New xenograft model for assessing experimental therapy of central nervous system tumors: human glioblastoma in the intrathecal compartment of the nude mouse.

Ten congenitally athymic "nude" mice and 10 immunocompetent mice underwent intrathecal inoculation with a human glioblastoma cell line (U87MG) via percutaneous lumbar puncture (5 x 10(5) cells/animal). All of the nude mice developed paraplegia with or without incontinence at 2 weeks and routinely died of inanition 3 weeks postimplantation. Histological examination confirmed extensive proliferation of neoplastic cells within the intrathecal space. A second group of animals was inoculated with 5 x 10(4) cells/animal: 20 nude mice, 10 cyclosporine A-immunosuppressed animals, and 10 immunocompetent control mice. The 20 mice were further divided into four subsets. Subset A did not receive chemotherapy, Subset B received 200 mg of carmustine (BCNU) per m2 by intraperitoneal injection, Subset C received a single dose of 4 mg of methotrexate (MTX) per m2 by intrathecal injection 4 hours after tumor inoculation, and Subset D received 12 mg of intrathecal MTX per m2. Decreasing the concentration of cells per animal by 1 log doubled the time interval required for the development of paralysis and incontinence to 4 weeks. Treatment with intrathecal MTX at a dose of 4 mg/m2 extended the symptom-free period by an additional week (to 5 weeks postinoculation), and a dose of 12 mg/m2 allowed an average of 6 weeks before the onset of neurological impairment. The xenografts did not grow in the immunocompetent control mice, the BCNU-treated group, or the cyclosporine A-immunosuppressed animals. An intrathecal xenograft model of central nervous system malignancies allows a novel approach to the evaluation of experimental chemotherapy.(ABSTRACT TRUNCATED AT 250 WORDS)