Effect of anticoagulants and heat on the detection of tumor necrosis factor in murine blood.

Assays for tumor necrosis factor (TNF) may be inhibited by nonspecific factors present in body fluids. We found that the ability to quantitate TNF is greatly affected by blood processing methods. Mice were anesthetized by inhalation of methoxyflurane before obtaining blood by cardiac puncture. The blood from a group of mice was allowed to clot before recovery of serum. Plasmas were obtained from three other groups of mice after collection of blood into tubes containing EDTA, sodium citrate, or preservative-free heparin. The pooled serum or plasmas were spiked with rhTNF to a final concentration of 1 microgram/ml and aliquoted for frozen storage. The serum and plasmas were divided into heated (56 degrees C for 30 min) and unheated portions prior to a standard L929 cytotoxicity assay. Comparison of absorbances at 595 nm after crystal violet staining of cells revealed differences in detection of TNF in plasmas compared to serum and in heated compared to nonheated samples. Citrated plasma clotted in the assay at dilutions at or below 1:25. EDTA plasma consistently produced unexplained lysis of L929 cells in both heated and unheated unspiked samples. We conclude that TNF levels should be determined only in heat treated serum samples, and that comparisons be made against both a TNF standard and a TNF standard prepared in normal homologous serum that can be heated and assayed in parallel with the test samples.