Innate, non-cytolytic CD8+ T cell-mediated suppression of HIV replication by MHC-independent inhibition of virus transcription

Abstract

MHC-I-restricted, virus-specific cytotoxic CD8+ T cells (CTLs) may control human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) replication via the recognition and killing of productively infected CD4+ T cells. Several studies in SIV-infected macaques suggest that CD8+ T cells may also decrease virus production by suppressing viral transcription. Here, we show that non-HIV-specific, TCR-activated non-cytolytic CD8+ T cells suppress HIV transcription via a virus- and MHC-independent immunoregulatory mechanism that modulates CD4+ T cell proliferation and activation. We also demonstrate that this CD8+ T cell-mediated effect promotes the survival of infected CD4+ T cells harboring integrated, inducible virus. Finally, we used RNA sequencing and secretome analyses to identify candidate cellular pathways that are involved in the virus-silencing mediated by these CD8+ T cells. This study characterizes a previously undescribed mechanism of immune-mediated HIV silencing that may be involved in the establishment and maintenance of the reservoir under antiretroviral therapy and therefore represent a major obstacle to HIV eradication.

Fig 1. TCR-stimulated CD8⁺ T cells but not activated B cells or monocytes mediate suppression of HIV replication in CD4⁺ T cells.

(A) CD4/CD8 Flow Cytometry Gating Strategy shows isolated CD4⁺ and CD8⁺ T cells separately labeled with CellTrace Violet dye (Vio⁺) and CellTrace Red dye (Red⁺). (B) Representative flow plots of uninfected (mock), and infected wells (HIV) depicting productively infected CD4⁺ T cells (live CD3⁺Vio⁺Red^-CD8^-eGFP⁺ cells) from CD4 mono- and CD4/CD8 co-cultures; the levels of suppression are represented by solid lines connecting the frequencies of eGFP⁺ cells from CD4 mono-culture wells (black), CD4/CD8 at 1:1 (blue) and 5:1 (red) E:T ratios from each subject (n = 10 subjects). (C) Co-culture with activated B cells or Monocytes. The frequencies of productively infected CD4⁺ T cells are shown from the CD4 mono-culture wells (black), CD4/CD19 at 1:1 (light orange) and 5:1 (dark orange) E:T ratios; CD4/CD14 at 1:1 (light green) and 5:1 (dark green) E:T ratios. eGFP⁺ levels are individual values from 8 HIV-negative distinct subjects with solid lines connecting values from mono- and co-cultures from each subject. (D) Infection with Env-defective NL4-3_D2eGFP virus. The levels of suppression are shown by dashed lines connecting the frequencies and mean fluorescence intensity (MFI) of D2eGFP⁺ cells from CD4 mono-culture wells (black), CD4/CD8 at 1:1 (blue) and 5:1 (red) E:T ratios from each subject (n = 7 subjects). Flow cytometric data for all the samples were performed on day 3 post- HIV infection, and comparisons between frequencies and MFI of HIV expression on mono- and co-cultures were carried out using Wilcoxon matched-pairs signed rank test.

Fig 2. Effect of TCR-stimulated CD8⁺ T cells on HLA-DR, HLA-E and proliferation levels of uninfected and infected CD4⁺ T cells.

Representative mean fluorescence intensity (MFI) of HLA-DR (A), HLA-E (B) and CellTrace Violet (C) is shown for non-productively infected and uninfected eGFP^-CD4⁺ T cells from mono- and co-cultures in the multiple round infection assay. (A and B) The aggregate MFI data are also shown for HLA-DR and HLA-E of uninfected CD4⁺ T cells (mock; indicated by triangles), non-productively infected and uninfected CD4⁺ T cells (eGFP^-; indicated by squares), and productively infected CD4⁺ T cells (eGFP⁺; indicated by circles) of 10 distinct subjects. Solid lines connecting the MFI values of activation markers (HLA-DR or HLA-E) from CD4 mono-culture wells (black), CD4/CD8 at 1:1 (blue) and 5:1 (red) E:T ratios from each subject represent the levels of suppression by CD8⁺ T cells. (C) Fold change in CellTrace violet MFI relative to CD4⁺ T cells alone followed by a f (x) = 1/x transformation are depicted for uninfected CD4⁺ T cells (mock; indicated by triangles), non-productively infected and uninfected CD4⁺ T cells (eGFP^-; indicated by squares), and productively infected CD4⁺ T cells (eGFP⁺; indicated by circles) of 10 distinct subjects. Black lines represent the mean and standard deviation. Flow cytometric data for all the samples were performed on day 3 post- HIV infection, and comparisons between MFI of activation and proliferation markers on co-cultures with that of positive control wells (CD4⁺ T cells alone) were carried out using Wilcoxon matched-pairs signed rank test. See corresponding single cycle infections data of NL4-3_eGFP virus (S2B Fig) and NL4-3_D2eGFP virus (S2C Fig).

Fig 3. TCR-stimulated CD8⁺ T cells protect target cells from virus cytopathic effects, lack MHC-restriction, and is correlated with increased levels of Th2 cytokines.

(A-E) In vitro infections were conducted using the replication competent NL4-3_eGFP. (A) Frequencies of live CD4⁺ T cells were determined using a more stringent viability cutoff to include only cells that are negative for Live/Dead-Aqua dye labelling (live CD3⁺Vio⁺Red^- cells) in the infected mono—or co-culture wells (n = 10 subjects). (B) Frequencies of HIV-induced cell death were calculated by subtracting the frequency of live CD4⁺ T cells in the infected mono—or co-cultures wells by the frequency of live CD4⁺ T cells in the corresponding uninfected mono—or co-cultures wells (mock infection; n = 10 subjects). (C) Anti-pan HLA-ABC blocking mAb W6/32 (20 μg/ml), anti-HLA-E mAb 3D12 (20μg/ml), IgG2 (20 μg/ml) or IgG1 (20 μg/ml) were added to the CD4/CD8 co-cultures after infection. Bars represent mean frequencies of productively infected CD4⁺eGFP⁺ cells and black lines represent the standard deviation (n = 8 subjects). (D) Heterologous CD8⁺ T cells (mismatched donors) were added to CD4⁺ T cells and frequencies of productively infected CD4⁺eGFP⁺ cells were compared to the autologous co-cultures (matched donors). Bars represent mean values and black lines represent the standard deviation (n = 8 subjects). (E) In the transwell assay, infected CD4⁺ T cells were placed in the top chamber and autologous CD8⁺ T cells were placed in the bottom chamber. Frequencies of productively infected CD4⁺eGFP⁺ cells were compared between mono and- co-cultures (n = 8 subjects). (F) Supernatant levels of IL-5, sST2, IL-4, and IL-13 from CD4 mono-cultures and CD4/CD8 co-cultures at 5:1 (E:T). Experiments conducted with replication competent NL4-3_eGFP virus treated with protease inhibitor Darunavir are indicated by squares (n = 8 subjects), and experiments with replication-incompetent Env-defective NL4-3_D2eGFP are indicated by circles (n = 8 subjects). Flow cytometric data for all the samples were performed on day 3 post- HIV infection, and comparisons between co-cultures and mono-cultures were carried out using Wilcoxon matched-pairs signed rank test.

Fig 4. Infected CD4⁺ T cells co-cultured with CD8⁺ T cells accumulate reservoir-harboring cells compared to CD4⁺ T cells alone.

All in vitro infections were conducted using the replication competent NL4-3_eGFP under single cycle condition with protease inhibitor (Darunavir), with the exception of 2 donors who were infected with the (Env)-defective NL4-3_eGFP in the inducible provirus experiment (indicated by triangle symbol). CD4⁺ T cells from mono- and co-cultures with CD8⁺ T cells at 1:1 E:T ratio were FACS-sorted based on eGFP⁺ and eGFP^- expression 3 days post-infection (see S3 Fig). Data from 8 subjects are shown. (A) Integrated HIV-DNA copies/10⁶ cells are shown for sorted non-productively infected and uninfected (eGFP^-) and sorted-productively infected (eGFP⁺) CD4⁺ T cells from mono—or co-cultures. (B) Sorted eGFP^-CD4⁺ T from mono—and co-cultures were reactivated by CD3/28 stimulation three days post-infection and controlled for i) de novo infection with protease inhibitor (Darunavir) and ii) pre-integration events with integrase inhibitor (Dolutegravir). Inducible provirus was determined by the frequency of eGFP⁺ cells 24 hours post-reactivation using flow cytometry. (C) Levels of HIV transcripts on a per-integrated-provirus basis were calculated by dividing the copy numbers of cell-associated multiply-spliced Tat-Rev HIV-RNA per copy numbers of integrated HIV-DNA per cell. Comparisons between co-cultures and mono-cultures were carried out using Wilcoxon matched-pairs signed rank test.

Fig 5. Transcriptome analysis of infected CD4⁺ T cells reveals systematic immune-suppressive activity by CD8⁺ T cells.

All in vitro infections were conducted using the replication competent NL4-3_eGFP under single cycle condition with protease inhibitor (Darunavir). CD4⁺ T cells from mono- and co-cultures with CD8⁺ T cells at 1:1 E:T ratio were FACS-sorted based on eGFP⁺ and eGFP^- expression 3 days post-infection (see S3 Fig). Data from 8 subjects are shown. (A) Principal-component analysis (PCA) of RNA-seq data shows CD4⁺ T sorted subsets derived from CD4 mono-cultures (eGFP^-, light green; eGFP⁺; dark green) and CD4⁺ T sorted subsets derived from CD4/CD8 co-cultures (eGFP^-, light magenta; eGFP⁺; dark magenta). Each point represents a sample labelled with the donor ID. Female subjects are indicated by circles, and male subjects are indicated by squares. (B) Venn diagrams demonstrating differentially expressed genes (DEGs) in eGFP^- and eGFP⁺CD4⁺ T subsets in response to CD8⁺ T cells (DESeq2: padj <0.05 and fold change threshold >1.5 (0.585 log) or <0.67 (-0.585 log). Intersection shows DEGs in common between eGFP^- and eGFP⁺CD4⁺ T subsets in response to CD8⁺ T cells. (C) Barplot shows the pathways significantly downregulated in eGFP^- and eGFP⁺CD4⁺ T subsets by the CD8⁺ T cells obtained from the MSigDB database BioCarta collection, GSEA. A normalized enriched score (NES) lower than 0 corresponds to a pathway for which member genes are downregulated in CD4/CD8 co-cultures. The 21 pathways depicted presented a NES < -1.4 in either eGFP^- or eGFP⁺CD4⁺ T subsets. The color scale denotes the maximum and minimum normalized enrichment score (NES), and the nominal p value alongside the bars, estimates the statistical significance of the enrichment score for a single gene set. For leading edge genes of Fas, Th1/Th2, Inflammatory and G2/M pathways (see S4 Fig). (D) Effect of CD8⁺ T cells on GATA3 and GFI1 transcriptional regulator genes associated with the Th2 pathway of CD4⁺ T helper differentiation (DESeq2 library size normalized read counts).