Neural stem cell transplantation is a promising strategy for the treatment of neurodegenerative diseases. To evaluate the differentiation potential of human neural progenitor cells (hNPCs) as a prerequisite for clinical trials, we intracerebrally transplanted in vitro expanded fetal mesencephalic hNPCs into hemiparkinsonian rats. On postnatal day one (P1), 17 animals underwent a unilateral intraventricular 6-hydroxydopamine injection into the right lateral ventricle. At P3, animals (n = 10) received about 100,000 hNPCs (1 microL) in the right striatum. Five weeks after birth, animals underwent behaviour tests prior to fixation, followed by immunohistochemistry on brain slices for human nuclei, glial fibrillary acidic protein, S100beta, neuronal nuclei antigen, neuron-specific enolase and tyrosine hydroxylase. Compared with the apomorphine-induced rotations in the lesioned-only group (7.4 +/- 0.5 min(-1)), lesioned and successfully transplanted animals (0.3 +/- 0.1 min(-1)) showed a significant therapeutic improvement. Additionally, in the cylinder test, the lesioned-only animals preferred to use the ipsilateral forepaw. Conversely, the lesioned and transplanted animals showed no significant side bias similar to untreated control animals. Transplanted human nuclei-immunoreactive cells were found to survive and migrate up to 2000 microm into the host parenchyma, many containing the pan-neuronal markers neuronal nuclei antigen and neuron-specific enolase. In the striatum, tyrosine hydroxylase-immunoreactive somata were also found, indicating a dopaminergic differentiation capacity of transplanted hNPCs in vivo. However, the relative number of tyrosine hydroxylase-immunoreactive neurons in vivo seemed to be lower than in corresponding in vitro differentiation. To minimize donor tissue necessary for transplantation, further investigations will aim to enhance dopaminergic differentiation of transplanted cells in vivo.