The association between antitumon activities and endotoxins, the major components of the cell walls of gram negative bacteria, was first reported in the late 1800s by Bruns [81 and by Coley [10] who described regression of certain tumors in patients either recovering from bacterial infections or in patients intentionally injected with mixtures of bacterial toxins. Similar observations were also reported in 1931 by Gratia and Linz and in 1943 by Shear and colleagues who demonstrated that endotoxins can induce hemorrhagic necrosis of certain transplanted tumors in mice [ 14,25]. Because endotoxins are not directly cytotoxic to tumor cells in vitro, it was generally believed by Algire [4] that the hemorrhagic necrosis was the result of endotoxin-induced hypotension leading to vascular, collapse ahd ischemia in.the tumor site.
The time which has elapsed since these observations were made has witnessed significant progress in our knowledge of the mechanisms involved in endotoxininduced tumor necrosis and the participation of cellular and humoral immune elements in the host response to tumor. In 1975, Carswell and colleagues showed that the sera of mice infected with Bacillus Calmette-Guerin (BCG) and subsequently treated with endotoxin, contain a substance which, independent of endotoxin, induces tumor necrosis [9]. They calledthis substance tumor necrosis factor (TNF) and advanced the hypothesis that endotoxin-induced tumor necrosis is mediated by the release of TNF from activated macrophages.
Although it has been shown that endotoxins may be useful agents to nonspecifically stimulate host immune defense against tumors [6], their clinical applications for the treatment of human malignancies has not been extensively explored because of their significant toxicity. Furthermore as demonstrated recently by Ribi and colleagues and by Bloksman and colleagues [7,20.1 the detoxified endotoxins have only minor antitumor activities.
The availability of large quantities of purified TNF therefore was an essential requirement in order to study the role of this lymphokine, independent of endotoxins, in tumor necrosis. Using the human promyelocytic cell line HL-60, Aggarwal and colleagues described the production, purification to homogeneity, and characterization ofa cytotoxic protein which they designated human TNF [2].