Because foetal bovine serum is a very complex mixture of nutrients including proteins, growth factors, vitamins, co-factors. bases, amino acids, metabolites and salts, removal of endotoxins should leave these nutrients largely intact and thereby retain most of the cell culture activity of the finished product.

Processes such as ultrafiltration, ion exchange chromatography (DEAE cellulose), charcoal, Seitz asbestos pad or Zeta plus filtration, and solvent extraction methods can be used to lower the endotoxin levels of solutions. however, when applied to serum, these methods either adsorb growth factors or effectively destroy the serum by protein denaturation.

Therefore, removing endotoxins from serum solutions and other cell culture products can be difficult and often expensive. Recently, solid-phase endotoxin-adsorbing reagents have become available, but their binding affinities are often too low to dissociate endotoxin-protein complexes effectively. Nonspecific binding of proteins and growth stimulating substances to these reagents is another common problem.

To address these issues, solid-phase reagents with a high specific affinity for endotoxin have been developed. These remove endotoxins from biological media, including water, buffers, cell culture media, and purified proteins. Affinity of the immobilized ligand for proteins and other growth substances is low.

Materials and Methods

Acticlean Etox, an endotoxin adsorbing ligand (Sterogene Bioseparations Inc. 5922 Farnsworth Ct Carlsbad, CA 92008 USA) on a stable, beaded agarose gel has become available and this was used in these trials to reduce the pyrogen level of the fetal serum in the laboratory. The process was then tested in production trials run under full scale manufacturing plant conditions.

Small scale trials were run in 12 mm Pharmacia columns using 5 mL of Acticlean Etox. Large scale runs were carried out using 2 liters of Etox Acticlean loaded in a 5 liter Amicon process scale column.

Acticlean Etox was pretreated and prepared prior to endotoxin removal as per the manufacturer''s instructions. The columns were regenerated with 5 column volumes or 1.0M NaOH (Sigma NaOH made up in endotoxin-free MilliQ water) at a flow rate equivalent to 125 mL/ min/liter of beads.

The regenerated beads were either left overnight in the 1.0M NaOH, or washed with 5 column volumes of sterile phosphate buffered saline (PBS)(Trace Biosciences Ltd, Sydney, Australia), and endotoxin removal attempted following the PBS wash.

For the small scale trials (5mL of Acticlean Etox) the serum, regeneration solutions and PBS were run through the column using gravity, and the flow rate was manually adjusted. For the large scale trials, the serum and washing solutions were pumped through the 2 liter column using a Master-Flex peristaltic pump. The serum was prefiltered by 0.5 micron filtration.

The endotoxin level was monitored using the Pyrotel LAL gel clot system (Cape Cod Associates, Ma, USA) against standard endotoxin positive controls, following heat treatment (2 min at 100 0C) after a 1/10 dilution of the serum with pyrogen-free saline.

Results: The results of the small scale column runs were summarized in Table 1. In trial #3 the beads were not regenerated, resulting in little endotoxin removal from that experiment.

Repeated use trials: Serum used in the small scale experiments was run through the 5mL column at an equivalent flow rate of 0.125L/liter Acticlean etox/min, using fresh regenerant and fresh serum with each cycle. Over 50 cycles were run with no significant reduction in capacity for endotoxin. These experiments also established the optimal endotoxin load for Acticlean Etox which was subsequently implemented in the large scale trials.

Large scale trials: The serum was prefiltered on a 0.5 micron Starclear filter from Pall-Filtron. the column used was a 140 mm diameter Amicon column packed with 2 liters of Acticlean Etox. The serum batch sizes varied from 47 to 60 liters. Eleven batches were processed in total. The total endotoxin bound per volume of resin increased with the endotoxin load, suggesting an equilibrium binding.

Cell culture data: A 100 mL sample of a cell culture tested batch of fetal bovine serum was run through a 5mL Acticlean Etox column. The initial endotoxin level was 130Eu/mL, and the final endotoxin level 7.5Eu/mL. When this batch of Acticlean Etox treated serum was tested against CHO, VERO and Sp2 cell lines, the results were within 5% of each other and of the control. This suggested that the Acticlean treatment did not change the cell growth promoting properties of fetal bovine serum.

Conclusions: Treatment with Acticlean Etox reduced the endotoxin level of the fetal bovine serum, provided the beads were adequately regenerated with 1.0M NaOH. In the small scale trial #3, the beads were not regenerated and the endotoxin removal was significantly lower. The flow rate of serum through the column must also be controlled. In large scale run #11 the flow rate was carefully monitored to ensure a flow rate of 0.125L/liter beads/min was maintained throughout, which led to excellent resin performance.

The optimal filtration configuration should include a Pall 0.5 micron profile filter followed by a sterile rated 0.2 micron final filter, followed by the Acticlean Etox column to polish the endotoxin below 10 EU/mL, or whatever value is targeted.

Courtesy: Trace Biosciences Ltd., Australia.