Abstract Body

The majority of the HIV-1 proviruses are defective, making it irrelevant to use HIV-1 DNA quantitation to measure the size of the latent reservoir. However, a decrease in HIV-1 proviral DNA levels has been reported in vivo using the shock-and-kill strategy. We have previously shown that defective HIV-1 proviruses can be transcribed and translated, making them potential targets for cytotoxic T lymphocyte (CTL) elimination. We hypothesize that HIV-1 proviral landscape is dynamic, and may be shaped by CTL selection pressure.

We examined the HIV-1 proviral landscape in virally suppressed HIV-1-infected individuals. We then examined whether defective HIV-1 proviruses can be recognized by CTLs in vitro and ex vivo.

In the HIV-1 proviral landscape of virally suppressed HIV-1-infected individuals, we found that the frequency of hypermutated proviruses correlates negatively with the duration of infection, while the frequency of proviruses containing large internal deletions correlates positively with the duration of infection, implying a dynamic process. To examine whether defective HIV-1 proviruses can be recognized by CTLs in vitro, we co-cultured CD4+ T cells transfected with defective HIV-1 proviruses with autologous CTL clones known to recognize specific CTL epitopes. CTL recognition of cells containing defective HIV-1 proviruses were measured by CD107a loading. Surprisingly, we found that cells containing hypermutated proviruses which have a premature stop codon or mutated start codon 5′ to an intact CTL epitope can be recognized by CTLs, while cells containing large internal deletions cannot be recognized by CTLs. To understand whether hypermutated HIV-1 proviruses may be recognized by CTLs ex vivo, we used targeted deep-sequencing to examine the frequency of APOBEC-mediated nonsense mutations in gag. We compared resting and activated CD4+ T cells treated with and without autologous prestimuated CTLs. We found that the frequency of HIV-1 RNA containing premature stop codons 5′ to the CTL recognition site decreased upon CTL co-culture, indicating that cells containing hypermutated proviruses can be eliminated by CTLs ex vivo.

We found that hypermutated HIV-1 proviruses may be targeted by CTLs. HIV-1 DNA measurement may in part reflect the effect of HIV-1 elimination strategies. HIV-1 proviral landscape may be shaped, at least partially, by CTL selection pressure.