Background:  The dynamic relationships between HIV-1 populations in cerebrospinal fluid (CSF) and the periphery over the full course of infection have not been well characterized. A better perception of how viral genotypes evolve in the CSF and peripheral blood over the course of infection may improve our understanding of the mechanisms of HIV-1 neuroinvasion and adaptation to the central nervous system (CNS). Like HIV infection in humans, simian immunodeficiency virus (SIV_sm) invades the CNS of infected macaques and can cause neurological disease in a subset of animals.

Methods:  We used an SIV_sm/macaque model of HIV-1 infection to characterize viral genetic populations in blood plasma and CSF in a longitudinal manner, from acute infection to end-stage disease. We challenged 3 rhesus macaques intravenously with the SIV_sm E660 inoculum, which contains a swarm of viral genotoypes that have not been purposely adapted to the CNS. Paired blood plasma and CSF samples were obtained every 2 to 4 weeks for the duration of infection. Viral genetic populations in plasma and CSF were characterized by using heteroduplex tracking assays targeting the V1/V2 region of env.

Results:  In all 3 macaques, virus was detected in both plasma and CSF at the earliest post-challenge sample time-points of approximately 2 weeks. The SIV_sm env genetic populations in plasma and CSF were complex, but indistinguishable at this time, indicative of a high level of equilibration between the 2 compartments. Viral genetic populations in plasma were unchanged until approximately Days 75 to 100 post-challenge, when major changes in the viral population structure were apparent. In 2 macaques, SIV_sm env genetic populations in CSF mirrored those in plasma over the entire course of infection. In the third macaque, a strikingly different dynamic pattern was observed, with env genetic populations in CSF diverging from those in plasma by day 31, which preceded any detectable change in the plasma viral population structure.

Conclusions:  SIV_sm populations are present in CSF early in infection, with no evidence of a genetic bottleneck for CSF invasion from the periphery. Furthermore, SIV_sm populations in CSF can emerge and turn over rapidly. Even so few animals, we observed different relationships between the blood plasma and CSF viral populations over the course of infection: close concordance over a prolonged period of time versus rapid discordance.