Dynamic Shifts in the HIV Proviral Landscape During Long Term Combination Antiretroviral Therapy: Implications for Persistence and Control of HIV Infections

Abstract

Combination antiretroviral therapy (cART) controls but does not eradicate HIV infection; HIV persistence is the principal obstacle to curing infections. The proportion of defective proviruses increases during cART, but the dynamics of this process are not well understood, and a quantitative analysis of how the proviral landscape is reshaped after cART is initiated is critical to understanding how HIV persists. Here, we studied longitudinal samples from HIV infected individuals undergoing long term cART using multiplexed Droplet Digital PCR (ddPCR) approaches to quantify the proportion of deleted proviruses in lymphocytes. In most individuals undergoing cART, HIV proviruses that contain gag are lost more quickly than those that lack gag. Increases in the fraction of gag-deleted proviruses occurred only after 1-2 years of therapy, suggesting that the immune system, and/or toxicity of viral re-activation helps to gradually shape the proviral landscape. After 10-15 years on therapy, there were as many as 3.5-5 times more proviruses in which gag was deleted or highly defective than those containing intact gag. We developed a provirus-specific ddPCR approach to quantify individual clones. Investigation of a clone of cells containing a deleted HIV provirus integrated in the HORMAD2 gene revealed that the cells underwent a massive expansion shortly after cART was initiated until the clone, which was primarily in effector memory cells, dominated the population of proviruses for over 6 years. The expansion of this HIV-infected clone had substantial effects on the overall proviral population.

Keywords: HIV clonal expansion; HIV persistence; ddPCR; proviruses.

Figure 1

All HIV-1 DNA regions analyzed decline when combination antiretroviral therapy (cART) is initiated. HIV-1 DNA copies per 1 million CD4+ T-cells shown in green (long terminal repeat (LTR)), purple (gag), gold (tat exon 1), and blue (tat/rev exon 2). HIV viral load as HIV RNA copies per mL plasma is shown in black; open symbols represent values that are below the limit of detection. Total CD4 count in cells/mL shown in lower panel in red. Grey background represents periods of cART. (A) A participant treated during recent HIV infection AVBIO2_14. (B) A participant treated during chronic HIV infection, AVBIO2_17. (C) A participant treated during AIDS, AVBIO2_08. Error bars for copy numbers are omitted for clarity; as described in Methods, range in copy number was ≤2-fold in multiple determinations of each sample. (D). HIV-1 provirus map denoting positions of primers amplifying LTR, gag, tat exon1, and tat/rev exon2.

Figure 2

The ratio of HIV-1 LTR DNA to internal HIV-1 DNA regions during cART. HIV viral load (HIV-1 RNA copies per ml plasma) is shown in black; open symbols represent values that are below the limit of detection. LTR:gag ratios (purple), LTR:tat exon 1 ratios (gold), LTR:tat/rev exon 2 ratios (blue) are also shown. Total CD4 counts (cells/mm³) are shown in the lower panel in red. Grey background represents periods of cART. (A) A participant treated during recent HIV infection AVBIO2_14. (B) A participant treated during chronic HIV infection, AVBIO2_17. (C) A participant treated during AIDS, AVBIO2_08.

Figure 3

HIV LTR to internal HIV-1 DNA target ratios increase in most individuals after prolonged cART. Box and whisker plots of HIV-1 LTR DNA to internal HIV-1 DNA ratios for all study participants were determined at pretherapy, after viral suppression, during 3^rd phase viral decay, and by prolonged cART (one-way ANOVA, Tukey HSD p < 0.05, ** p < 0.01; outlier values included as indicated; all other pairwise comparisons were not significant, LTR:tat exon 1 p = 0.59, and LTR:tat/rev exon 2 p = 0.58). (A) HIV-1 LTR DNA to HIV-1 gag DNA ratios, (B) HIV-1 LTR DNA to HIV-1 tat exon 1 DNA ratios and (C) HIV-1 LTR DNA to HIV-1 tat/rev exon 2 DNA ratios.

Figure 4

Cells containing deleted proviruses are present after treatment interruption and resuppression. The effects of treatment interruption and subsequent resuppression on HIV-1 LTR and gag ratios in four individuals who experienced treatment discontinuation. LTR DNA to gag DNA ratios before, during, and after treatment interruption. Error bars represent the standard deviation of ratios calculated from 3 to 12 replicates.

Figure 5

Dynamics of the clonal expansion of cells harboring a solo LTR integrated in HORMAD2. Abundance of the solo LTR provirus in the HORMAD2 gene in AVBIO2_21 over time measured with the Droplet Digitial PCR (ddPCR) integration site specific assay (Figure S6). HIV viral load (black diamonds), HIV-1 LTR DNA copies per million CD4+ T cells (green circles), HORMAD2 provirus copies per 1 million CD4+ T cells (blue diamonds), total CD4 count (red diamonds) are shown in the lower panel. Shaded areas indicate periods of cART, blank areas periods of treatment interruption (labeled TI-1, -2,-3). Open symbols indicate values that are less than the limit of detection.

Figure 6

An expanded cell clone harboring a solo LTR in HORMAD2 is enriched in the effector memory subset. Sorted T cell subset analysis showed that the HORMAD2 provirus is primarily in the effector memory subset of participant AVBIO2_ 21: A) Frequency of the LTR (green circles) and gag (purple circles) DNA and the expanded cell clone with the provirus in HORMAD2 (blue diamonds) in the three T cell subsets. Dashed bars and open symbols indicate the limit of detection. B) HIV-1 LTR to gag DNA ratios in the sorted cells (one-way ANOVA, Tukey HSD * p < 0.05, *** p < 0.001). C) Pie charts of integration sites obtained by ISA for cell subsets. The size of the chart reflects the absolute number of integration sites recovered. Integration sites obtained only once are indicated in the grey segment, while colored portions represent integration sites identified more than once (cell clones). Tables list the frequency of integration sites that were found in both the CTM and EM T cell subsets.

Figure 7

Proportion of proviruses that are deleted increases during cART. Prior to cART, non-deleted proviruses dominate the proviral landscape. After a period of viral suppression, defective viruses, including deleted and hypermutant proviruses are enriched; after prolonged cART, such proviruses dominate the proviral landscape.