sites (see Table 6). In three other experiments, TIL from murine tumors in 
the liver were less effective than TIL from cutaneous sites. 
In the human, other workers have shown that small cutaneous tumor auto 
inoculations can provoke significant immune responses. Hoover and Hanna have 
published work in which colorectal cancers from primary sites were irradiated 
and utilized for autologous immunization of patients in the adjuvant setting 
(22). Patients with Dukes B2 and C tumors demonstrated improved overall 
survival in a randomized study, presumably due to an effective immune response 
to the immunization. In addition, Berg and Mastrangelo have investigated the 
immunization of melanoma patients with irradiated autologous tumor, and have 
demonstrated the induction of significant T-cell infiltrates in tumors as well 
as rare clinical responses in patients with metastatic disease (23,24). These 
studies generally utilized irradiated tumor cell inoculation and therefore 
result in no tumor for the generation of TIL. The presence of tumor is vital 
for generating optimal TIL in that T-cells separated from fresh tumors will 
grow in vitro with IL-2 but will show decreased in vivo efficacy if not re- 
exposed to tumor in culture (presumably due to the requirement of cultured T- 
cells for antigen exposure) (25) . Furthermore, irradiated or non-viable tumor 
and tumor extracts produce immune responses in animals that are typically 
inferior to the responses to viable tumor (26). Therefore the injection of 
small amounts of viable tumor at a cutaneous site might not only result in 
tumor for TIL production, but also generate an immune response superior to 
that demonstrated using irradiated or non-viable tumor. This immune response 
may not only be seen at the tumor site, but pre-clinical models show that T- 
cells can be recovered in lymph nodes draining the site of tumor inoculation, 
which can be expanded in culture and show in vivo anti-tumor activity. This 
aspect of the cellular immune response to tumor immunization will be discussed 
later in the protocol. 
Because tumor growth at the transplant site is necessary for TIL 
production, it is important to know if that will occur and what are the 
potential risks involved. Southam and Brunschwig inoculated a series of 
[ 110 ] 
Recombinant DNA Research, Volume 15 
