Faculty Bibliography

Using variable-length deletion constructs of the 5'-flanking region of the human interleukin-6 (IL-6) gene linked to the chloramphenicol acetyltransferase gene, we showed that the region from positions -109 to -50 mediated the bulk of the response to tumor necrosis factor (TNF) or interleukin-1 (IL-1), while it was less responsive to forskolin. DNA mobility shift assays and DNase I footprinting analysis identified a nuclear protein from TNF- or IL-1-treated fibroblasts that bound to a region comprising a kappa B-like element located between positions -72 and -63 on the IL-6 gene. On the basis of these and other experiments, we conclude that TNF and IL-1 apparently activate IL-6 gene expression by closely related mechanisms involving activation of a NF-kappa B-like factor, whereas the pathway of IL-6 induction by forskolin is, in part, different

A lambda cDNA library was prepared from polyadenylated RNA isolated from quiescent human diploid FS-4 fibroblasts stimulated with tumor necrosis factor for 3 h. Differential screening was used to isolate cDNA sequences that are stimulated by tumor necrosis factor. Eight distinct tumor necrosis factor-stimulated gene sequences (designated TSG-1, -6, -8, -12, -14, -21, -27, and -37) were partially sequenced and compared with known sequences from GenBank. TSG-1 was identical to the gene for interleukin-8. TSG-8 corresponded to the gene for monocyte chemotactic and activating factor. TSG-21 and -27 were identical to the genes for collagenase and stromelysin, respectively. The other four sequences showed no homologies with known genes. Patterns of induction of mRNAs corresponding to the eight cloned cDNAs by various cytokines, growth factors, and activators of second messenger pathways were analyzed in FS-4 cells

Incubation of the human U937 histiocytic lymphoma cell line with granulocyte-macrophage colony stimulating factor (GM-CSF) rendered the cells responsive to induction of TNF by LPS. Treatment with IL-6 reduced TNF production in GM-CSF-primed U937 cells. The inhibitory effect was most pronounced (approximately equal to 80%) when IL-6 was added either along with GM-CSF or within the first 3 h of GM-CSF treatment. Both GM-CSF or IL-6 inhibited [3H]TdR uptake in U937 cells, and simultaneous treatment with GM-CSF and IL-6 resulted in an additive inhibitory effect on cell proliferation. However, the inhibition of TNF production could not be explained by the inhibitory effect of IL-6 on cell growth, nor was it due to a reduction in cell viability. An inhibition of TNF production by IL-6 was also demonstrated in cultured human peripheral blood monocytes. Treatment with IL-6 also resulted in a dose-dependent reduction of the 17-kDa TNF band revealed by SDS-PAGE after labeling monocytes with [35S]cysteine and immunoprecipitation with anti-TNF mAb. In addition, treatment with IL-6 resulted in a reduction of monocyte in vitro cytotoxicity for tumor target cells. Finally, in mice sensitized by the administration of Bacillus Calmette-Guerin, the injection of IL-6 significantly reduced the levels of TNF found in the serum upon challenge with LPS. Inasmuch as TNF is known to be an inducer of IL-6, the inhibitory action of IL-6 on TNF production may represent the negative arm of a regulatory circuit. The inhibitory action of IL-6 on TNF production is consistent with a predominantly antiinflammatory role of IL-6 in the intact organism

Nuclear protein IRF-1 (interferon regulatory factor 1) was earlier shown to bind to cis-acting regulatory elements present on interferon (IFN)-alpha/beta genes and some IFN-inducible genes. Here we show that in both human FS-4 and murine L929 cells, steady-state levels of IRF-1 mRNA were increased by treatment with tumor necrosis factor (TNF), interleukin 1 (IL-1), poly(I).poly(C), or IFN-beta. IRF-1 mRNA induction was also demonstrated in cells treated with calcium ionophore A23187 or with phorbol 12-myristate 13-acetate, but not with epidermal growth factor, dibutyryl-cAMP, or the adenylate cyclase activator forskolin. To determine whether stimulation of IRF-1 mRNA levels correlates with IFN-beta induction, we compared IRF-1 and IFN-beta mRNA levels in cells exposed to various stimuli. In L929 cells, treatment with poly(I).poly(C) under conditions that failed to induce significant levels of IFN-beta mRNA led to a very low induction of IRF-1 mRNA, but 'priming' cells with IFN prior to the addition of poly(I).poly(C) greatly increased both IRF-1 and IFN-beta mRNAs. In FS-4 cells an increase in IFN-beta mRNA (examined by the polymerase chain reaction) was seen after treatment with TNF, IL-1, A23187, or poly(I).poly(C), but not with IFN-beta, epidermal growth factor, dibutyryl-cAMP, or forskolin. Thus, all treatments that increased steady-state levels of IFN-beta mRNA also enhanced IRF-1 mRNA levels. However, treatment with IFN-beta, which caused a marked stimulation in IRF-1 mRNA, failed to produce a detectable increase in IFN-beta mRNA. It appears that IRF-1 may be necessary but not sufficient for IFN-beta induction. The ability of TNF and IL-1 to increase both IRF-1 and IFN-beta mRNAs may be responsible for some similarities in the actions of TNF, IL-1, and the IFNs

In addition to its cytotoxic/cytostatic action on many tumor cells in vitro, tumor necrosis factor (TNF) was recently shown to stimulate the growth of some types of cells in culture. We examined the action of TNF in BALB/c 3T3 cells which have been used extensively to study growth regulation. In subconfluent, rapidly dividing 3T3 cultures, murine (Mu) TNF was cytotoxic, while human (Hu) TNF had virtually no antiproliferative action on the cells. In contrast, in density-arrested BALB/c 3T3 cells maintained in a chemically defined, serum-free medium, MuTNF produced a dose-dependent stimulation of DNA synthesis. In stimulating DNA synthesis, MuTNF acted synergistically with both epidermal growth factor or platelet-derived growth factor. While stimulating DNA synthesis in quiescent 3T3 cultures, high doses of MuTNF (100 or 10 ng/ml) were also cytotoxic for a portion of the cells in the same cultures. Cytotoxicity was apparent 2 h after the addition of MuTNF, well before the onset of DNA synthesis. BALB/c 3T3 cell variants selected for their resistance to the cytotoxic action of MuTNF retained the capacity to respond to the mitogenic action of MuTNF, indicating that the stimulation of DNA synthesis by TNF is not a consequence of a TNF 'wounding effect.' Addition of TNF to density-arrested 3T3 cells resulted in the release of free arachidonic acid and palmitic acid from the cells. Quinacrine, a phospholipase inhibitor, inhibited both cytotoxicity and DNA synthesis in response to TNF, and melittin, a phospholipase activator, mimicked both the cytotoxic and mitogenic actions of TNF in quiescent BALB/c 3T3 cells. These results suggest that phospholipid breakdown is part of the essential early signal transduction events required both for the cytotoxic and mitogenic response to TNF action