Endotoxin vs sterility: why sterile does not mean endotoxin free

Sterility and endotoxin testing answer different questions. A sterility test looks for viable contaminating microorganisms, while an endotoxin assay looks for bacterial material that can remain after those organisms are gone.[7][1] In peptide research, that material can complicate the interpretation of a cell response: the intended reagent may not be the only thing the cells are reacting to.[5]

Reader offer · IQON Labs

Buy 1, Get 1 Free

Pick any research compound. The second one is on us.

FREE

Applied automatically at checkout, no code needed

Claim my free vial →

Lower priced item free; accessories excluded. For research use only.

An explanation of laboratory evidence. The studies discussed do not test or endorse IQON Labs.

The bacteria can be gone before the endotoxin is

Endotoxin comes from the outer membrane of Gram-negative bacteria. It is associated with a molecule called lipopolysaccharide, or LPS.[3] It is bacterial material, not a living organism. A test designed to find viable microbes and a test designed to detect LPS therefore have different targets.[7][3]

Removing the organisms does not necessarily remove that material. In a 2019 laboratory study, researchers examined a commercial antibody preparation that already contained endotoxin.[4] Sterile filtration made only a small difference to the measured endotoxin activity.[4] The example involved an antibody rather than a peptide, and the authors did not perform additional replicates or statistical analysis for that demonstration.[4] It nevertheless shows the distinction in a measured sample: filtration and endotoxin removal were not interchangeable.

“Pyrogen” is broader still. It describes a substance that can produce fever, and includes substances other than bacterial endotoxin.[1][2] A result for one of these targets should not be silently turned into a result for all of them.

When a contaminant supplies part of the response

The relevance to peptide research was apparent in a 2007 screen of laboratory reagents used to study microglia, the brain's resident immune cells.[5] The researchers found endotoxin in some synthetic peptide preparations, including peptides used experimentally to activate receptors.[5]

They also asked whether contamination could account for cellular effects attributed to other reagents. For that part of the work, they selected proteins and tested their effects on a microglial cell line and primary mouse microglia.[5] The comparisons included conditions with polymyxin B and conditions without it.[5] This compound binds endotoxin and inhibits many of its effects.[5]

Some inflammatory responses fell when the inhibitor was present. Others persisted, leaving room for genuine activity of the intended protein or other contaminants.[5] The experiment did not turn every unexpected response into an endotoxin artifact. It showed how an apparently active preparation could contain more than one cause of activity.

The distinction between the paper's two parts matters: it measured endotoxin in peptide samples, but those functional follow-up experiments used selected proteins.[5] It was a study of particular laboratory materials, not a survey of today's retail peptide supply.

Why two endotoxin assays missed what cells could detect

A separate study in 2017 exposed a harder problem. Researchers produced the same recombinant protein in ordinary E. coli and in an engineered strain whose modified LPS did not trigger the usual response in their test systems.[3] Neither protein preparation produced a detectable endotoxin signal in the LAL assay or the recombinant Factor C assay used in the study.[3]

LAL, short for Limulus amebocyte lysate, is an endotoxin assay whose reaction depends on Factor C. The recombinant assay uses a manufactured version of that protein.[3] Using both methods gave the researchers two measurements, but the tests shared a biological recognition step.

Human monocytes, immune cells isolated from donated blood, responded differently. After 24 hours, the protein preparation from ordinary E. coli increased inflammatory signaling molecules, while the preparation from the engineered strain did not show that effect.[3]

To investigate the discrepancy, the team blocked TLR4, a receptor involved in recognizing LPS. The inflammatory response decreased.[3] They also used engineered cells that produced a light signal when the LPS recognition pathway was activated. That result supported endotoxin as the source of the activity missed by the Factor C assays.[3]

This was research on a recombinant protein and laboratory cells, not a test of commercial peptide products or a clinical safety study. Its lesson concerns the measurement: changing the assay name does not necessarily remove a limitation the methods share.

A fresh control does not tell the whole history

A laboratory can check for interference by adding a known amount of endotoxin to the sample and seeing whether the assay recovers it.[1][3] This is called a spike control. If the added material fails to produce the expected response, the sample may be suppressing the test.

The researchers found a more subtle pattern. LPS held in certain buffer mixtures became poorly detectable in Factor C assays, while fresh endotoxin spikes added shortly before measurement were still recovered.[3] The test could respond to the fresh addition even though it missed much of the endotoxin that had been sitting in the mixture.

That time-dependent loss of detectable activity is called endotoxin masking, a cause of low endotoxin recovery.[3] It is different from showing that the contaminant has been removed. Buffer components can change how LPS is arranged and presented to the assay's recognition machinery.[3]

The researchers then tested whether masked LPS could still activate primary human monocytes. Their cytokine comparison used eight independent experiments, with measurements after 24 hours.[3] These were laboratory repetitions with donor-derived cells, not participant counts in a clinical trial. Masked LPS still stimulated inflammatory responses, although some responses were weaker than those to unmasked LPS.[3]

The effect also varies with the endotoxin itself. The 2019 study found different masking behavior across bacterial sources and growth conditions, including preparations that remained comparatively resistant to masking under the tested conditions.[4] Neither paper supports treating every peptide solution as a masking problem. They explain why the relevant sample conditions have to be examined rather than assumed.

About the evidence

The 2017 study received Austrian Science Fund support and assay reagents from Hyglos; the authors declared no competing financial interests.[3] The 2019 paper included Microcoat Biotechnologie affiliations and declared no external funding or conflicts.[4] The 2007 study acknowledged public and nonprofit grants, along with donated recombinant proteins from biotechnology companies.[5] These studies provide experimental examples, not supplier endorsements.

What makes a negative result informative

A low endotoxin reading becomes more useful when its detection limit and the sample's compatibility with the assay are known.[2] FDA's March 2026 guidance also addresses what happens before measurement: storage and handling can affect the ability to detect endotoxin, so laboratories need evidence that assayable endotoxin remains stable through those steps.[2]

Dilution can help when sample ingredients interfere with a test, but it also lowers the amount of endotoxin presented to the assay. The guidance therefore connects dilution to the level the method must still be able to detect.[2] Diluting until the reading is low is not the same as establishing a meaningful negative result.

These are principles of laboratory interpretation, not a way to establish that a research product is suitable for people. The studies do not supply a universal acceptable endotoxin level for peptide experiments or a human safety threshold.

Our guide to reading a peptide certificate of analysis covers the separate task of matching a report to its sample. For an endotoxin result, the next question is whether the method could detect the contaminant in that sample, including after the relevant handling and storage.[2]

The masking experiments show why that question matters: a fresh spike could be detected while older endotoxin remained poorly visible to the same assay and still activated immune cells.[3] Resolving that discrepancy helps a researcher decide what caused the cellular response, before assigning it to the peptide under investigation.

Frequently asked questions

Is endotoxin the same as bioburden?

No. Bioburden concerns the level of microbial contamination, while an endotoxin result concerns bacterial material that may remain after microorganisms have been removed. FDA's technical guide explains that lowering the microbial level does not necessarily produce a similar reduction in endotoxin. A low bioburden result is therefore not a substitute for an endotoxin measurement.

[1]

Can a pooled sample conceal a contaminated individual sample?

It can make the contamination harder to detect. FDA's guidance on pooling finished product samples warns that combining a unit containing more endotoxin with units containing less can dilute the signal. The guidance consequently recommends accounting for that dilution and describes situations where individual testing is more appropriate. A pooled result needs its sampling context; it is not a separate measurement of every container.

[2]