Caspase-3 and -8 activation and cytokine removal with a novel cellulose triacetate super-permeable membrane in an in vitro sepsis model

Pro-apoptotic molecules are generated during sepsis which may be responsible for alteration of organ function in sepsis. Removal of systemic apoptotic activity may affect recovery from sepsis. Current high flux membranes might not be sufficiently permeable to eliminate pro-apoptotic factors. We evaluated the elimination of pro-apoptotic factors induced by LPS in human whole blood by a super-permeable cellulose triacetate membrane (SUREFLUX FH 150, Nipro, Osaka, Japan ) in comparison to a standard high flux cellulose triacetate membrane (UT 700, Nipro, Osaka, Japan) and a polyethersulfone plasmafilter (Bellco, Mirandola Italy) in an in vitro blood circulation. We spiked human whole blood with lipopolysaccharide from Escherichia coli (Serotype 026-86, 10 mg/ml), incubated it for 3 hours to allow cytokine generation and recirculated it at 300 ml/min for 3 hours. The UF line was first returned to the blood module at 10 min. After this, the UF was drained from 10 to 60 min at a rate of 1000 ml/h. Zero balance was obtained by re-infusion of bicarbonate buffered hemofiltration fluid. Apoptosis was assessed on U937 monocytes (incubated with plasma or ultrafiltrate) by fluorescence microscopy dyes (Hoechst 33342, propidium iodide) and annexin V flow cytometry. Caspase-3 and Caspase-8 activity was assessed on the recirculated blood monocytes by spectrophotometric methods. IL-2, IL-10 and TNFα were determined by commercially available ELISAs. Sieving coefficients and clearances were determined for the different cytokines. Caspase-3 and Caspase-8 were activated by LPS and remained either stable or increased during in vitro circulation. Apoptosis activity of U937 cells, when incubated with the ultrafiltrate, increased in parallel with arterial plasma values (for Uf: UT700=23.1%; Sureflux FH150=42.5%). However, by 60 min the apoptotic activity recorded with the ultrafiltrate was reduced to the levels of arterial plasma (for Uf: UT700=19.8% ; Sureflux FH150=11.2%). Sieving coefficients in the super-permeable membrane were significantly higher for all measured cytokines in comparison to the standard high flux membrane (e.g. TNFα 0.72 vs 0.03 p<0.001) and close to the values observed for the plasmafiltration membrane. Nevertheless protein losses measured by albumin leakage were much lower with the Sureflux filter in comparison to the plasmafilter. In conclusion, pro-apoptotic factors can be eliminated by dialytic membranes with the removal rate maximized by using super high flux dialysers which may represent a compromise between hemofiltration and plasmafiltration membranes.