We have engineered immunotoxins into exquisitely cell type specific regents with promise for cancer therapy. Exploring their applications in vivo, we have found that 1) there are powerful pharmacologic barriers that limit protein access to tumor cells; 2) this problem is exacerbated in the brain where the blood-brain barrier prevents macromolecule movement into the brain tissue; 3) the plant and bacterial toxins used for construction of immunotoxins are highly immunogenic and soon after treatment antibodies arise that inactivate reagent. Thus, to overcome the problems of delivery and immunogenicity, we have pursued regional delivery of immunotoxins to the brain as a way to treat brain tumors. Since cancer can spread and grow in the cerebrospinal fluid, a condition known as leptomeningeal carcinomatosis, immunotoxins were initially injected directly into the cerebrospinal fluid to access tumor cells and were found to kill 99% to 99.9% of the tumor cells in vivo with occasional animals cured. An intriguing dose limiting toxicity was found specifically related to this route of administration. Purkinje cells were killed by diphtheria toxin derived immunotoxin in guinea pigs and ricin derived immunotoxins in rats and monkeys. Another protein, called the eosinophil-derived neurotoxin is homologous to RNases A and also selectively kills Purkinje cells. 4) Comparing a family of homologous RNases, we found 5000-fold variation in cytotoxicity. The molecular basis of toxicity was explored and cell binding, RNase inhibitor sensitivity and/or enzyme activity all appear to contribute. 5) We have determined the dose limiting toxicity of immunotoxins in three model species, guinea pigs, rats and rhesus monkeys, and in man. 6) We have explored the potential of diffusion of transferrin (Tf)- CRM107 for brain tumor therapy. Tf-CRM107 is an immunotoxin we designed that has extremely potent and specific tumor cell toxicity. Direct intratumoral injection of brain tumors grown in the flanks of nude mice with Tf-CRM107 effected frequent cures of animals. The toxicity of Tf- CRM107 infused into brain was determined. Concentrations of Tf-CRM107 below 1 x 10-8 M were not toxic whereas higher concentrations caused brain infarction. The normal brain tolerated doses of Tf-CRM107 3000- fold higher than the doses that were toxic to a number of human brain tumor cell lines in vitro. The promising bioactivity of the drug and favorable safety profile suggested Tf-CRM107 may be a valuable therapeutic drug. 7) Recent findings of similar bioactivity and structure shared by diphtheria toxin and Bcl-2 family members will be used to design new apoptosis-regulating immunotoxins; 8) we have identified ways to protect brain endothelial cells from the side effects caused by the Tf-CRM107 thus widening the therapeutic window.
