Residual Viremia in an RT-SHIV Rhesus Macaque HAART Model Marked by the Presence of a Predominant Plasma Clone and a Lack of Viral Evolution

Highly active antiretroviral therapy (HAART) significantly reduces HIV-1 replication and prevents progression to AIDS. However, residual low-level viremia (LLV) persists and long-lived viral reservoirs are maintained in anatomical sites. These reservoirs permit a recrudescence of viremia upon cessation of therapy and thus HAART must be maintained indefinitely. HIV-1 reservoirs include latently infected resting memory CD4+ T-cells and macrophages which may contribute to residual viremia. It has not been conclusively determined if a component of LLV may also be due to residual replication in cells with sub-therapeutic drug levels and/or long-lived chronically infected cells. In this study, RT-SHIVmac239 diversity was characterized in five rhesus macaques that received a five-drug HAART regimen [tenofovir, emtricitabine, zidovudine, amdoxovir, (A, C, T, G nucleoside analogs) and the non-nucleoside reverse transcriptase (RT) inhibitor efavirenz]. Before maximal viral load suppression, longitudinal plasma viral RNA RT diversity was analyzed using a 454 sequencer. After suppression, LLV RT diversity (amino acids 65-210) was also assessed. LLV samples had viral levels less than our standard detection limit (50 viral RNA copies/mL) and few transient blips <200 RNA copies/mL. HAART was discontinued in three macaques after 42 weeks of therapy resulting in viral rebound. The level of viral divergence and the prevalence of specific alleles in LLV was similar to pre-suppression viremia. While some LLV sequences contained mutations not observed in the pre-suppression profile, LLV was not characterized by temporal viral evolution or apparent selection of drug resistance mutations. Similarly, resistance mutations were not detected in the viral rebound population. Interestingly, one macaque maintained a putative LLV predominant plasma clone sequence. Together, these results suggest that residual replication did not markedly contribute to LLV and that this model mimics the prevalence and phylogenetic characteristics of LLV during human HAART. Therefore, this model may be ideal for testing HIV-1 eradication strategies.