Recombinant DMA Advisory Committee - 2/4/91 
Dr. Rosenberg said that tests in mice had shown TNF to be remarkably effective as an 
antitumor agent. However, in a study of 39 patients receiving intravenous TNF, as well 
as in over a dozen other studies performed around the world with TNF in humans, no 
antitumor effects were shown at all with TNF. A possible explanation for this is that 
humans cannot tolerate a large enough dose of TOT to produce the antitumor effects. 
Humans can tolerate a systemic injection of only 8-10 mg/kg of TOT, whereas the effects 
in mice were shown with a dose of 400 mg/kg. Therefore the TNF^jl protocol is an 
attempt to bypass the toxicity to normal cells and allow for a high accumulation of TOT 
at the localized tumor site. 
Dr. Rosenberg said that preclinical evidence for the possibility of this experiment 
succeeding was provided by transducing the TOT gene into mouse and human tumor 
cells and then transferring these into syngeneic mice or human tumors which 
spontaneously regressed due to local accumulation of TNF being produced by the 
tumors. A further study was done at the behest of the NIH Institutional Biosafety 
Committee that showed that anti-TNF antibody could abrogate the regression of the 
tumor, thus demonstrating indeed it was the TNF production by these tumor cells that 
led to the regression. Studies also were performed showing that local injection of TOT 
into tumors could lead to tumor regression. 
Dr. Rosenberg said a two-gene construct was developed for the protocol using the TNF 
gene promoted by the retroviral LTR and the neomycin resistance gene with the SV-40 
promoter, although the initial protocol did not call for selection of these cells. Studies 
were performed which showed that when these genes were inserted into TILS, secretion 
of TNF by these transformed cells was increased 100-fold. 
Dr. Rosenberg said that TNF was very toxic, causing severe hypotension and low blood 
pressure in patients receiving it intravenously in high doses. Therefore, the investigators 
were cautious in selecting an appropriate starting level of TNF which was calculated to 
be .07 mg/kg based on a 70 kg human receiving 10^° TNF-transduced TILs. It had been 
shown previously that a human can withstand a dose of 8 mg/kg of soluble TOT without 
toxic consequence, but since TILs can accumulate in liver and other organs, this 
technique may result in different local toxicity. Further, humans had previously been 
exposed to 1.2 mg/kg of TOT as a result of LAK cell therapy. Therefore, the chosen 
level was significantly lower than that already shown to be safely tolerated in humans. 
Dr. Rosenberg said the protocol calls for infusion of escalating doses of transduced TILs 
at 3-week intervals in combination with IL-2 given at 180,000 iu/kg every 8 hours. The 
dose escalation then would be escalated to reach 2-3 X 10^^ cells, at which level one 
would predict enough TNF would be produced at the tumor to cause tumor regression. 
Subsequent stages of the protocol would then call for administration of selected cells and 
then higher doses of IL-2. 
[456] 
Recombinant DNA Research, Volume 14 
