Preclinical Studies I nvolving Irradiated. Genetically Modified 
Tumor Cells 
The decision to employ irradiated tumor cells expressing GM-CSF for our 
clinical protocol was based upon collaborative studies between Dr. Mulligan's 
laboratory at the Whitehead Institute and M.I.T. and Dr. Pardoll's laboratory 
at Johns Hopkins. One rationale for these studies was to assess the relative 
potency of the many different molecules previously identified as possessing 
immuno stimulatory activity, and to identify new gene products capable of 
inducing systemic antitumor immunity. No previous study had compared the 
activity of different gene products in a single tumor model, nor examined the 
activity of a single gene product in multiple tumor models. A second 
rationale was the need to understand, in a general way, some of the mechanisms 
underlying the immunostimulatory activity of cytokines and other gene 
products . 
The important finding that irradiated, transduced tumor cells were as 
capable of inducing potent systemic anti-tumor immunity as live, transduced 
cells was made through an early series of experiments in which the effects of 
the expression of a variety of different cytokines and adhesion molecules upon 
the tumorigenicity of the mouse melanoma cell line, B16, was examined. To 
facilitate the transfer of a variety of gene products into tumor cells, a 
recombinant retrovirus vector, MFG, was utilized (see Appendix 14. B) . Studies 
from Dr. Mulligan's laboratory had shown that this vector consistently yielded 
high titers of helper-free recombinant virus, and high levels of the desired 
gene product in infected cells (Appendix 14. A. 3). Ten different molecules 
were examined for their effect upon the tumorigenicity of B16 cells. These 
included IL-2, IL-4, IFN-gamma, and TNF-alpha, as well as a number of gene 
products not previously examined (IL-5, IL-6, IL-IRA, CD2, ICAM-1, and GM- 
CSF). Of the ten molecules tested, only IL-2 promoted the complete rejection 
of B16 cells. Several of the other molecules caused a partial delay in tumor 
growth. Expression of a number of other gene products, including TNF , IL-6, 
and GM-CSF, actively induced systemic toxicities, due to the progressive 
growth of tumor cells expressing those products. 
Although IL-2 transduced B16 cells were effectively rejected, they did 
not protect animals from a subsequent challenge with non-transduced cells. 
Since only IL-2 expression resulted in rejection of B16 cells, challenge 
experiments were next performed using cells doubly infected with an IL— 2 
expressing virus and an additional construct. It was through this analysis 
that GM-CSF was first identified as a powerful inducer of systemic anti-tumor 
immunity. Since B16 cells expressing GM-CSF alone had been shown to grow 
progressively, it was possible that in the case of cells expressing both IL-2 
and GM-CSF, IL-2 expression served primarily to cause rejection of the 
vaccinating cells. This was shown to be the case, since B16 cells that were 
expressing GM-CSF alone, but irradiated with 3500 rads to prevent their 
outgrowth in vivo, stimulated potent systemic anti-tumor immunity (Figure 1A) . 
Non-transduced, but irradiated B16 cells showed little or no stimulation of 
Recombinant DNA Research, Volume 17 
[293] 
