July 30, 1990, Human Gene Therapy Subconmittee 
He said he also had some minor concerns about the informed consent form. 
Dr. Parkman said he had some concern as to whether local production of TNF would 
actually down-modulate the indigenous capacity of TIL cells to kill their appropriate 
targets, and whether the efficacy of the TIL cells was affected by introduction of the 
TNF gene. 
Dr. Leventhal said she wanted to see some dose-response data on decreases in toxic 
effects due to lowering of IL-2 doses and the relationship to toxic effects of the TILs 
themselves. She said it was important to know what effects are due to the TNF and 
what are due to the TIL plus the IL-2. Further, she wanted to know what methods were 
envisioned to ameliorate chronic TNF toxicity as a result of the TIL infusion should it 
occur. 
Dr. Walters called on Dr. Rosenberg to respond to the questions presented by the 
subcommittee. Dr. Rosenberg said that the protocol was an attempt to improve upon 
adoptive cellular therapy which had begun with LAK cells (treating almost 200 patients) 
and now with TILs, which had demonstrated that almost half the patients show an 
objective regression in their cancers, albeit of a limited duration. He said that one 
obvious way to improve upon this result is to try to alter the TIL to improve their anti- 
cancer properties. Dr. Rosenberg said it had been shown in previous studies that TIL 
traffic to tumor deposits and concentrate there, and further that indium^^^ studies have 
allowed for calculation of how many cells get to the tumor site. He said that calculations 
have shown that TIL can be used as a potential vehicle to deliver TNF to tumors in 
sufficient quantity to improve antitumor effects. 
Dr. Rosenberg said TNF was selected as the first gene to be put into TIL on the basis of 
extensive animal data from experiments, looking at biologic and chemotherapeutic effects 
of TNF in mice. It was found to be the most effective agent for causing tumor 
dissolution in the murine model, with dissolution of tumor occurring in 15-30 minutes 
after administration. He said these effects were probably due to effects of TNF locally 
which affect the vasculature and injure recirculation, possibly due to local formation of 
oxygen-free radicals, resulting in tumor reduction. 
Dr. Rosenberg said that despite these successful animal experiments, TNF administration 
in humans had been an abysmal failure. Injection of TNF into 39 patients by both 
continuous and bolus infusion at maximum tolerated levels failed to produce antitumor 
effects. It is thought that the reason for this is a difference in the ability to achieve local 
concentrations of TNF in the mouse and the human. He noted that it was possible to 
inject inbred mice bred to endure endotoxin-mediated infections at a level of 400 mg/kg. 
Therefore, they tolerate TNF much better than humans, in whom the maximum tolerated 
dose by i.v. or bolus infusion is 8-10 mg/kg per day. However, if a high enough 
concentration of TNF capable of mediating tumor reduction could be achieved at the 
[188] 
Recombinant DNA Research, Volume 14 
