A cell-culture failure traced to one missing number

Investigators reporting in Regenerative Therapy traced a failed endotoxin test in their cell-sheet culture to a purchased medium supplement whose certificate stated no endotoxin concentration.

What the investigators were making

The team was fabricating autologous oral mucosal epithelial cell sheets, thin layers of cells grown from a donor's own oral tissue, as a cell-based product under clinical study. Every batch of sheets passes a set of quality-control tests before release. One of them is the bacterial endotoxin test on the culture supernatant, the liquid medium the cells are grown in. The acceptance criterion the investigators set for that test was less than 1.0 EU/mL.

The failure

During validation, the supernatant from one volunteer's culture measured 6.68 EU/mL. During the clinical study, a participant's culture measured 1.31 EU/mL, and a third-party laboratory confirmed it at 1.74 EU/mL. Both exceeded the limit. Fresh medium prepared for that participant and never used for culture measured 2.97 EU/mL, which placed the source in the materials rather than in the handling of the cells.

Finding the source

The investigators measured endotoxin in each reagent used to prepare the medium. The source was a commercial cholera toxin solution, a standard supplement in epithelial culture media. Three lots of the product each measured above 300 EU/mL. The product had been supplied without a certificate of analysis stating its endotoxin concentration, and the reagents had not been tested for endotoxin before use.

The resolution was a number

The investigators moved to cholera toxin from a second manufacturer whose certificate stated the endotoxin content of each lot: 12, 45 and 5.0 EU/mg across three lots. From those figures they calculated the endotoxin the supplement would add to the finished medium, between 0.0004 and 0.0038 EU/mL, well under their limit. Cultures prepared after the change showed no detectable endotoxin.

Why the number mattered

A certificate that reports endotoxin as "pass" states that a measurement fell under a threshold the reader cannot see. The investigators' limit applied to their finished medium, not to the supplement on its own. The figure they needed was one they could carry into their own calculation: endotoxin per milligram, multiplied by the amount added to the medium. A pass/fail verdict cannot be converted into a concentration in the medium. A value in EU/mg can.

The authors recommend that endotoxin in fabrication reagents be measured by the facility itself or confirmed by an adequate certificate of analysis from the manufacturer, and that it be checked again whenever a lot or a manufacturer changes.

Reading the endotoxin line on a certificate

Three figures can appear on the endotoxin line of a certificate, and only one of them describes the material. The assay sensitivity describes the lysate reagent used for the test. The limit is the threshold the verdict was judged against. The measured value, a number with a unit, is the figure a laboratory compares with its own criterion. The unit matters as much as the number: a value per milligram, per milliliter or per vial describes the same material in different terms, and only a stated unit lets the figure be carried into a calculation like the one these investigators made.

The Crescent Foundation walks through each line of a real certificate, including the endotoxin page, at crescentbioresearch.info.

The certificate that resolved the failure printed the endotoxin value for every lot.