What Is Endotoxin Testing?

Endotoxin testing checks a biological material for lipopolysaccharide, a fragment of the outer membrane of Gram-negative bacteria that can trigger a strong immune reaction even after the bacteria themselves are dead. The standard method, in use since the 1970s, runs the sample against a reagent made from horseshoe crab blood cells, because those cells clot in the presence of even trace amounts of it.

The test shows up on everything from injectable drugs to medical devices to water used in manufacturing, and it shows up, as a separate line item, on some peptide certificates of analysis. The reasons for that are more specific than "contamination is bad," and they're worth separating from the rest of what a certificate of analysis covers.

What endotoxin actually is

Endotoxin is not a toxin a bacterium secretes on purpose. It's lipopolysaccharide (LPS), a molecule built into the outer membrane of every Gram-negative bacterium, and it's released when the cell wall breaks apart, including when the bacterium dies. That matters because sterilizing a material kills bacteria but does nothing to destroy the LPS already shed into it. A sample can be sterile — no living organisms — and still carry enough endotoxin to cause a reaction.

The reaction it causes is a strong one. LPS is recognized by a receptor complex on immune cells built around a protein called TLR4, and that recognition is exactly what the immune system evolved it to do: treat LPS as a reliable sign that Gram-negative bacteria are present, and respond accordingly. In a living body above a threshold dose, that response is fever and, at high enough levels, septic shock, which is why endotoxin is classified as a pyrogen — a fever-causing substance — and why injectable drugs, implanted devices and the water used to make them are required to be tested for it.

How the test actually works

Before the 1970s, the only practical way to test a drug for pyrogens was to inject it into a rabbit and watch for a fever over several hours — a slow, animal-intensive assay that companies ran by keeping large rabbit colonies on-site. The replacement came out of a surprising spot: Johns Hopkins researcher Frederik Bang noticed in 1953 that horseshoe crab blood would clot dramatically in the presence of a particular marine bacterium, and that even dead bacteria triggered it. About a decade later, Bang and hematology fellow Jack Levin traced the reaction to amebocytes, the only cellular component of horseshoe crab blood, and found the clotting cascade inside them responded specifically to endotoxin rather than to the bacteria as a whole. They published the finding in 1964, and the resulting test, using a reagent called Limulus amebocyte lysate (LAL), reached full FDA approval as a finished-product test in 1983 after a period of review over how sustainably the crabs could be bled for it.

The mechanism is a cascade: endotoxin activates a clotting-factor enzyme inside the lysate, which activates the next enzyme in line, which eventually converts a soluble protein called coagulogen into an insoluble gel or, depending on the test format, releases a detectable color. The amount of signal tracks the amount of endotoxin present, which is what turns a yes/no clotting reaction into a number. Horseshoe crabs are bled, not killed, for the reagent, and the specific chemistry is different enough from mammalian biology that it responds to bacterial LPS with a precision a whole-animal fever response can't match — and it returns a result in under an hour instead of several.

Three ways labs run the assay

The compendial standard — USP <85> in the United States, with equivalent chapters in the European and Japanese pharmacopeias — recognizes three LAL-based formats. Gel-clot is the original and simplest: the lysate either visibly gels or it doesn't, against a defined concentration limit, and it's the method pharmacopeias default to if a dispute arises over a result from one of the other two. Turbidimetric and chromogenic methods are both quantitative, tracking cloudiness or a color change over time to produce an actual endotoxin concentration rather than a pass/fail call, each expressed in endotoxin units (EU).

All three depend on wild-caught horseshoe crab blood, which has driven two things in parallel: growing attention to whether the harvest is sustainable, and the development of recombinant alternatives that replace the natural lysate with cloned versions of the same clotting-factor enzymes, made without bleeding a crab at all. A 2021 review in Biomedicines by Hiroshi Tamura, Johannes Reich and Isao Nagaoka traces regulatory recognition of recombinant Factor C (rFC) reagents through the European Pharmacopeia (chapter 2.6.32, adopted 2020) and the Japanese Pharmacopeia (2021), alongside a proposed — not yet finalized, as of that review — USP chapter for the same approach. The authors frame conservation as a direct driver of that work, calling horseshoe crab blood "a natural, irreplaceable, and precious resource" worth protecting through reduced reliance on it.

Whatever the format, every run also has to pass an interference check specific to that sample. A test article can falsely suppress or exaggerate the clotting reaction depending on what else is in it, so labs spike samples with a known amount of endotoxin and confirm the assay still recovers it before trusting a result from that matrix.

Why it shows up on a peptide report at all

Peptides aren't injectable drugs, and a research-use peptide isn't held to a pharmacopeial endotoxin limit the way a licensed parenteral product is. So why does any lab bother testing one for it?

Because the people actually working with the material often need it low for a reason that has nothing to do with drug approval: endotoxin is one of the most potent, promiscuous confounders in cell-based and animal research. A 2014 study in PLoS ONE by Schwarz and colleagues tested commercially available recombinant proteins and found some carried measurable endotoxin, in a few cases above what the supplier's own data sheet claimed — then showed that a specific class of human dendritic cell could be activated by LPS at concentrations in the same range as that contamination. Because the TLR4 receptor that detects LPS sits on far more than just the immune cells an experiment is trying to study, trace contamination in a reagent can quietly activate unrelated pathways and produce a result that looks like a finding about the peptide, when it's actually a finding about what was riding along with it.

That's the practical case for a peptide COA listing an endotoxin result: it's not a drug-safety claim, it's a data-quality one, aimed at researchers running cell culture or animal work who need to rule out LPS as the actual cause of whatever they observed. It's also a different claim than the purity and identity results Torrey Labs' guide to reading a peptide lab report walks through, which speak to what the peptide is; an endotoxin result speaks to what else might have come along with it.

What's known, and what's still unsettled

What's well established: the mechanism (LPS activates a defined clotting cascade in horseshoe crab blood, and a defined receptor pathway in mammalian immune cells), the compendial methods (gel-clot, turbidimetric, chromogenic), and the regulatory baseline for drugs, biologics and devices under USP <85> and USP <161>.

What's still moving: how far recombinant reagents will displace wild-caught lysate, and how fast regulators formalize that shift — the Biomedicines review describes European and Japanese adoption as further along than the American one as of its 2021 publication, and that kind of harmonization timeline is the sort of thing that keeps changing year to year. There's also no single universal endotoxin limit the way there's no single universal purity threshold: the acceptable level in a pharmacopeial test depends on the product, the route of administration and the maximum dose, and research-use material sits outside that framework entirely, which is exactly why a report that includes an endotoxin number is making a narrower claim than it might first appear to — a measurement against the lab's own method, not a certification against a regulatory ceiling that doesn't apply to the material in the first place.

What this means for researchers

An endotoxin result on a peptide report is worth reading as what it actually is: evidence about a specific potential confounder, generated by a specific test format, on a specific lot. It says nothing about identity or net content — those are separate tests covered elsewhere on a certificate of analysis — and for material that isn't a licensed drug, it isn't being measured against a pharmacopeial pass/fail line in the first place.

For anyone whose work actually depends on ruling out LPS contamination — cell culture, animal dosing, any immune-relevant readout — treat the presence of a real, lot-specific, numeric endotoxin result as the thing to check for, the same way identity and purity results are checked, through Torrey Labs' public lab-report lookup rather than taken on description alone. A report that's silent on endotoxin isn't evidence the material is clean; it's evidence the test wasn't run.