antibodies. Several workers have attempted to produce active immunization by subcutaneous injection 
of autologous glioma cells or cultured glioma cell lines (28). In this study 20 patients were injected with 
irradiated human glioma cell lines (not autologous) along with 500 mg of bacille Calmette-Guerin cell 
wall. A trend to increased survival with inoculation was described. Importantly, no allergic 
encephalomyelitis resulted (29). 
Previous Work with IGF-I and Glial Tumors Many tumor types express high 
amounts of IGF-I or the related growth factor IGF-II (30). The expression of IGF-I and IGF-II genes is 
markedly enhanced in glial tissue from glioblastoma and anaplastic astrocytoma as compared to 
normal adult brain (31). This over-expression of selective growth factors including IGF-I and II has 
been implicated in uncontrolled growth and maintenance of the tumor phenotype. (32,33). The 
expression of IGF-I and II is also high in fetal brain. Thus the expression in tumors of genes that are 
active during embryonic or fetal development could represent re-enactment of an early embryonic or 
fetal program, enabling the cells to in some way evade immune recognition (34). Most data bearing on 
this point have been obtained in reference to IGF receptor status. IGF-I binding was demonstrated by 
competition-inhibition binding assays in plasma membrane preparations from meningiomas, 
glioblastomas, and normal adult brain. High affinity 125IGF-I binding was greatest in gliomas and in 
meningiomas,and minimal in normal adult brain (35). Immunohistochemistry staining using mouse 
monoclonal antibodies to IGF binding protein has demonstrated this protein intracellularly in 4 of 6 
gliomas (36). Scatchard analysis demonstrated increased specific IGF-I binding to cell membrane 
preparations from 18 gliomas (33). 
Other experimental data address the relative expression of IGF genes. We and other 
have shown that the rat C6 glioma expresses high levels of IGF-I, produces large quantities when 
grown in serum-free medium, and forms rapidly growing tumors in syngeneic animals. In addition, 
when examined by immunocytochemical techniques, high levels of IGF-I expression were detected in 20 
of 20 fresh frozen human glioma samples (Ilan, unpublished observation). 
In studies by others the expression of both IGF I and IGF II were increased in comparison 
to normal brain tissue in one series of 10 freshly obtained human astrocytomas including 2 grade IV 
glioblastomas (51), and in a second series of 4 brain cancers including 3 glioblastomas and 1 anaplastic 
astrocytoma (31). However, in a study that evaluated relative transcription rates for a number of 
growth factors in cultured cells, many of which were established cell lines, results were less clear. Both 
IGF I and IGF II were highly expressed in some glioblastoma cell lines but a uniform pattern of 
overexpression was not obtained (52). The results are consistent with findings by Glick et. al. who 
reported no significant production of IGFs by specific radioimmunoassay in tumor cell conditioned serum 
free media of 12 glioblastomas (53). However, in our experience the cellular characteristics of cultured 
glioblastoma cells differ significantly after several cell passages in comparison to those of fresh cells. 
Moreover, IGF protein is not secreted and/or not retained in stable form within the cell conditioned 
media, but rather accumulates within the cell (see Fig 3, ref 34). 
Previous studies have looked for a role of IGF-I and IGF-II in driving proliferation of 
tumor cells. Growth in animal models has been blocked with specific antibodies to IGF'(37,38). We 
chose to block IGF-I expression through the introduction of antisense IGF-I cDNA into C6 rat glioma 
cells. An episome-based plasmid, pAntiIGF-I, was constructed that incorporates Epstein-Barr virus 
replicative signals and the ZnS04-inducible metallothionein I transcriptional promoter. (See 
Attachments H, I and J) The transfectants express high levels of IGF-I mRNA and protein when 
cultured in absence of ZnS04. Addition of ZnS04 to culture media resulted in high levels of antisense 
transcripts and a dramatic decrease in level of endogenous IGF-I mRNA and IGF-I protein within these 
cells. The subcutaneous injection of these transfected glioma cells into syngeneic rats brought about the 
induction of an immune response against the transfected glioma cells. Moreover, the induced CD8 T 
lymphocytes were capable of recog nizin g and destroying exisbnc tumor even when the transfected cells 
were injected at a point distant from the tumor, whether localized intra or extra-cranially. We 
concluded that this IGF-I antisense strategy may reverse a phenotype which allows C6 glioma cells to 
evade immune surveillance (30). 
Recombinant DNA Research, Volume 18 
[133] 
