July 30, 1990, Human Gene Therapy Subconmittee 
tumors? 
Dr. Rosenberg said this was one of the original problems they considered in thinking 
about the TIL protocols. However, a very stringent procedure was devised, and the FDA 
approved a set of criteria whereby any cells that are to be infused are carefully examined 
by a cytologist using melanoma-specific antibodies. If there is the trace of even a single 
tumor cell, then the cells are not administered. Further, in light of the extremely heavy 
burden of tumor cells in the patients who will receive the TILs, the number of any tumor 
cells that would escape detection would be at most on the order of 1/100,000 of the 
normal tumor burden. 
Dr. Mahoney questioned the occurrence of a cancer in a mouse that received retroviral 
supernatant intraperitoneally. Dr. Rosenberg said that Dr. Anderson had studied the 
tumor and that it did not contain any retrovirus. Since it was an old mouse which has a 
3-5% incidence rate of spontaneous cancer, it was deemed not to have been as a result 
of the administration of the retroviral supernatant. 
Dr. Rosenberg said he wished to address Dr. Leventhal's questions as to the number of 
patients per dose level and how often measurements for toxicity would be taken. He 
said page 154 of the protocol contains a schedule of measurement of each of the toxicity 
determinants. He noted that each patient is his own control and therefore if any patient 
achieves intolerable toxicity, at any given dose, the escalation would be stopped. Dr. 
Leventhal said she did not see any form for annotation of this information. Dr. 
Rosenberg said that a toxicity form is filled out daily on all patients and that a nurse 
would enter the dosage received by each patient. 
Dr. Leventhal asked what the stop criteria would be for dose escalation. Dr. Rosenberg 
said that any grade 4 toxicity that were not reversible by appropriate medications would 
be cause for cessation of escalation. 
Dr. Epstein said that the "suicide tumor cells" in the murine model are interesting. 
However, he asked if a better model would involve injecting tumor cells into a nude 
mouse and then administering TILs to mirror the human experiment. Dr. Rosenberg 
said that Drs. Mule and Jacob have tried to do this in nude mice, BNX mice and SCID 
mice with no success, due to inability to achieve appropriate trafficking of human cells in 
the mouse. He suggested there may be endothelial recognition mechanisms which will 
not allow the human cells to traffic in the mouse. He added that such a model would be 
helpful in evaluating TIL subpopulations of gene-modified TIL. 
Dr. Epstein said he was surprised to see the protocol being described as a Phase I study 
and asked where Phase I would end and Phase II would begin. Dr. Rosenberg said that 
"Phase I" was a term denoting a trial for determining maximum tolerated doses of a 
therapeutic reagent through dose escalation. He said the design of this study was a 
Recombinant DNA Research, Volume 14 
[197] 
