dose of TNF in both Surgery Branch, NCI and other studies is approximately 8 
ug/kg/day. Thus when injected intravenously only 2% of the TNF dose required 
to mediate antitumor effects in the mouse can be administered to man. 
Because of the unique effectiveness of TNF in the treatment of a variety 
of murine malignancies, we have sought means to selectively increase the local 
concentration of TNF at the tumor site. Because TIL can traffic directly to 
tumor deposits and concentrate at those sites we hypothesized that TIL 
transduced with the gene for TNF and producing large amounts of TNF in the 
local tumor microenvironment might have substantially increased antitumor 
effects compared to normal TIL. This hypothesis served as the basis for the 
protocol dealing with the administration of TIL transduced with the gene for 
TNF that is now underway. 
In the course of these studies we used retroviral vectors to introduce 
the gene coding for TNF into murine and human tumors. These transduced tumor 
cell lines produced up to 15 ng TNF/10 < cells/24 hours. An example of the 
production of TNF over a two month period by high and low producer tumor lines 
from a sarcoma in C57BL/6 mice is shown in Figure 2. Extensive studies 
introducing the TNF gene construct into human melanoma cell lines were 
similar. These transduced cells contained one to two copies per cell of an 
unrearranged TNF vector genome and expressed transcripts homologous to the TNF 
cDNA corresponding to the proviral transcribed full length message (see Fig. 
3). The human tumor cell lines produced up to 15 ng TNF/10 6 cells/24 hr. 
Extensive experiments were performed to test the immunogenic ity of these 
tumor lines in both syngeneic mice as well as in nude mice. In murine models 
we found that unmodified parenteral tumor cells grew aggressively when 
implanted subcutaneously into syngeneic mice, although tumor cells transduced 
with the TNF gene regressed in a significant number of animals after an 
initial phase of growth. An example of this phenomenon is shown in Fig. 4. 
This effect correlated with the amount of TNF produced and could be blocked 
with a specific anti-TNF antibody (see Fig. 5). The regression of these TNF 
producing cells was not associated with any demonstrable toxicity in the mice 
[62] 
Recombinant DNA Research, Volume 15 
