NEUROINVASION BY SIMIAN IMMUNODEFICIENCY VIRUS COINCIDES WITH INCREASED NUMBERS OF PERIVASCULAR MACROPHAGES MICROGLIA AND INTRATHECAL IMMUNE ACTIVATION/

During peak viremia and initial antibody response, rhesus macaques inf ected with pathogenic and nonpathogenic isolates of SIV show distinct differences in viral load and tissue distribution. Animals infected wi th pathogenic isolates of SIV invariably have virus in the CSF and bra in parenchyma by two weeks postinoculation, whereas animals infected w ith nonpathogenic isolates do not. Mechanisms underlying neuroinvasion by SIV and HIV are unknown, but recruitment of latently infected mono nuclear cells from the peripheral circulation (Trojan horse theory) is frequently proposed, Circulating monocytes, from which perivascular m acrophage/microglia are derived, are a likely vehicle for cell-associa ted transport of virus across the blood-brain barrier. This transport and the kinetics of perivascular macrophage/microglial turnover in the CNS likely depend on endothelial and leukocyte adhesion molecules suc h as vascular cell adhesion molecule-1 (VCAM-1), which has previously been shown to be upregulated on cerebrovascular endothelium in SIV enc ephalitis. To investigate the role of peripheral monocyte recruitment into the perivascular macrophage/microglial cell pool at the time of i nitial viral neuroinvasion, we examined the temporal relationships amo ng perivascular macrophage/microglia density, endothelial VCAM-1 expre ssion and localization of viral nucleic acid in the CNS of macaques ac utely infected with pathogenic and nonpathogenic molecular clones of S IV. The concentration of CSF quinolinic acid, a marker of intrathecal immune and macrophage activation, was examined concurrently. We found that significant increases in the density of perivascular macrophages/ microglia coincided with viral neuroinvasion and marked elevations in CSF quinolinic acid. Furthermore, combined in situ hybridization and i mmunohistochemistry demonstrated that infected perivascular cells were macrophages/microglia. These findings provide evidence suggesting tha t neuroinvasion occurs through an influx of infected monocytes which t ake up residence in the CNS as perivascular macrophages/microglia. VCA M-1 expression, however, was not clearly correlated with these events, thus its contribution to initial viral neuroinvasion is unclear.