# Endotoxin testing systems

The enzymatic cascade behind the LAL test, the Factor G cross-reaction that recombinant Factor C removes, and the low endotoxin recovery problem that no reagent chemistry has solved.

Why a horseshoe crab clotting cascade detects lipid A, what β-glucans do to it, and why an endotoxin unit is not a mass.

Source: https://en.bioecon.ru/docs/bioproduction-equipment/analytics-pat/endotoxin-testing-systems/
Updated: 2026-09-04



Bacterial endotoxin is the lipopolysaccharide of the Gram-negative outer membrane, and the pyrogenic part of it is lipid A — the acylated diglucosamine anchor. It is heat-stable in a way vegetative cells are not: an autoclave sterilises without depyrogenating, which is why glassware is depyrogenated by dry heat, conventionally 250 °C for 30 minutes to demonstrate a three-log reduction. Sterility and freedom from endotoxin are separate problems requiring separate steps.

## The cascade is the detector

The Limulus amebocyte lysate test borrows an innate immune mechanism. In the horseshoe crab, circulating amebocytes carry a zymogen cascade whose first member, Factor C, is a serine protease that is itself activated by direct binding to lipid A. Activated Factor C cleaves Factor B, activated Factor B cleaves the proclotting enzyme, and the proclotting enzyme converts coagulogen into coagulin, which polymerises into a gel. Each step is catalytic, so a few picograms of lipid A produce a macroscopic result. The original gel-clot readout is a simple pass/fail at the point where the tube no longer flows. Quantitative formats replace coagulogen with a synthetic peptide carrying para-nitroaniline, cleaved to a chromophore read at 405 nm, or track the turbidity rise of the forming gel. Because the signal is amplified through a cascade, the calibration is log–log against onset time, not a linear absorbance relation.

## What else activates it

The lysate contains a second, parallel pathway. Factor G is activated by (1→3)-β-D-glucans — not endotoxin at all, but polysaccharides shed by cellulosic depth filters, some membrane materials and fungal contamination. The cascade converges downstream, so a glucan-driven clot is indistinguishable from an endotoxin-driven one and reads as a failure the product does not have. The classical fix is a glucan-blocking buffer that inhibits Factor G. The structural fix is recombinant Factor C: a single expressed protein with a fluorogenic substrate, containing no Factor G pathway to cross-react and requiring no horseshoe crabs. The European Pharmacopoeia describes the recombinant Factor C test in chapter 2.6.32, and the compendial route to using it is now established rather than exceptional.

## Units, and the limit that is not solved

An endotoxin unit is defined by bioactivity against an international reference standard, not by mass. This is unavoidable: lipid A potency depends on acylation pattern, which differs between species, and on aggregation state, since LPS forms micelles and lamellar structures whose presentation to Factor C changes with ionic strength and detergent. Pharmacopoeial limits are dose-based — 5 EU per kilogram per hour for most parenteral routes, with a far tighter 0.2 EU/kg for intrathecal administration, because the blood–brain barrier is bypassed.

The unresolved problem is low endotoxin recovery. In formulations that combine a chelator such as citrate or phosphate with a non-ionic surfactant such as a polysorbate, endotoxin deliberately spiked into the product becomes progressively undetectable over days. The accepted explanation is masking: the formulation disaggregates the LPS supramolecular structure into a form the cascade no longer recognises, without removing it. The material is still there and still potentially pyrogenic. There is no universal demasking method, and hold-time spike recovery studies exist because the assay cannot be trusted to see what it is looking for in every matrix.