Background: HIV-1 transcription during suppressive antiretroviral therapy (ART) is not well understood. This is problematic as latency-reactivating agent-based HIV-1 eradication trials utilize changes in viral transcription as an efficacy biomarker.
Methods: We conducted an observational cohort study enrolling aviremic, HIV-1-infected adults on long-term ART. Cell-associated unspliced (CA-US) HIV-1 RNA and total HIV-1 DNA were quantified in unfractionated CD4 T cells monthly for a total of six consecutive visits. Random-effects models were used to determine the following: (i) proportion of variation attributable to intra-individual versus inter-individual changes; (ii) range estimate for random samples from any participant or cohort-matched individual (95% prediction interval); and (iii) range estimate for random samples from the same person (95% variation intervals expressed as fold change).
Results: Among our cohort of 26 HIV-1 patients, 10.4% of variation in CA-US HIV-1 RNA was attributable to intra-individual fluctuations. Similarly, intra-individual changes also accounted for minor proportions of the variation in total HIV-1 DNA (5.1%) and RNA/DNA (28.3%). The 95% prediction interval (per 10 CD4 T cells) for CA-US HIV-1 RNA and HIV-1 DNA were each approximately 2 log10. Finally, model-derived 95% variation intervals indicate that spontaneous changes above 2.11-fold in CA-US HIV-1 RNA would occur in less than 5% of repeated measurements in an individual on long-term ART.
Conclusion: The individual CA-US HIV-1 RNA levels are remarkably stable during ART. Importantly, the observed variations were less than the reported changes for latency-reactivating agent trials. These data will serve as a foundation for planning and interpreting future eradication trials.