Background:
The capacity of HIV-1 to replicate during antiretroviral therapy (ART) remains controversial. Limitations of clinical sampling and the high background level of genetic diversity in most persons living with HIV make it challenging to detect additional substitutions on ART. To gain greater sensitivity to detect even small changes to the virus population on ART, we used an NHP model, which allows precise control over the level of pre-ART evolution and subsequent changes to the viral population.
Methods:
We infected 21 rhesus macaques with the barcoded virus SIVmac239M to compare near-full-length (nFL) SIV DNA single genome sequences from PBMCs (and in some cases lymph nodes and spleen) obtained near the time of ART initiation and those present after long-term ART. Animals started ART between 10-27 days post infection (dpi) and were treated for 285-977 dpi. We obtained 25-200 intact nFL sequences per animal. We then assessed whether the viral populations differed significantly between the two time points in (i) genetic diversity, (ii) population structure, and (iii) genetic divergence from the founder.
Results:
We found no evidence for replication on ART for any of the animals. The median average pairwise distance did not differ significantly between baseline and long-term ART samples (0.015% vs 0.015%; p-value = 0.23; two-tailed Wilcoxon-signed rank test.) The probability that sequences from the two samples came from different populations was not statistically significant in any animal (test for panmixia). Finally, the regression slope of p-distance over time did not differ significantly from zero in any animal. Neighbor-joining trees were also consistent with a lack of viral evolution, displaying no clustering of sequences from different time points, or longer branches associated with sequences obtained from long-term ART.
Conclusions:
Overall, these data are consistent with the most recent papers in the field, which did not find evidence for evolution of HIV on ART. We cannot exclude the possibility of low-level ongoing viral replication not detectable in our study despite intensive sampling. For example, computational modeling suggests that ongoing replication in a theoretical sanctuary site could lead to viral evolution while maintaining the level of plasma viremia below the limit of detection. However, the potential contribution of such a small replicating population to the rebound-competent reservoir or as a source of drug resistance escape mutations would be limited.