Endotoxin in cell assays
Endotoxin is lipopolysaccharide from the outer membrane of Gram-negative bacteria. It is a potent activator of innate immune signaling, it is stable to autoclaving, and it sticks to plastic and to proteins. In a cell assay it produces cytokine release, changes in viability and changes in surface marker expression, all of which look exactly like the biology people are hoping to see.
Where it comes from
Bacterial expression is the obvious source, and His-tagged material purified from bacteria needs attention by default. Less obvious sources cause more trouble: water, buffers made up in advance, columns and resins that have sat wet, plasticware, and plasmid preparations, which is why transfection-grade preparations are specified for cells.
Mammalian expression is not automatically clean, because the reagents and the handling can introduce it.
Measuring it
The limulus amebocyte lysate assay in its several formats is the standard, along with newer recombinant factor C assays that use a cloned enzyme instead of animal-derived lysate. Both are sensitive and both are interfered with by many buffer components, so samples usually need dilution, and the dilution has to be validated by spiking a known amount and checking recovery.
Report a concentration per milligram of protein, not just per milliliter, since what reaches the cells depends on how much protein is dosed.
Keeping it out
- Dedicated glassware and plasticware, and endotoxin-free water and buffers for anything destined for cells.
- Clean columns and resins, with a documented sanitization step. A shared resin is a shared history.
- Remove it when it is there. Polymyxin B resins, anion exchange in the right mode, phase separation with a mild detergent, and repeated washing on an affinity column all work to varying degrees. Efficiency depends on the protein, so measure after removal rather than assuming.
- Test the final material, in the final buffer, at the concentration that will be dosed.
The control that saves the experiment
Run a polymyxin B control arm, or an endotoxin spike on a clean sample, whenever a cell result is surprising and a protein preparation is involved. If the effect disappears when endotoxin is neutralized, the effect was endotoxin.
For panels, the giveaway is a candidate that behaves differently in cell assays and normally in binding assays. Before writing that down as a mechanism, check the endotoxin numbers for those lots.
How much is too much
It depends on the assay. Primary immune cells respond to very low levels. Cell lines are less sensitive, and a viability assay tolerates more than a cytokine readout. Rather than a universal threshold, set one for the assay from a titration of purified endotoxin, and hold every lot below it.