Intracerebral grafting combined with gene transfer may provide a powerful technique for local delivery of therapeutic agents into the CNS. The present study was undertaken to: (i) develop a reliable and reproducible automated cell implantation system, (ii) determine optimal implantation parameters of cells into the striatum, (iii) determine upper safe limits of cellular implantation into the neostriatum of monkeys. Autologous fibroblasts were infused into six sites of the striatum in nonhuman primates (Macaca mulatta, n = 11). Twenty-six-gauge cannulae were inserted vertically through cortical entry sites into the striatum (two sites in the caudate nucleus and four sites in the putamen) at predefined coordinates based on magnetic resonance imaging (MRI). The cannulae were guided by an electronically operated, hydraulic micropositioner and withdrawn at controlled rates, while cells (5, 10, 20, 40, or 80 microl/site) were infused simultaneously. Varying infusion rates and cell concentrations were also evaluated. Visualization and evaluation of graft placement were performed using contrast MRI at 3-5 days postsurgery. Animals were monitored for signs of clinical complications and sacrificed 2 weeks following surgery. Postimplantation MRI revealed a tissue mass effect of the implant with shifting of midline, edema, and infiltration of the white tracts at 40 and 80 microl/site. In addition, these animals developed transient hemiparesis contralateral to the implant site. MRI of animals grafted with 20 microl/site exhibited columnar-shaped implants and evidence of infiltration into white matter tracts possibly due to a volume effect. No clinical side effects were seen in this group. At 14 days postsurgery, MRI scans showed consistent columnar grafts (measuring approximately 5 mm in height) throughout the striatum in animals implanted with 5 or 10 microl/site. No signs of clinical side effects were associated with these volumes and postmortem histological examination confirmed MRI observations. Optimal surgical parameters for delivery of cells into the striatum consist of a graft volume of 10 microl/site, an infusion rate of 1.6 microl/min, a cell concentration of 2.0 x 10(5) cells/microl, and a cannula withdrawal rate of 0.75 mm/min. These results show that infusion of cells into the striatum can be done in a safe and routine manner.