passed, and the quantity of material the figure is expressed against. A bare number is a claim rather than a measurement. And like every certificate entry it describes one batch at one moment; endotoxin introduced afterwards, by the diluent or vessel at reconstitution, lies outside it.

What published analyses report

Published product analysis bears on this directly. A 2024 study in the Journal of Medical Internet Research obtained semaglutide products from online sellers operating without a prescription and put them through laboratory analysis; among its findings, endotoxin was detected in the vials tested 3. That is what the study reports, on the specific units it purchased. It supports no proportion for any wider market.

The broader analytical literature on illegally supplied peptide preparations has concentrated on identity and content — whether the labelled compound is present at all, and in what quantity 4. Microbiological attributes are examined far less often, so the absence of published endotoxin findings for a class of preparation reflects what has been looked for. The working position is that endotoxin status is unknown until measured on the material in hand.

Deciding whether testing is warranted

1. Classify the readout. If the endpoint is immune, inflammatory or cytokine-based, or the system uses monocytes, macrophages or primary cells, treat testing as part of preparing the material.
2. Establish whether the preparation has ever been tested. Look for an entry carrying method, dilution and control outcome. No entry means untested, not clean.
3. Account for everything added after the vial. A tested peptide reconstituted in untested water is an untested preparation.
4. Test the preparation as it will be applied — working dilution, final buffer — not the neat solid.
5. Run a spike recovery control alongside, every time. Without it, an inhibited assay returns a clean result on a contaminated sample.
6. If recovery fails, dilute further and repeat until it passes. Record the dilution reached and the resulting detection limit.
7. Where testing cannot be arranged, discriminate by control instead: a vehicle-only arm carrying the same diluent and vessels, and a parallel arm with an endotoxin-neutralising agent.
8. Record the outcome against the batch — method, dilution, spike recovery, date — so a later anomaly can be checked against a measurement.
9. Re-test on a new batch, or after a change of diluent or vessels. A result does not transfer between preparations.

The decision this procedure supports is narrow. Endotoxin testing does not make a preparation suitable for anything; it establishes one quality attribute of one batch under stated conditions, so a readout can be attributed to the compound with some confidence rather than none 1. Where the endpoint is immunological, an untested preparation carries an alternative explanation that further replication cannot remove.