Human immunodeficiency virus type 1 (HIV-1) infection is highly specific for its human host. In order to study HIV-1 infection of the human nervous system, we have established a small animal model in which second-trimester (11-17.5 weeks) human fetal brain or neural retina is transplanted into the anterior chamber of the eye of immunosuppressed adult rats (Epstein et al., 1992; Cvetkovich et al., 1992), and more recently in immunodeficient (SCID) mice. The human xenografts survive for many months, vascularize and form a blood-brain barrier. Immunohistochemistry with PGP 9.5 identified neuronal cell bodies and neuritic processes. Electron microscopy revealed axonal growth cones and synaptic junctions. Infection of these xenografts with cell-free HIV-1 proved difficult, however co-engraftment with HIV-1-infected human monocytes resulted in characteristic pathological changes, including the formation of syncytial giant cells, neuronal loss, and astroglial proliferation, supporting the hypothesis that these cells can mediate neurotoxicity. In other studies, xenografts of human fetal retinal tissue were readily infected with cell-free human cytomegalovirus (HCMV) strain AD169. These grafts contained cells with intracytoplasmic and intranuclear inclusions typical of HCMV infection. Productive infection within these grafts was demonstrated by the presence of immediate early, and late (capsid) HCMV antigens, by recovery of HCMV on human fibroblast cultures, and by serial passage of virus to additional retinal xenografts (DiLoreto et al., 1994). The aim of these studies is to develop a small animal model to study direct and indirect effects of HIV-1 infection on human neural tissues, and to study interactions between HIV-1 and other opportunistic pathogens such as HCMV. This model should prove useful in evaluating antiviral therapies. © 1994 Elsevier Science Ltd.