Background: The initiation of drug therapy can result in a reduction of virus load, causing a contraction in the virus population. This contraction represents a potential genetic bottleneck for HIV-1. While the bottleneck directly affects the gene targeted by the drug, it may also have an impact on other regions. We have used V1/V2- and V3-HTAs to determine the effect of a protease inhibitor-induced genetic bottleneck on env gene populations.

Methods: The V1/V2 and V3 regions of the HIV-1 env gene were amplified from plasma samples using RT-PCR. PCR products were annealed to radiolabeled probes designed from a molecular clone spanning the same region. Heteroduplexes were resolved using native PAGE and quantified with storage phosphor autoradiography.

Results: We analyzed 15 subjects with HIV-1 infection and low CD4⁺ T-cell counts. Subjects were classified into groups based on the magnitude and duration of the drop in virus load with the initiation of therapy, although the drop in all subjects was transient. All subjects evolved resistance-associated mutations in pro. Subjects with a strong initial drop in virus load exhibited a loss of heterogeneity in the env population at virus load rebound along with the emergence of new env variants. In contrast, subjects with a weak initial drop exhibited little to no loss of heterogeneity at rebound and retained the same env variants. Prolonged drops in virus load were associated with slower population diversification and the further emergence of new variants, whereas short-term drops in virus load were associated with rapid population diversification and the return of entry variants. No changes were observed in env when protease developed further resistance mutations after virus load rebound.

Conclusions: The extent of env population disruption is dependent on the magnitude of the virus load drop, not just passage through a drug-induced genetic bottleneck. Low virus load was correlated with contractions in env heterogeneity and the emergence of new env variants. High virus load was correlated with little to no contraction or change in env variants. We propose that genetic recombination is limited during times of low virus load and this slows the reappearance of pre-existing env variants into the drug-resistant background. One benefit of suppressive therapy may be to limit the level of recombination thus making a more genetically stable target for drug and immune selection.