HIV Infected T Cells Can Proliferate in vivo Without Inducing Expression of the Integrated Provirus

Abstract

Background: HIV-1 proviruses can persist during ART in clonally-expanded populations of CD4+ T cells. To date, few examples of an expanded clones containing replication-competent proviruses exist, although it is suspected to be common. One such clone, denoted AMBI-1 (Maldarelli et al., 2014), was also a source of persistent viremia on ART, begging the question of how the AMBI-1 clone can survive despite infection with a replication-competent, actively-expressing provirus. We hypothesized that only a small fraction of cells within the AMBI-1 clone are activated to produce virus particles during cell division while the majority remain latent despite division, ensuring their survival. To address this question, we determined the fraction of HIV-1 proviruses within the AMBI-1 clone that expresses unspliced cell-associated RNA during ART and compared this fraction to 33 other infected T cell clones within the same individual.

Results: In total, 34 different clones carrying either intact or defective proviruses in "Patient 1" from Maldarelli et al. (2014) were assessed. We found that 2.3% of cells within the AMBI-1 clone contained unspliced HIV-1 RNA. Highest levels of HIV-1 RNA were found in the effector memory (EM) T cell subset. The fraction of cells within clones that contained HIV-1 RNA was not different in clones with intact (median 2.3%) versus defective (median 3.5%) proviruses (p = 0.2). However, higher fractions and levels of RNA were found in cells with proviruses containing multiple drug resistance mutations, including those contributing to rebound viremia.

Conclusion: These findings show that the vast majority of HIV-1 proviruses within expanded T cell clones, including intact proviruses, may be transcriptionally silent at any given time, implying that infected T cells may be able to be activated to proliferate without inducing the expression of the integrated provirus or, alternatelively, may be able to proliferate without cellular activation. The results of this study suggest that the long, presumed correlation between the level of cellular and proviral activation may not be accurate and, therefore, requires further investigation.

Keywords: CARD-SGS; HIV reservoir; SGA; cell-associated HIV RNA; expanded clones; latent; latently-infected cells; proviral expression.

FIGURE 1

(A) Longitudinal viral load plot of Patient 1 in Maldarelli et al. (2014) and Simonetti et al. (2016) from the time of ART initiation to his death from cancer. Viral loads below limit of detection (<50 copies per mL) are indicated with open black circles. The red arrow in the right panel indicates the timepoint that was analyzed by CARD-SGS (viral load of 134 copies/ml). (B) NJ tree showing plasma RNA sequences at three marked timepoints on the viral load plot in A. Solid circles indicate sequences with no drug resistance mutations and open circles indicate drug-resistant variants. The ART regimens employed are shown in the boxes at the top. ^∗Indicated that the drug regimen was simplifed for the figure.

FIGURE 2

NJ distance tree of PBMC viral DNA sequences from the timepoint indicated in red in Figure 1. The tree contains possible clones with matching ca-HIV RNA (indicated with numbers) including the replication-competent clone AMBI-1, and the possible clones Outgrowth-1 and Outgrowth-2 in red, defective clones in purple, non-induced possible clones in blue, and DR proviruses with matching ca-HIV RNA in orange. Proviruses where no ca-HIV RNA was detected are shown in gray. A total of 183 proviral sequences were analyzed.

FIGURE 3

NJ distance tree of proviral sequences from CD4+ T cell memory subsets from the first timepoint (red) in Figure 1. PBMC were sorted into naïve (pink), central-transitional memory (black), and effector memory (orange) CD4+ T-cell subsets. Possible clones from which ca-HIV RNA was recovered are numbered, including VOA outgrowth viruses. DR mutations are indicated.

FIGURE 4

Plot of the fraction of cells that contained ca-HIV RNA for each of the 34 different clones/possible clones. Replication-competent clones/possible clones (shown in red) are intact and outgrowth was observed in VOA. Defective clones (shown in purple) were hypermutated and contained stop codons. Non-induced possible clones (shown in blue) did not contain obvious defects but outgrowth of the matching virus was not observed by VOA. About 50–1200 cells within each clone/possible clone were assayed for ca-HIV RNA.

FIGURE 5

NJ distance tree of drug-resistant variants. Plasma SGS are shown as red, open circles, PBMC proviruses with matching ca-HIV RNA are shown as orange diamonds, ca-HIV RNA in single cells are shown in different colored squares, PBMC proviruses with no ca-HIV RNA detected are open diamonds. The number of single cells that contained ca-HIV RNA is indicated by the number of different colored squares. The number of ca-HIV RNA molecules in each single cell is indicated by the number of squares of each color. One cell that contained a high level of ca-HIV RNA was detected (bright pink squares) and the sequence of this ca-HIV RNA matched the sequence of the most prevalent variant in the plasma and the sequence of the Drug Resistant Outgrowth-1 variant. Only RNA sequences identical to DR proviral DNA sequences are shown.

FIGURE 6

Levels of ca-HIV RNA in single cells that carried the wild-type (blue) and drug resistant (red) proviruses. The number of ca-HIV RNA copies per cell in the wild-type population ranged from 1 to 16. The number of ca-HIV RNA copies per cell in the drug resistant population ranged from 1 to 65. Data were analyzed from 820 cells with wild-type proviruses and 324 cells with drug resistant proviruses. The remaining infected cells did not contain any detectable ca-HIV RNA. The high-expressing cells are indicated with the black arrow.