HIV-1 Infection Impairs Regulatory T-Cell Suppressive Capacity on a Per-Cell Basis

Abstract

The impact of CD4⁺ regulatory T cells (Tregs) on human immunodeficiency virus type 1 (HIV-1) pathogenesis remains incompletely understood. Although it has been shown that Tregs can be infected with HIV-1, the consequences of infection on a per-cell basis are still unknown. In vitro HIV-GFP infected and noninfected Tregs were isolated by flow-based cell-sorting to investigate Treg suppressive capacity and gene expression profiles. Our data show that HIV-1-infected Tregs were significantly less suppressive than noninfected Tregs and demonstrated down-regulation of genes critical to Treg function. This impaired function may have detrimental consequences for the control of generalized immune activation and accelerate HIV disease progression.

The exact impact of CD4⁺ thymus-derived regulatory T cells (Tregs) on human immunodeficiency virus type 1 (HIV-1) pathogenesis, including potential benefits through suppression of HIV-1-associated generalized immune activation and control of HIV replication vs possible deleterious effects on viral control via suppression of HIV-1-specific immunity, remains incompletely understood. While human Tregs can serve as targets for HIV-1 infection, their fate is unclear.

We have previously shown that bulk Tregs isolated from chronic HIV progressor blood retain their suppressive function in vitro when compared to Tregs isolated from healthy donors [1]. However, considering the low frequency of infected Tregs in peripheral blood of HIV-1-infected individuals (<0.7% of peripheral Tregs [2]), it is unlikely that the effects of HIV-1 infection can readily be assessed directly ex vivo from peripheral blood. Previous studies investigated the function of bulk in vitro infections on Treg function and found slightly contradictory results [2, 3]. In addition, the consequences of HIV-1 infection on Tregs on a per-cell basis have not been investigated to date, as viable infected Tregs of high purity could not be separated and tested for their function in prior studies. The current study focused on thymus-derived human Tregs and results may not be directly applicable to peripherally derived (induced) Tregs. We here used a vesicular stomatitis virus envelope glycoprotein (VSVg)-pseudotyped HIV-green fluorescent protein (GFP) reporter to determine the impact of HIV-1 infection on Tregs by investigating the suppressive function and gene expression profiles of flow-sorted HIV-1-positive and HIV-1-negative Tregs.

METHODS

Isolation and Culture of Regulatory T cells

Human CD4⁺ T-cell-enriched PBMC were isolated by density centrifugation from peripheral blood of healthy donors (median age = 25.5 years; interquartile range = 23–31 years) using a RosetteSep isolation kit (STEMCELL technologies) and stained with anti-CD3-Phycoerythrin-Cyanine-7 (PE-Cy7) (clone SK7), CD4-fluorescein isothiocyanate (FITC) (clone RPA-T4), CD25-allophycocyanin (APC) (clone BC96), CD127-phycoerythrin (PE) (clone hIL-7R-M21). Fresh CD3⁺CD4⁺CD25⁺ CD127^low Tregs were sorted on a FACSAria (BD Biosciences), activated with anti-CD3/anti-CD28 coated microbeads (Invitrogen) at a 1:1 bead-to-cell-ratio and cultured in X-VIVO15 media (Lonza) supplemented with 10% human serum (Gemcell), penicillin/streptomycin (50 U/mL) and IL-2 (300 U/mL).

HIV-GFP Infection of Tregs

The VSVg-pseudotyped HIV-GFP was produced by co-transfecting HEK 293T cells (ATCC) with a VSVg expression vector (Harvard Gene Therapy Initiative) and R5-tropic NL4-3-IRES-GFP [4] using Lipofectamine 2000 (Invitrogen). Cells were infected on day 5 (MOI = 1) by spin infection in the presence of polybrene.

Treg Suppression Assay

On day 2 post-infection, Tregs were labeled with LIVE/DEAD Fixable Violet Dead Cell Stain Kit (Invitrogen) and stained with anti-CD3-PE-Cy7. Viable GFP^pos-infected Tregs and GFP^neg-noninfected Tregs were flow-sorted on a FACSAria.

Live autologous CD3⁺ responder T cells were flow-sorted from cryopreserved PBMC, labeled with CellTrace Violet (Invitrogen) and co-cultured with HIV-1-infected or noninfected Tregs at different responder to Treg ratios, in the presence of anti-CD2/anti-CD3/anti-CD28 microbeads (Miltenyi Biotec) at a 1:1 bead-to-cell-ratio. After 4 days of co-culture, cells were stained with anti-CD3-PECy7, anti-CD4-APC (clone SK3) and anti-CD8-Alexa Fluor 700 (clone RPA-T8). Data were acquired on a LSRII flow-cytometer (BD Biosciences) and analyzed with FlowJo (Treestar).

The proliferation of the responder T cells was quantitated using the FlowJo software proliferation platform and the percent of suppression was calculated by comparison with the activated responder T-cell responders alone.

Gene Expression Profiling

Transcriptional profiling of 6 matched flow-sorted HIV-1-infected and uninfected Tregs was established using the digital multiplexed nCounter GX Human Immunology Kit containing 511 immunology-related genes (nanoString Technologies). Normalizations and quality controls were performed using R packages (R Development Core Team). Unsupervised analysis was performed on normalized data using principal component analysis and hierarchical clustering analysis. Differentially expressed genes were identified using Limma: linear model microarray analysis software package where genes were ranked by t statistic using a pooled variance. Differentially expressed genes were identified on the basis of false discovery rate (FDR) adjusted P value. The FDR was controlled using the Benjamini and Hochberg algorithm. The pathways, functions, and interactive networks analysis was performed on differentially expressed genes using Ingenuity Pathways Analysis (IPA 7.0; http://www.ingenuity.com/) and cytoscape software packages.

RESULTS AND DISCUSSION

To understand the consequences of HIV-1 infection on regulatory T cells, we first investigated the infection rate of CD4⁺ Tregs isolated from HIV-negative donors (Figure 1A, left panel) using a VSVg-pseudotyped HIV-GFP reporter in vitro. We used a pseudotyped virus in order to reach high infection rates regardless of the CCR5 expression on Tregs [2, 3] and to obtain sufficient numbers of infected Tregs for the experiments outlined. Indeed the Treg HIV-1 infection rates reached a median of 8.35% (interquartile range: 4.92%–10.85%; Figure 1A, middle panel) and provided us with the unique opportunity to isolate viable infected Tregs (Figure 1A, right panel) in large enough quantities to perform further characterization in functional assays.

Figure 1.

Impact of HIV-1 infection on Treg function. A, Example (left panel) of dot plots and gating strategy used to sort CD4⁺ CD25⁺ CD127^low regulatory T cells. Frequency of HIV-GFP expression (middle panel) in regulatory T cells (n = 12) 2 days after infection with VSVg-pseudotyped HIV-GFP at a MOI of 1. The box and whiskers plot represents the median, 25 and 75 percentiles (box), min and max values (whiskers). Example (right panel) of dot plot and gating strategy used to sort GFP_pos HIV-1-infected (solid gate) and GFP^neg-HIV-1-uninfected (hatched gate) Tregs 2 days post-infection. B, Example of activated CD8⁺ (left panel) and CD4⁺ (right panel) T-cell proliferation followed by CellTrace Violet dilution after 4 days of culture alone (gray histogram), in the presence of matched HIV-1-infected Tregs (Solid histogram) or uninfected Tregs (blue histogram) at a ratio of 4:1 Responder T cells to Tregs. C, Frequency of suppression of activated CD8⁺ (left panel) and CD4⁺ (right panel) T cells when co-cultured in the presence of matched HIV-1-infected Tregs (green dots) or uninfected Tregs (blue dots) at a ratio of 4:1 responder T cells to Tregs. P value was calculated using the Wilcoxon matched-pairs signed rank test and considered significant when inferior to .05. D, Example of percent of suppression of activated CD8⁺ (left panel) and CD4⁺ (right panel) T cells when co-cultured in the presence of HIV-1-infected Tregs (green hatched line) or uninfected Tregs (blue hatched line) at different Responder T cell to Treg ratios. Abbreviations: GFP, green fluorescent protein; HIV-1, human immunodeficiency virus type 1; VSVg, vesicular stomatitis virus envelope glycoprotein.

We next determined the suppressive capacity of matched sorted HIV-1-infected and -uninfected Tregs and found HIV-1-infected Tregs to be significantly less potent in suppressing autologous CD8⁺ and CD4⁺ T-cell proliferation than uninfected CD4⁺ Tregs (Figure 1B–D), suggesting that HIV-1 infection impairs Treg suppressive capacity. These data confirm and extend results from a recent study by Pion et al [3], which found that HIV-1 dysregulates Treg suppression. However, our results are unique in that we were able to isolate and separately investigate pure HIV-1-infected and -uninfected Treg populations, on which our findings are based. These 2 studies slightly contradict an earlier study by Moreno-Fernandez et al [2] that reported no impairment in Treg function after HIV-1 infection of Treg cultures. This discrepancy is likely explained by the fact that our study investigated pure flow-sorted infected Tregs, whereas Moreno et al used bulk infected Tregs, which could have diluted out the impact of HIV-1 infection on individual Tregs due to low frequencies of infected Tregs in the bulk culture, a fact the authors allude to in the paper discussion. Functional impairment of Tregs in autoimmune diseases results in chronic inflammation. In the context of HIV-1, such impairment may promote HIV-associated generalized immune activation and inflammation. Notably, we did not find a difference in the suppressive function of bulk Tregs isolated from healthy donors and chronic HIV-1 progressors in our previous study [1]. This is likely explained by the fact that the frequency of infected Tregs is very low at this stage of infection [2] and is not sufficient to impact the function of bulk Tregs. However it has been suggested that in tissues of HIV-1-infected individuals, viral burden and Treg enrichment is enhanced. In these tissue compartments increased frequencies could have an impact on the capacity of Tregs to control HIV-1 infection in macrophages and conventional T cells [5], thereby facilitating disease progression. Further studies are warranted to better understand the consequences of HIV-1 infection of Tregs in vivo in the periphery and the tissues of human study subjects.

In order to investigate how HIV-1 impairs Treg function, we next compared the gene expression profiles of matched uninfected and HIV-1-infected Tregs using an ultrasensitive multiplexed nanoString gene expression assay. Principal component (PC) analysis revealed that uninfected and HIV-1-infected Tregs form 2 distinct clusters (Figure 2A), which were primarily separated on PC2, demonstrating that HIV-1 modifies the Treg transcriptome following infection. Supervised analysis identified statistically significant differential expression of 23 genes between the 2 groups on the basis of multiple test corrected P value (FDR ≤ 0.05; Figure 2B). Pathway enrichment analysis depicted significant down-regulation of pathways linked to T-cell receptor (TCR; PTPRC, CD3E, IKBKE), which may decrease the ability of Tregs to respond to activation stimuli, CD27 (IKBKE, TRAF5, CD27) [6], which was found to be positively correlated with Treg suppressive function and PPAR-gamma (TRAF6, STAT5A, IKBKE, PDGFR) [7], which was reported to enhance Treg function. Several other genes involved in Treg lineage and function were significantly down-regulated (ETS1, TRAF6, STAT5A, LEF-1) [8–11], although a number of genes known to inhibit Treg function were up-regulated (IL18RAP, LTA, IL7, BCL6) [12–15] after HIV-1 infection and may explain their impaired suppressive function. The cell cycle inhibitor p21 (CDKN1A) was also over-expressed in Tregs after HIV-1 infection and may have anti-apoptotic effects. Pion et al reported a decrease in FOXP3 protein expression following HIV-1 infection that we did not observe at the gene expression level; nevertheless, our results suggest a possible destabilization of FOXP3 by down-regulation of ETS-1 [8]. Moreover, LEF-1, which acts in synergy with FOXP3 to activate the Treg transcriptional signature and enhances occupancy of its genomic targets [11], as well as STAT5a [10], which binds to the FOXP3 gene, were down-regulated. Overall our data suggest that HIV-1 infection of Tregs does not affect a single pathway that could potentially be individually compensated but likely multiple pathways simultaneously. Another relevant future analysis to complement our data set could be to investigate matched infected CD4⁺ Tregs and CD4⁺ conventional T cells (Tconvs). In addition to our current findings, such comparison could illustrate which Treg-specific-genes are differentially expressed after HIV-1-infection and may warrant further studies. However, several molecules we identified are not exclusive to Tregs (eg, ETS-1, LEF-1, CD27, etc) and a Treg-specific analysis could have masked the regulation of these interesting molecules in the present study.

Figure 2.

Transcriptional changes induced in Tregs following HIV-1 infection. A, Unsupervised analysis of HIV-1-infected (green dots) and uninfected Treg (blue dots) gene expression performed on normalized data by Principal Component Analysis. Principal component analysis was performed on normalized data generated using a three step process including background correction using negative spike-in controls, signal correction based on positive spikes in controls and housekeeping genes to bring all chips to similar expression levels. B, Heat map of significantly differentially expressed genes following HIV-1 infection in Tregs. Columns represent the samples, with rows representing genes. Gene expression levels are shown as a pseudocolor scale (−3 to 3) with red denoting high expression level and green denoting low expression level. Genes with FDR ≤ 0.05 were considered differentially expressed. Abbreviations: FDR, false discovery rate; HIV-1, human immunodeficiency virus type 1.

In conclusion, Treg infection using the HIV-GFP reporter allows for isolation and characterization of HIV-1-infected Tregs. HIV-1 alters Treg suppressive capacity by regulating multiple critical genes and may play a detrimental role in HIV-1 immunopathogenesis. Targeting regulatory pathways may thereby represent a viable approach for future immunotherapy aimed at reversing detrimental HIV-1 associated immune activation.