HIV-1 genome is often defective in PBMCs and rectal tissues after long-term HAART as a result of APOBEC3 editing and correlates with the size of reservoirs.

OBJECTIVES Precise characterization of viruses present in reservoirs in long-term pretreated patients will be a major issue to consider in the context of viral eradication. We assessed the frequency of defective viruses present in cellular reservoirs. METHODS Peripheral blood mononuclear cells (PBMCs) and rectal biopsy samples were compared between five patients on successful long-term highly active antiretroviral therapy (HAART) (>7 years without blips) and five untreated patients. Molecular cloning and sequencing of the reverse transcriptase region were used to detect the presence of and quantify in-frame stop codons in HIV quasi-species. The relationship between the size of the reservoir and the frequency of defective genomes was assessed. RESULTS Defective genomes were systematically detected in all patients on long-term HAART in both compartments (PBMCs and rectal tissues), with a higher level of defective genomes per sample compared with PBMCs of untreated patients. A high level of defective genomes was correlated with a small size of HIV proviral DNA. Regarding the nucleotide context, guanine (G) to adenine (A) substitution at tryptophan positions was responsible for the appearance of 89% of all in-frame stop codons in the context of G-to-A hypermutation, likely reflecting APOBEC3 footprints on the viral genome. CONCLUSIONS We propose a scenario whereby defective genomes accumulate during HAART treatment, eventually reaching a viral extinction threshold. In the context of viral eradication, measurement of the relative amounts of defective and non-defective viruses (by molecular cloning and ultradeep sequencing) should be used as a new criterion for eradicating HIV.