Late infantile neuronal ceroid lipofuscinosis (LINCL), a pediatric autosomal recessive neurodegenerative lysosomal storage disorder, results from mutations in the CLN2 gene and consequent deficiency in tripeptidyl-peptidase I (TPP-I) and progressive destruction of neurons. We have previously demonstrated that CNS gene transfer of AAV2(CU)hCLN2 (an AAV2-based vector expressing the human CLN2 cDNA) in rats and nonhuman primates mediates long-term TPP-I expression in the CNS neurons [Sondhi, D., Peterson, D.A., Giannaris, E.L., Sanders, C.T., Mendez, B.S., De, B., Rostkowski, A., Blancard, B., Bjugstad, K., Sladek, J.R., Redmond, D.E., Leopold, P.L., Kaminsky, S.M., Hackett, N.R., and Crystal, R.G. (2005). Gene Ther. 12, 1618-1632]. The present study tests the hypothesis that direct CNS administration of a clinical-grade AAV2(CU)hCLN2 vector to the CNS of rats and nonhuman primates at doses scalable to humans has a long-term safety profile acceptable for initiating clinical trials. Fischer 344 rats were injected bilaterally via the striatum with 2 x 10(10) particle units (PU) of AAV2(CU)hCLN2, using saline as a control. At 13, 26, and 52 weeks, vector and phosphate-buffered salineinjected rats were killed (n = 6 per time point), and blood, brain, and distant organs were assessed. There were no biologically significant differences between control and vector groups for complete blood count, serum chemistry, and neutralizing anti-AAV2 antibody levels. CNS administration of AAV2 CUhCLN2 did not result in any pathological changes in the brain that were attributable to the vector, although microscopic changes were observed along the track consistent with needle trauma. A total dose of 3.6 x 10(10) or 3.6 x 10(11) PU of AAV2(CU)hCLN2 was administered to the CNS of African Green monkeys at 12 locations, targeting the caudate nucleus, hippocampus, and overlying cortices. Monkeys (n = 3 at each dose) were killed 1, 13, 26, or 52 weeks after injection. Controls included sham-injected, saline-injected, and AAV2(CU)Null-injected (3.6 x 10(11) PU) monkeys. There were no biologically significant differences among vector-injected and control groups in any parameter of the general assessment, complete blood count, or serum chemistry assessed at multiple time points after vector administration. Importantly, no abnormal behavior was observed in any group in videotaped neurological assessment, where behaviors were quantified before administration and at multiple time points afterward. Histopathological examination of the CNS demonstrated that 1 week after administration, AAV2(CU)hCLN2 produced transient minor white matter edema with reactive glial cells in the corona radiata of the cerebrum along the injection track and in the surrounding white matter. This abnormality was not observed at 13, 26, or 52 weeks. Together with the long-term gene expression after gene transfer, these findings supported the initiation of clinical trials to assess the safety of AAV2(CU)hCLN2 administration to individuals with LINCL.