as tolerated to a maximum of 10 mg every four hours. If the patient should continue to deteriorate, or if 
impending transtentorial herniation is evident, decompressive craniotomy may be offered to the 
patient. 
As with any invasive procedure, infection is a risk. Patients with glial tumors are often 
immunocomprorruzed by their tumors. In addition they may have increased susceptibility to infection 
because of chronic glucocorticoid administration. The injection process itself carries little risk of 
infection. However contamination of cells during culture, passage or transfection will be carefully 
avoided by extensive precautions against contamination of cultured material. Tumor cells will be 
harvested in the operating room at University Hospitals of Cleveland and immediately placed in 
culture media in a sealed sterile vial. Samples will be transported to the culture lab which is 
dedicated exclusively to this project. There, under a Category 2 laminar flow hood, samples will be 
minced and dissociated as described. Glioma cultures will be maintained without antibiotics, and will 
be screened regularly by culture for bacteria, mycoplasma and fungi. 
Prior to injection, transfected tumor cells will be irradiated. Therefore growth of transfected 
cells at the site of local injection should not occur. Bloom injected untransfected human glioma cells 
under the skin of the same patients from whom they had been obtained (50). Ten of 12 patients 
developed tumor growth at the site of injection. However Mahaley's group showed that untransfected 
glioma cells injected with BCG caused an occasional severe local reaction, but no autoimmune 
encephalitis or tumor metastasis (29). Since cells in this protocol will be rendered non-viable by 
radiation prior to injection, this should be unlikely to occur. Because the cells to be reinjected in this 
study will have a major growth factor pathway suppressed by antisense IGF-I, unrestrained growth of 
the tumor cells is theoretically less likely. Our previous experience with the rat model showed that 
transfected cells never became tumors, because they are unable to produce IGF-I and in addition 
propagate a T cell mediated immune response within the host (30). Nevertheless as part of their 
follow-up, patients will be monitored for tumor growth at the site of injection. Should a tumor mass 
develop, its size and growth rate will be monitored on a frequent basis. If the mass does not regress, or if 
functional limitation or pain develops, the mass will be entirely resected, followed by local radiation 
therapy. Even if no tumor growth occurs at the site of injection, needle biopsy of the subcutaneous pocket 
will be performed 4 weeks following injection of transfected cells to assess for the presence of viable 
tumor, CD8+ killer T cells and other immune mediators. 
Immunologic attack on the tumor and the injected transfected cells could lead to local immune- 
mediated phenomena such as swelling and pain at the site of injection. Clinically this might be 
difficult to distinguish from a bacterial or fungal infection at the injection site. Under such 
circumstances, blood cultures will be obtained and appropriate intravenous antibiotics will be 
administered. If cultures are negative, and the condition does not respond to broad spectrum antibiotics, 
biopsy of the injection site will be performed to establish a diagnosis. Ulceration occurred in the study 
in which tumor cells were injected with BCG adjuvant (29). Severe reactions are less likely in the 
absence of adjuvant. 
Although CD8+ killer T cells are involved, the precise mechanism by which antisense 
transfection works in the rat model is not known. Expression of the antisense vector and subsequent 
downregulation of IGF-I may lead to an unmasking or increased expression of oncofetal antigens on the 
surface of transfected cells. These antigens might also be present in smaller amounts or in a slightly 
different form on untransfected tumor cells. Therefore it is possible that other antigens present on both 
tumor and normal glia could be affected and become susceptible to immune-mediated killing. Under 
such circumstances, toxicity could conceivably affect normal cells within the central nervous system, 
producing an allergic encephalomyelitis or peripheral neuritis. Target antigens shared by other organ 
systems could similarly be attacked. Such a situation, while unlikely, would be initially difficult to 
detect. The potentially serious effects of this scenario would be treated with immunosuppressant 
therapy. To date no evidence for such a phenomenon has been detected in laboratory animals in which 
antisense therapy has been studied or in humans who have undergone other forms of immunotherapy for 
glioma (29). There are however examples of allergic encephalomyelitis in animals immunized with 
antigens derived form autologous nervous tissue combined with Freund's adjuvant. 
[142] 
Recombinant DNA Research, Volume 18 
