Although cells of monocytic lineage are the primary source of human immunodeficiency virus type 1 (HIV-1) in the brain, other cell types in the central nervous system, including astrocytes, can harbor a latent or persistent HIV-1 infection. In the present study, we examined whether immature, multipotential human brain-derived progenitor cells (nestin positive) are also permissive for infection. When exposed to IIIB and NL4-3 strains of HIV-1, progenitor cells and progenitor-derived astrocytes became infected, with peak p24 levels of 100 to 500 pg/ml at 3 to 6 days postinfection. After 10 days, virus production was undetectable but could be stimulated by the addition of tumor necrosis factor alpha (TNF-␣). To bypass limitations to receptor entry, we compared the fate of infection in these cell populations by transfection with the infectious HIV-1 clone, pNL4-3. Again, transfected progenitors and astrocytes produced virus for 7 days but diminished to low levels beyond 8 days posttransfection. During the nonproductive phase, TNF-␣ stimulated virus production from progenitors as late as 5 weeks posttransfection. Astrocytes produced 5-to 20-fold more infectious virus (27 ng of p24/10 6 cells) than progenitors at the peak of 3 days posttransfection. Differentiation of infected progenitors toward an astrocyte phenotype increased virus production to levels consistent with infected astrocytes, suggesting a phenotypic difference in viral replication. Using this cell culture system of multipotential human brain-derived progenitor cells, we provide evidence that progenitor cells may be a reservoir for HIV-1 in the brains of AIDS patients.

Progressive multifocal leukoencephalopathy is a demyelinating disease of the human central nervous system that results from lytic infection of oligodendrocytes by the polyomavirus JC (JCV). Originally, JCV was thought to replicate exclusively in human glial cells, specifically oligodendrocytes. However, we have recently shown that JCV can replicate in cells of lymphoid origin such as hematopoietic precursor cells, B lymphocytes, and tonsillar stromal cells. To determine whether tonsils harbor JCV, we tested a total of 54 tonsils, 38 from children and 16 from adult donors. Nested PCRs with primer sets specific for the viral T protein and regulatory regions were used for the detection of JCV DNA. JCV DNA was detected in 21 of 54 tonsil tissues, or 39% (15 of 38 children and 6 of 16 adults) by using regulatory-region primers and in 19 of 54 tonsil tissues, or 35% (13 of 38 children and 6 of 16 adults) by using the T-protein primers. The DNA extracted from children’s nondissected tonsil tissue, isolated tonsillar lymphocytes, and isolated stromal cells that demonstrated PCR amplification of the JCV regulatory region underwent cloning and nucleotide sequencing. Of the regulatory-region sequences obtained, nearly all contained tandem repeat arrangements. Clones originating from nondissected tonsil tissue and tonsillar lymphocytes were found to have sequences predominantly of the Mad-1 prototype strain, whereas the majority of clones from the DNA of tonsillar stromal cells had sequences characteristic of the Mad-8br strain of JCV. A few clones demonstrated structures other than tandem repeats but were isolated only from tonsillar lymphocytes. These data provide the first evidence of the JCV genome in tonsil tissue and suggest that tonsils may serve as an initial site of viral infection.

No abstract