Mechanism of TNF-α cytotoxicity in a leukemia virus transformation model

Abelson murine leukemia virus (A-MuLV)-induced transformation was investigated to determine whether cells not sensitive to TNF-α could be made sensitive to the cytolytic action of TNF-α when infected with this retrovirus. Mouse embryonic fibroblast cell line CL.7 was found to be relatively insensitive to TNF-α. Upon transformation with A-MuLV, these cells gave rise to a clone (3R.1) which was found to be insensitive to TNF-α and another clone (6R.1) which had an increased sensitivity to TNF-α. The differential cytotoxicity was observed when cells were treated with TNF-α, for 18 hr, at 0 to 100 units/ml, at 37°C.

The mechanism of this differential cytotoxicity was further investigated. Thus, TNF-R levels on the cell surface were found to be not correlated with the differential TNF-α response. The A-MuLV transformation suppressed the epidermal growth factor-receptor (EGF-R) in 3R.1 clone and induced its levels significantly in the 6R.1 clone (p<0.05). Cell surface EGF-receptor (EGF-R) levels in CL.7 and 3R.1 clones were lower than the 6R.1 clone (p<0.05). Although the EGF-R levels in all the clones were induced with TNF-α, the expression of EGF-R correlated with the susceptibility to TNF-α.

The role of antioxidants, such as α-tocopherol and β-carotene, (known anti-cancer agents) in modulating TNF-α-induced EGF-R expression was investigated. In both the untransformed and the transformed clones, f-carotene suppressed the constitutive and the TNF-α induced EGF-R levels whereas α-tocopherol was found to have an enhancing effects. Studies with metabolic inhibitors on TNF-R and EGF-R expression indicate that inhibitors of the arachidonic acid cascade and modulators of protein kinase-C (PK-C), could influence the binding and internalization of TNF-α and thereby controlling the physiologic future of the cells.

The A-MuLV specific V-abl protein, p120, tyrosine phosphorylation was determined by a radio-labelled anti-phosphotyrosine antibody in an antigen capture assay. TNF-α had little effect on p120 phosphotyrosine levels of TNF-α insensitive CL.7 and 3R.1 clones. The, TNF-α sensitive, 6R.1 clone, however, was found to induce its p120 specific phosphotyrosine upon exposure to TNF-α for 8 hr. Thus, TNF-α modulated the tyrosine phosphorylation of p120 only in the TNF-α-sensitive cell line.

The mitochondrial toxicity of TNF-α was determined by monitoring the rate of quenching of a cationic spin probe CAT 16. Mitochondrial preparation from CL.7 and 3R.1 clones had higher ability to quench CAT 16 signal with TNF-α incubation time than mitochondria from the 6R.1 cells. This indicates that the differential TNF-α cytotoxicity manifested in A-MuLV transformed clones may, in part, be due to the differential mitochondrial toxicity of this cytokine.

The hypothesis that TNF-α cytotoxicity was mediated via an oxidative process was tested on the TNF-α sensitive L929 cells. Using a flow cytometric detection system it was determined that TNF-α produced intracellular hydrogen peroxide in these cells which was sensitive to concentration and incubation time of TNF-α. Superoxide radicals were also generated during TNF-α action on L929 cells, as determined by the use of the spin trap PBN in conjunction with EPR spectroscopic techniques. The PBN-OOH spin adduct spectrum peaked at 9 hr of TNF-α incubation and was inhibitable upto 30 % with 10 µM of desferral-Mn complex (a known SOD mimic). These data indicate that superoxide and hydrogen peroxide are common events in TNF-α dependent cell killing process.

The differential TNF-α cytotoxicity was found to depend on differences in the antioxidant status of the target clones. Thus, it was found that Cu/Zn-SOD, Mn-SOD, GSH-Peroxidase and GSH-Reductase enzymes were all induced significantly in the CL.7 clone (p<0.05) upon incubation with 100 units/ml of TNF-α for 18 hrs. TNF-α had little effect on the antioxidant enzymes of both 3R.1 and 6R.1 cells. However, the constitutive levels of most antioxidant enzymes were found to be higher in 3R.1 cells than in the 6R.1 cells. Therefore, the susceptibility of 6R.1 to TNF-α may, in part, be due to a low level of antioxidant enzymes present in this clone.

In conclusion we found that the differential cytotoxicity of TNF-a may, in part, due to: (1) differential EGF-R expression, (2) differential mitochondrial cytotoxicity, and (3) differential ability to modulate the tyrosine phosphorylation in untransformed and A-MuLV transformed cells and (4) differential antioxidant status of these cells to handle oxidative stress imposed by TNF-α.