Optimization of the L-M cell bioassay for quantitating tumor necrosis factor alpha in serum and plasma.

In attempting to quantitate TNF-alpha in body fluids using current bioassay protocols, we discovered several factors which adversely affect reliability, sensitivity and specificity. The objective of this study was to establish an optimum assay for use with serum and plasma samples. While adopting the commonly used L-M cell bioassay to test serum and plasma samples, we noted non-specific staining of cell debris and protein by crystal violet dye, even if the wells were washed prior to staining. However, when we replaced crystal violet with tetrazolium salts to measure cell viability, we discovered that both serum and plasma, from a variety of species, non-specifically reduced both MTT and XTT tetrazolium salts to a colored formazan product resulting in absorbance values significantly higher than those of medium and serum controls. This effect was particularly pronounced with fetal bovine serum which showed significant color development with as little as 6% serum in the test wells. Maximum sensitivity can only be obtained by eliminating these false negative readings. Therefore the serum or plasma containing supernatant must be removed from the test wells and replaced with fresh serum-free medium prior to addition of the substrate. This finding should be applicable to other body fluids such as urine, milk, and synovial fluid as well as any tetrazolium based assay for cell viability which uses serum-supplemented culture medium. Additionally, when the substrate XTT, which reduces to a water-soluble formazan product, was compared to MTT which reduces to a water-insoluble product, XTT was found to be superior since elimination of the solubilization process resulted in reduction of assay time. Also, overall absorbance readings using XTT were higher than with MTT, without any loss of sensitivity. There were no differences in detectability of TNF-alpha between serum and plasma and use of preservative-free heparin was found not to have adverse effect on TNF-alpha assay. Heat treatment of both serum and plasma seemed to inactivate factors that contributed to the non-specific lysis of the L-M cells when their concentrations exceeded 25%.