HIV-1 infection is associated with depletion and functional impairment of Mycobacterium tuberculosis-specific CD4 T cells in individuals with latent tuberculosis infection

Abstract

Co-infection with HIV is the single greatest risk factor for reactivation of latent Mycobacterium tuberculosis infection (LTBI) and progression to active tuberculosis (TB) disease. HIV-associated dysregulation of adaptive immunity by depletion of CD4 T helper cells likely contributes to loss of immune control of LTBI in HIV-infected individuals, although the precise mechanisms whereby HIV infection impedes successful T cell-mediated control of Mtb have not been well defined. To further delineate mechanisms whereby HIV impairs protective immunity to M. tuberculosis (Mtb), we evaluated the frequency, phenotype, and functional capacity of Mtb-specific CD4 T cells in HIV-infected and HIV-uninfected adults with LTBI. HIV infection was associated with a lower total frequency of cytokine-producing Mtb-specific CD4 T cells, and preferential depletion of a discrete subset of Mtb-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells. Mtb-specific CD4 T cells in HIV-infected individuals expressed significantly higher levels of Ki67, compared with HIV-uninfected individuals, thus indicating recent activation and turnover of these cells in vivo. The ex vivo proliferative capacity of Mtb-specific CD4 T cells was markedly impaired in HIV-infected individuals, compared with HIV-uninfected individuals. Moreover, HIV infection was associated with increased Mtb Ag-induced CD4 T cell death ex vivo, indicating a possible mechanism contributing to impaired proliferative capacity of Mtb-specific CD4 T cells in HIV-infected individuals. These data provide new insights into the parameters of Mtb-specific CD4 T cell immunity that are impaired in HIV-infected individuals with LTBI, which may contribute to their increased risk of developing active TB disease.

Introduction

The vast majority of immunocompetent individuals infected with Mycobacterium tuberculosis (Mtb) never develop symptoms of clinical disease and are considered to have latent Mtb infection (LTBI). Although the correlates of protective immunity to Mtb have not been clearly defined, co-infection with human immunodeficiency virus type 1 (HIV) has long been recognized as the single greatest risk factor for reactivation of LTBI and development of TB disease. HIV-infected individuals not on effective antiretroviral therapy are at >20-fold higher risk of developing active TB disease than HIV-uninfected individuals (1–4); moreover, the increased risk of active TB is evident within the first year of HIV seroconversion (5), suggesting impairment to Mtb immunity occurs early after HIV infection. Approximately 11% of new TB cases occur in HIV-infected individuals worldwide, although the HIV prevalence among TB cases in high-burden countries in southern Africa exceeds 50% (2).

Although both innate immunity and Mtb-specific T cell immunity are clearly important in maintaining successful immune control of Mtb (6), the immune parameters that are modified by HIV co-infection and subsequently contribute to loss of immune control of Mtb infection have not been well defined. Chronic HIV infection is associated with dysregulation of innate and adaptive immunity, including decreased frequencies of dendritic cells (7), functional impairment of natural killer (NK) cells (8–10), persistent immune activation (11, 12), disrupted maturation and differentiation of memory T cells (13–17), and T cell exhaustion (18–20). Importantly, HIV infection leads to profound depletion of CD4 T helper cells, a critical component of the immune response to Mtb (21). The introduction of antiretroviral therapy (ART) has had a tremendous impact on reducing the risk of TB disease in HIV-infected individuals, with meta-analyses estimating a risk reduction of 67% (22, 23). However, HIV-infected individuals with preserved CD4 T cell counts remain at higher risk for development of TB disease, compared with HIV-uninfected individuals (24–27), thus providing compelling evidence that additional mechanisms of immune dysfunction contribute to the increased risk of TB disease in HIV-infected individuals.

Mounting evidence indicates Mtb-specific CD4 T cells producing IFN-γ are preferentially depleted in individuals who are co-infected with HIV (28–30), and that depletion of these cells is evident within one year of HIV seroconversion (28). A potential mechanism contributing to loss of Mtb-specific CD4 T cells has been suggested by reports that Mtb-specific CD4 T cells are preferentially infected by HIV (31), although additional mechanisms leading to depletion of Mtb-specific CD4 T cells have yet to be defined. In addition to depletion of Mtb-specific CD4 T cells, impaired mycobacteria-specific CD4 T cytokine production has been described in cells isolated from bronchoalveolar lavage of HIV-infected individuals, compared with HIV-uninfected individuals (32). Moreover, IL-2 production capacity by Mtb-specific T cells has been inversely correlated with HIV viral load (33), suggesting Mtb-specific T cells are further impaired in the setting of high levels of viremia and immune activation that are characteristic of chronic HIV infection. Taken together, these data suggest HIV infection impairs protective Mtb-specific CD4 T cell responses early after HIV infection, thus potentially contributing to increased susceptibility to active TB disease.

To further define mechanisms of HIV-associated dysregulation of Mtb-specific CD4 T cell immunity, we measured the frequency, phenotypic profiles, and functional capacity of Mtb-specific CD4 T cells in HIV-infected and HIV-uninfected adults with LTBI in South Africa. We identified a discrete subset of cytokine⁺ Mtb-specific CD4 T cells that are preferentially depleted in HIV-infected individuals, and that are predictive of poor proliferative capacity of Mtb-specific CD4 T cells in HIV-infected individuals. Mtb-specific CD4 T cells were more activated and more susceptible to Ag-induced cell death in HIV-infected individuals, compared with HIV-uninfected individuals, thus providing further insights into the mechanisms of Mtb-specific CD4 T cell immune dysfunction in the setting of HIV co-infection.

Materials and Methods

Study participants and sample collection

Blood samples were collected from HIV-uninfected and HIV-infected and individuals with LTBI in the Cape Town region of South Africa. Individuals with LTBI included in the study were asymptomatic, with no symptoms of active TB disease (cough, fever, weight loss, night sweats), had no previous history of diagnosis or treatment for active TB disease, and had a positive response to CFP-10 and/or ESAT-6 pooled peptides by IFN-γ production following overnight stimulation of whole blood. HIV infection status was determined using the Alere Determine™ HIV-1/2 Ag/Ab Combo test, a rapid finger-prick test that detects the presence of p24 Ag and antibodies to both HIV-1 and HIV-2. Absolute CD4 T cell counts and plasma HIV-1 viral loads were measured in HIV-infected participants, all of whom were antiretroviral therapy naïve at the time of analysis. (At the time the study was done, South Africa had a policy of providing antiretroviral therapy only to patients with CD4 counts <200 cells/mm³.) Blood samples from all participants were collected in sodium heparin Vacuette® tubes for analysis of Ag-specific T cell cytokine production capacity, phenotype, and proliferative capacity, as described below.

Whole blood intracellular cytokine staining (ICS) assay

Immediately after collection, whole blood was incubated with the following Ags: pooled, overlapping 15-mer peptides corresponding to the sequences of CFP-10, ESAT-6, and human cytomegalovirus (HCMV) pp65 (1.25 μg/ml/peptide), and Mtb purified protein derivative (PPD; Staten Serum Institut; 10 μg/ml). The HCMV pp65 peptide pool was obtained through the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH (34–36). Blood incubated with no Ag served as a negative control; blood incubated with staphylococcal enterotoxin B (SEB; Sigma-Aldrich; 1 μg/ml) served as a positive control. After incubation for 3 hours at 37°C, Brefeldin A (BFA) was added (10 μg/ml) and the incubation continued for an additional 5 hours at 37°C. Red blood cells were lysed and white cells fixed with FACS Lysing Solution (BD Biosciences). Cells were then cryopreserved in freezing medium (50% RPMI, 40% fetal bovine serum [FBS] and 10% DMSO), and stored in liquid nitrogen until use. Upon thawing, cells were washed in Perm/Wash Buffer (BD Biosciences) and stained with monoclonal Abs in two separate flow cytometry panels. The first panel included anti-CD3 Pacific Blue (UCHT1), anti-CD8 PerCP-Cy5.5 (SK-1), anti-IFN-γ Alexa Fluor 700 (B27), anti-IL-2 FITC (5344.111), all from BD Biosciences, and anti-CD4 Qdot605 (S3.5; Life Technologies), anti-TNF-α PE-Cy7 (Mab11; eBiosciences), and anti-PD-1 PE (EH12.2H7; BioLegend). The second panel included anti-CD3 Brilliant Violet 605™ (OKT3), anti-CD4 Brilliant Violet 570™ (RPA-T4), anti-IFN-γ Brilliant Violet 711™ (4S.B3), anti-Ki67 PE (Ki67), all from BioLegend, and anti-CD8 PerCP-Cy5.5 (SK-1), anti-IL-2 FITC (5344.111), anti-CCR7 BD Horizon™ PE-CF594 (150503), and anti-CD45RA BD Horizon™ BV421 (HI100), all from BD Biosciences. For both Ab panels, cells were stained for 1 hr at 4°C, washed with Perm/Wash Buffer, and acquired on a BD LSRII flow cytometer.

PBMC Proliferation assay

Peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood by density centrifugation with Ficoll-Hypaque (Sigma-Aldrich), within 4 hours of collection. Freshly isolated PBMC were labeled with 0.5 μg/ml CellTrace™ Oregon Green® 488 carboxylic acid diacetate, succinimidyl ester (OG; Life Technologies), as previously described (37). Cells were stimulated for 6 days with PPD (1 μg/ml), pooled CFP-10 and ESAT-6 peptides (0.1 μg/ml/peptide), or SEB (0.1 μg/ml). Negative controls consisted of cells incubated with media alone. No exogenous cytokines were added to the proliferation assay cultures. On day 6, cytokine production capacity of proliferating cells was determined by re-stimulation of cells with the same Ags for the final 6 hrs of the assay as follows: PPD-stimulated cells were re-stimulated with PPD (5 μg/ml); CFP-10/ESAT-6-stimulated cells were re-stimulated with CFP-10/ESAT-6 peptide pool (1 μg/ml/peptide), and SEB-stimulated cells were re-stimulated with SEB (1 μg/ml). BFA (10 μg/ml) was added for the final 5 hours of the assay. Cells were then washed with PBS and stained with LIVE/DEAD Fixable Violet Dead Cell Stain (Vivid; Life Technologies). Cells were then fixed with BD FACS Lysing Solution, permeabilized with Perm/Wash Buffer, and stained with the following monoclonal Abs for 1 hr at 4°C: anti-CD3 APC-H7 (SK7), anti-CD8 PerCP-Cy5.5 (SK-1), anti-IFN-γ Alexa Fluor 700 (B27), anti-IL-2 FITC (5344.111), all from BD Biosciences, and anti-CD4 Qdot605 (S3.5; Life Technologies) and anti-TNF-α PE-Cy7 (Mab11; eBiosciences). Stained cells were then acquired on a BD LSRII flow cytometer.

Flow cytometry data analysis

Multiparameter flow cytometry data were analyzed using FlowJo software (v9.7.6; Treestar). Doublet cell populations were excluded by plotting forward scatter area (FSC-A) versus forward scatter height (FSC-H). Single-stained anti-mouse Ig, κ beads (BD Biosciences) were used to calculate compensation. Lymphocyte populations were gated by morphological characteristics; viable cells were identified as Vivid^lo cells. Combinations of cytokine-producing cells were determined using Boolean gating in FlowJo. Background cytokine production in the negative control of ICS assays was subtracted from each Ag-stimulated condition; background proliferation in the negative control condition of the PBMC proliferation assay was subtracted from Ag-stimulated conditions.

Data analysis and statistics

Responses in whole blood ICS assays were considered positive if the frequency of cytokine⁺ cells was >0.01% of CD4 T cells, and greater than 3 times the frequency of cytokine⁺ CD4 T cells in the unstimulated condition. Phenotypic analysis was conducted only for individuals with positive responses to a given Ag in the ICS assay. Responses in 6-day proliferation assays were considered positive if the frequency of proliferating (OG^lo) CD4 T cells in the Ag stimulated cells was greater than 3 times the frequency of proliferating CD4 T cells in the negative control. Cytokine production capacity of Mtb-specific proliferating CD4 T cells was evaluated in individuals with a positive proliferative response.

Statistical testing was performed using GraphPad Prism v7.0b software. For analysis of whole blood ICS assay and phenotyping data, differences between HIV-uninfected and HIV-infected groups was assessed using the Mann-Whitney U test. The Bonferroni method was applied to the analysis of multiple cytokine producing subsets to correct for multiple comparisons. P values that did not maintain significance after applying the Bonferroni correction are denoted in the figures by “#”. A simple linear regression was calculated to predict proliferative capacity based on the frequency and proportion of ex vivo Mtb-specific cytokine⁺ CD4 T cell subsets. Differences in CD4 T cell viability in Ag-stimulated versus unstimulated conditions of the proliferation assay were assessed using the Wilcoxon matched-pairs signed rank test.

Ethical approval

All subjects provided written informed consent for participation in the study, which was approved by the Human Research Ethics Committee at the University of Cape Town, the Western Cape Province Department of Health, and the Emory University Institutional Review Board.

Results

Study Participants

Blood samples were collected from 52 HIV-uninfected and 20 HIV-infected participants with LTBI enrolled in the Cape Town region of South Africa (Table 1). The median absolute CD4 T cell count of HIV-infected participants was 471 cells/μl; the median viral load was 16,736 HIV-1 RNA copies/ml plasma. There was no difference in sex characteristics in the two groups of participants, although HIV-uninfected participants were younger than HIV-infected participants.

Table I.

Characteristics of study population

Participant Group	n	Age, years (range)a	Sex (% male)	CD4 Count, cells/μlc (IQR)d	HIV viral load, copies RNA/ml plasmac (IQR)d
HIV− LTBI	52	28 (18 – 50)	40	N/A	N/A
HIV+ LTBI	20	33 (21 – 46)b	25	471 (286 – 585)	16,736 (4,994 – 44,869)

^aValues denote mean age in years (range)

^bp < 0.05, compared with HIV-uninfected LTBI

^cValues denote median

^dIQR – Interquartile range

N/A, not applicable

HIV infection is associated with depletion of a distinct subset of Mtb-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells

Chronic HIV infection is characterized by global depletion of CD4 T cells. To determine whether Mtb-specific CD4 T cells are preferentially depleted in HIV-infected individuals with LTBI, we stimulated whole blood with PPD, and measured expression of IFN-γ, IL-2, and TNF-α by flow cytometry (Fig 1a). The total frequency of PPD-specific CD4 T cells expressing IFN-γ, IL-2, and/or TNF-α was lower in HIV-infected individuals, compared with HIV-uninfected individuals with LTBI (Fig 1b).

Figure 1. Distinct subsets of cytokine⁺ Mtb-specific CD4 T cells are depleted in HIV-infected individuals with LTBI.

Whole blood from HIV-uninfected (n=34) and HIV-infected (n=20) individuals with LTBI was stimulated with PPD, HCMV pp65 pooled peptides, and SEB for 8 hours. Intracellular expression of IFN-γ, IL-2, and TNF-α was measured by flow cytometry. (A) Representative flow cytometry data from an HIV-uninfected individual with LTBI (top row), and an HIV-infected individual with LTBI (bottom row). Plots are shown gated on CD3⁺CD4⁺ T cells. Grey dots represent the total cytokine-negative CD4 T cell population; black dots represent CD4 T cells producing IFN-γ, IL-2 and/or TNF-α. (B) Composite data of the total frequency of PPD-specific CD4 T cells producing IFN-γ, IL-2 and/or TNF-α in HIV-uninfected and HIV-infected individuals with LTBI. Horizontal lines represent the median. Differences were compared using a Mann-Whitney U test. (C – E) Comparison of the frequencies of PPD-specific (C), HCMV pp65-specific (D), and SEB-stimulated (E) CD4 T cells producing IFN-γ, IL-2 and/or TNF-α (white boxes: HIV-uninfected LTBI; grey boxes: HIV-infected LTBI). Boxes represent the median and interquartile ranges; whiskers represent the 10^th and 90^th percentiles. Data are shown after subtraction of background cytokine production in the unstimulated negative control condition. Differences in the frequencies of each cytokine-producing CD4 T cell subset between HIV-uninfected and HIV-infected individuals were assessed using a Mann-Whitney U test. # indicates p values that did not maintain significance after applying the Bonferroni correction for multiple comparisons. P values are shown for cytokine subsets with a median frequency of ≥0.01% of CD4 T cells in both groups of participants.

To determine whether particular cytokine⁺ subsets of PPD-specific CD4 T cells were depleted in HIV-infected individuals, we next evaluated the contribution of each cytokine producing subset to the total PPD-specific CD4 T cell response in HIV-infected and HIV-uninfected individuals with LTBI. In both groups, polyfunctional cells producing IFN-γ, IL-2, and TNF-α simultaneously constituted the dominant subset of Mtb-specific CD4 T cells (Fig 1c). Compared with HIV-uninfected individuals, HIV-infected individuals had significantly lower frequencies of one discrete population of Mtb-specific CD4 T cells: those producing IFN-γ and TNF-α, but lacking IL-2 production (IFN-γ⁺IL-2⁻TNF-α⁺) (Fig 1c). Similar results were found following stimulation of whole blood with CFP-10 and ESAT-6 pooled peptides (Fig S1a), and with TB10.4 pooled peptides (data not shown). Moreover, depletion of Mtb-specific IFN-γ⁺IL-2⁻TNF-α⁺ cells was unique to CD4 T cells, as there was no evidence for depletion of any cytokine producing subset of Mtb-specific CD8 T cells in HIV-infected individuals (Fig S1b).

We next asked if the depletion of IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells in HIV-infected individuals was Ag dependent. There was no evidence of depletion of this particular cytokine⁺ subset of HCMV pp65-specific CD4 T cells in HIV-infected individuals, measured concurrently with Mtb-specific responses in the same assay in the same individuals (Fig 1d). Moreover, HIV-infected individuals had a higher frequency of IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells following stimulation with the superantigen SEB (Fig 1e), compared with HIV-uninfected individuals, thus indicating there is no inherent defect in IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells from HIV-infected individuals to respond to mitogenic TCR stimulation. These data indicate that depletion of Mtb-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells in HIV-infected individuals with LTBI is particular to Mtb Ag stimulation.

In addition to measuring the frequencies of cytokine⁺ Mtb-specific CD4 T cells circulating ex vivo, we also evaluated the proportion of each cytokine⁺ subset contributing to the total Mtb-specific CD4 T cell response. Consistent with the ex vivo frequencies of cytokine⁺ Mtb-specific CD4 T cell subsets, the proportion of PPD-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells was significantly lower in HIV-infected individuals, compared with HIV-uninfected individuals with LTBI (Fig S2a). There was no difference between HIV-infected and HIV-uninfected individuals in the proportion of HCMV pp65 or SEB-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells contributing to the total cytokine-positive response (Fig S2b, c).

HIV infection modifies activation status but not memory differentiation profiles of Mtb-specific CD4 T cells

The above data indicate HIV infection is associated with depletion of distinct cytokine⁺ subsets of Mtb-specific CD4 T cells in individuals with LTBI. To determine if the depletion of this cytokine⁺ subset is reflective of a shift in effector and central memory T cell distribution in HIV-infected individuals, we evaluated the memory differentiation profiles of Mtb-specific CD4 T cells by analysis of CD45RA and CCR7 expression (Fig 2a). In both HIV-uninfected and HIV-infected individuals with LTBI, PPD-specific CD4 T cells displayed a predominately CCR7⁻CD45RA⁻ effector memory phenotype, followed by CCR7⁺CD45RA⁻ central memory cells (Fig 2b). There were no significant differences in the memory differentiation profiles of PPD-specific CD4 T cells between HIV-infected and HIV-uninfected individuals with LTBI.

Figure 2. HIV infection modifies activation status, but not memory differentiation or exhaustion phenotype, of PPD-specific CD4 T cells in individuals with LTBI.

Whole blood from individuals with LTBI was stimulated as described in Figure 1. Expression of CD45RA, CCR7, PD-1, and Ki67 was measured by flow cytometry on total CD4 T cells and Ag-specific CD4 T cells. (A) Representative flow cytometry data of CD45RA and CCR7 expression from an HIV-uninfected individual with LTBI (top row), and an HIV-infected individual with LTBI (bottom row). Plots are shown gated on CD3⁺CD4⁺ T cells. Grey dots represent the total cytokine-negative CD4 T cell population; black dots represent CD4 T cells producing IFN-γ and IL-2. (B) Composite data of the memory differentiation phenotype of PPD-specific CD4 T cells in HIV-uninfected (n=23) and HIV-infected individuals (n=16) with LTBI. Boxes represent the median and interquartile ranges; whiskers represent the 10^th and 90^th percentiles. (C) Representative flow cytometry data of PD-1 expression from an HIV-uninfected individual with LTBI (top row), and an HIV-infected individual with LTBI (bottom row). Plots are shown gated on CD3⁺CD4⁺ T cells. Grey dots represent the total cytokine-negative CD4 T cell population; black dots represent CD4 T cells producing IFN-γ, IL-2, and TNF-α. (D) MFI of PD-1 expression on total CD4 T cells and cytokine⁺ PPD-specific CD4 T cells (E) in HIV-uninfected (n=33) and HIV-infected individuals (n=20) with LTBI. (F) Representative flow cytometry data of Ki67 expression from an HIV-uninfected individual with LTBI (top row), and an HIV-infected individual with LTBI (bottom row). Plots are shown gated on CD3⁺CD4⁺ T cells. Grey dots represent the total cytokine-negative CD4 T cell population; black dots represent CD4 T cells producing IFN-γ and IL-2. (G) Composite data of Ki67 expression by PPD-specific CD4 T cells in HIV-uninfected (n=23) and HIV-infected individuals (n=16) with LTBI. (H) Composite date of Ki67 expression by HCMV pp65-specific CD4 T cells in HIV-uninfected (n=12) and HIV-infected individuals (n=16) with LTBI. Horizontal lines in panels D, E, G, and H represent the median. Differences in panels B, D, E, G, and H were compared using a Mann-Whitney U test.

HIV infection has been associated with T cell dysfunction and upregulation of markers associated with immune exhaustion, including programmed cell death 1 (PD-1) receptor (18–20). To evaluate whether immune exhaustion in chronic HIV infection extends to Mtb-specific T cells, we measured expression of PD-1 on total CD4 T cells and cytokine⁺ PPD-specific CD4 T cells (Fig 2c). Consistent with previous reports (18), PD-1 expression was significantly upregulated on the total CD4 T cell population in HIV-infected individuals, compared with HIV-uninfected individuals (Fig 2d). However, PD-1 expression was not significantly higher on PPD-specific CD4 T cells in HIV-infected individuals, compared with HIV-uninfected individuals with LTBI (Fig 2e). These data suggest that, as measured by PD-1 expression, Mtb-specific CD4 T cells in HIV-infected individuals with LTBI are not more exhausted than in HIV-uninfected individuals.

We next evaluated the activation status of CD4 T cells by measurement of Ki67, a nuclear protein expressed by proliferating cells but not by resting cells (38, 39). Ki67 expression was significantly upregulated in PPD-specific CD4 T cells from HIV-infected individuals, compared with HIV-uninfected individuals (Fig 2f, g). While there was evidence of increased cell cycling of PPD-specific CD4 T cells in HIV-infected individuals, Ki67 expression in HCMV pp65-specific CD4 T cells was not different between HIV-infected and uninfected individuals (Fig 2h). Taken together, these data suggest that HIV infection is associated with increased turnover of Mtb-specific CD4 T cells in vivo in an Ag-dependent manner.

HIV infection is associated with significant impairment of Mtb-specific CD4 T cell proliferative capacity

The ability of Ag-specific T cells to proliferate upon secondary encounter with Ag is a characteristic trait of highly functional memory T cells. To evaluate whether HIV infection modulates Mtb-specific T cell proliferative capacity, we stimulated Oregon Green (OG)-labeled PBMCs with PPD and SEB for 6 days and measured the percentage of proliferating (OG^lo) CD4 T cells by flow cytometry (Fig 3a). While the overall proliferative capacity of CD4 T cells to SEB stimulation was approximately 2-fold lower in HIV-infected individuals, compared with HIV-uninfected individuals, the proliferative capacity of PPD-specific CD4 T cells was even more markedly reduced at 12-fold lower in HIV-infected individuals with LTBI, compared with HIV-uninfected individuals (Fig 3b). Moreover, the median percentage of CD4 T cells proliferating to CFP-10/ESAT-6 pooled peptides was 45-fold lower in HIV-infected individuals, compared with HIV-uninfected individuals (Fig S3a, b), thus demonstrating profound impairment of Mtb-specific CD4 T cell proliferative capacity in HIV-infected individuals. There was no relationship between Mtb-specific CD4 T cell proliferative capacity and either HIV-1 viral load or absolute CD4 T cell count (data not shown).

Figure 3. Mtb-specific CD4 T cell proliferative capacity is significantly impaired in HIV-infected individuals with LTBI.

Freshly isolated PBMCs from HIV-uninfected (n=39) and HIV-infected (n=17) individuals with LTBI were labeled with Oregon Green (OG) and incubated with PPD and SEB for 6 days. Viable proliferating cells were defined as Vivid^loCD3⁺CD4⁺OG^lo cells. (A) Representative flow cytometry data of proliferating PPD-specific CD4 T cells from an HIV-uninfected individual with LTBI (top row) and an HIV-infected individual with LTBI (bottom row). Plots are shown gated on Vivid^loCD3⁺CD4⁺ T cells. Percentages indicate the frequency of proliferating (OG^lo) CD4 T cells. (B) Proliferative capacity of PPD-specific and SEB stimulated CD4 T cells in HIV-uninfected and HIV-infected individuals with LTBI. Horizontal lines represent the median. Differences were compared using a Mann-Whitney U test. (C – E) Simple linear regression plots between the following: total ex vivo frequency of cytokine⁺ PPD-specific CD4 T cells and proliferative capacity (C); ex vivo frequency of PPD-specific IFN-γ⁺IL-2⁺TNF-α⁺ CD4 T cells and proliferative capacity (D); ex vivo frequency of PPD-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells and proliferative capacity (E). (F, G) Simple linear regression plots between the ex vivo proportion of PPD-specific IFN-γ⁺IL-2⁺TNF-α⁺ CD4 T cells and proliferative capacity (E), and the ex vivo proportion of PPD-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells and proliferative capacity. Dotted lines in panels C – G represent the 95% confidence bands of the best-fit line.

The percentage of Ag-specific proliferating T cells could be reflective of the input number of Ag-specific cells in the proliferation assay. To address this possibility, we used a simple linear regression to determine whether the frequency of PPD-specific CD4 T cells ex vivo was predictive of the frequency of PPD-specific proliferating CD4 T cells. Although we observed a positive correlation between the total ex vivo frequency of cytokine⁺ PPD-specific CD4 T cells and proliferative capacity, the frequency of two distinct subsets of PPD-specific CD4 T cells, IFN-γ⁺IL-2⁺TNF-α⁺ and IFN-γ⁺IL-2⁻TNF-α⁺, were stronger predictors of proliferative capacity than the total frequency of cytokine-producing CD4 T cells (Fig 3c– e). The ex vivo frequencies of the other PPD-specific CD4 T cell cytokine⁺ subsets did not predict proliferative capacity (data not shown). The qualitative CD4 T cell response to PPD was analyzed as well to determine the contribution of each cytokine producing subset to the total PPD response. While the ex vivo frequency of PPD-specific IFN-γ⁺IL-2⁺TNF-α⁺ CD4 T cells was predictive of proliferative capacity, the proportion of polyfunctional cells, generally thought to be functionally superior, was not a predictor of proliferative capacity (Fig 3f). By contrast, the proportion of PPD-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells ex vivo was the only cytokine subset to significantly predict PPD-specific proliferative capacity (Fig 3g and data not shown). These data indicate that depletion of a discrete subset of cytokine⁺ of PPD-specific CD4 T cells is an independent predictor of poor PPD-specific T cell proliferative capacity, and further demonstrate the functional relevance of depletion of Mtb-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells in HIV-infected individuals with LTBI.

Proliferating Mtb-specific CD4 T cells maintain cytokine production capacity

To further define the functional capacity of proliferating Mtb-specific CD4 T cells, we evaluated cytokine production by re-stimulation of OG-labeled cells with PPD for the final 6 hours of the proliferation assay. Intracellular expression of IFN-γ, IL-2, and TNF-α was measured by proliferating (OG^lo) CD4 T cells in HIV-uninfected and HIV-infected individuals with a positive proliferative response (Fig 4a). Among individuals with a positive proliferative response, PPD-specific OG^lo cells in both HIV-uninfected and HIV-infected individuals maintained Ag-specific cytokine production capacity, predominately IFN-γ and TNF-α (Fig 4b). A small subset of OG^lo PPD-specific CD4 T cells maintained Ag-specific polyfunctional cytokine production capacity (IFN-γ⁺IL-2⁺TNF-α⁺), which was surprisingly higher in HIV-infected individuals, compared with HIV-uninfected individuals (Fig 4b). These data indicate that, although the frequency of PPD-specific proliferating CD4 T cells is low in HIV-infected individuals, the few CD4 T cells that are capable of proliferating also maintain cytokine production capacity at levels comparable to HIV-uninfected individuals.

Figure 4. Mtb-specific proliferating CD4 T cells maintain cytokine production capacity in individuals with LTBI.

Proliferation assays were performed as described in Figure 3. On day 6 of incubation, cells were re-stimulated with PPD for 6 hours and assessed for intracellular expression of IFN-γ, IL-2, and TNF-α. (A) Representative flow cytometry data of cytokine production capacity by proliferating PPD-specific CD4 T cells from an HIV-uninfected individual with LTBI (top row) and an HIV-infected individual with LTBI (bottom row). Flow plots in the left column are gated on Vivid^loCD3⁺CD4⁺ cells (total viable CD4 T cells); flow plots in the right column are gated on Vivid^loCD3⁺CD4⁺OG^lo cells (proliferating PPD-specific CD4 T cells). (B) Summary data of the percentage of proliferating PPD-specific CD4 T cells producing the indicated subsets of IFN-γ, IL-2 and/or TNF-α. Cytokine production capacity was evaluated for individuals with PPD-specific proliferative responses above background (n=28 HIV-uninfected and n=8 HIV-infected individuals with LTBI). (C) Representative flow cytometry data of proliferation and cytokine production by PPD-stimulated CD4 T cells from an HIV-uninfected individual with LTBI (top row) and an HIV-infected individual with LTBI (bottom row). Flow plots are shown gated on Vivid^loCD3⁺CD4⁺ cells (total viable CD4 T cells). (D) Summary data of the percentage of PPD-specific CD4 T cells producing the indicated subsets of IFN-γ, IL-2 and TNF-α following a 6-hour re-stimulation of cells with PPD on day 6 of the proliferation assay (n=29 HIV-uninfected and n=17 HIV-infected individuals with LTBI). Boxes in panels B and D represent the median and interquartile ranges; whiskers represent the 10^th and 90^th percentiles. Differences in the frequencies of each cytokine⁺ CD4 T cell subset between HIV-uninfected and HIV-infected individuals were assessed using a Mann-Whitney U test.

Approximately half of HIV-infected individuals did not have a detectable PPD-specific CD4 T cell proliferative response. To address the possibility that PPD-specific CD4 T cells are present in the 6-day assay that maintain cytokine production but lack proliferative capacity, we evaluated PPD-specific cytokine production by total viable CD4 T cells, including both proliferating and non-proliferating cells. PPD-specific cytokine production was largely restricted to OG^lo CD4 T cells that had proliferated, with little evidence of PPD-specific cytokine production by non-proliferating (OG^hi) CD4 T cells (Fig 4c). Overall, the frequency of PPD-specific cytokine⁺ CD4 T cells was significantly reduced in HIV-infected individuals, compared with HIV-uninfected individuals (Fig 4d), and was reflective of the reduced frequency of PPD-specific proliferating CD4 T cells in HIV-infected individuals (Fig 3b). Similar results were observed with CFP-10/ESAT-6 pooled peptides (Fig S3c, d), thus suggesting that Mtb-specific CD4 T cells that lack proliferative capacity in HIV-infected individuals do not remain functionally viable throughout the 6-day stimulation period.

Mtb Ag stimulation induces CD4 T cell death in HIV-infected individuals with LTBI

The above data suggest that the low level of Mtb-specific proliferating CD4 T cells detectable in HIV-infected individuals may be due to loss of these cells via Ag-induced cell death during the 6-day assay, either before the cells have the opportunity to proliferate, or early after PPD-specific CD4 T cells have undergone initial cycles of proliferation. To address this possibility, we analyzed cell staining patterns with Vivid, an amine-reactive viability dye, on day 6 of the proliferation assay.

As shown in Figure 5a, CD4 T cells that did not proliferate and are non-viable can be identified as OG^hiVivid^hi cells; cells that proliferated initially and are subsequently non-viable by day 6 are identified as OG^loVivid^hi cells. Following stimulation with PPD, HIV-uninfected individuals had a small but significantly higher frequency of proliferated, non-viable CD4 T cells, compared with HIV-infected individuals (Fig 5b), indicative of activation-induced cell death. By contrast, a significantly higher frequency of non-proliferating, non-viable CD4 T cells were observed in HIV-infected individuals, compared with HIV-uninfected individuals (Fig 5c). Stimulation with PPD induced a significantly higher frequency of non-proliferating, non-viable CD4 T cells in HIV-infected individuals only; there was no difference in the viability of non-proliferating CD4 T cells in unstimulated and PPD-stimulated PBMCs in HIV-uninfected individuals (Fig 5d). Significantly increased frequencies of non-proliferating, non-viable CD4 T cells were similarly observed in HIV-infected individuals following stimulation with CFP-10/ESAT-6 pooled peptides (Fig 5d), thus suggesting induction of CD4 T cell death is consistent across different Mtb Ags, and is not unique to the PPD Ag preparation. Moreover, induction of CD4 T cell death was Mtb-specific, as there was no evidence of induction of CD4 T cell death in HIV-infected individuals following non-specific T cell activation with SEB (Fig 5e). Together these data suggest that impaired Mtb-specific CD4 T cell proliferation in HIV-infected individuals is associated with induction of cell death early following stimulation with Mtb Ags, prior to the opportunity of these cells to proliferate.

Figure 5. Mtb Ag stimulation induces CD4 T cell death in HIV-infected individuals with LTBI.

Proliferation assays were conducted as described in Figure 3. CD4 T cell viability of proliferating and non-proliferating cells was assessed on day 6 by analysis of Vivid and OG staining. (A) Representative flow cytometry data of the viability of proliferating (OG^lo) and non-proliferating (OG^hi) CD4 T cells in an HIV-uninfected individual with LTBI (top row) and an HIV-infected individual with LTBI (bottom row). Plots are shown gated on CD3⁺CD4⁺ T cells. (B) Percentage of proliferating, non-viable (OG^loVivid^hi) CD4 T cells and percentage of non-proliferating, non-viable (OG^hiVivid^hi) CD4 T cells (C) following stimulation of PBMCs with PPD for 6 days. Horizontal lines represent the median. Differences in the frequencies of non-viable CD4 T cells in HIV-uninfected and HIV-infected individuals were assessed using a Mann-Whitney U test. (D – F) Paired comparison of the frequency of non-proliferating, non-viable CD4 T cells between unstimulated (no Ag) and PPD-stimulated PBMCs (D), unstimulated and CFP-10/ESAT-6 peptide pool-stimulated PBMCs (E), and unstimulated and SEB-stimulated PBMCs (F). HIV-uninfected individuals are represented by open circles, HIV-infected individuals are represented by filled circles. Differences in panels D – F were assessed using the Wilcoxon matched-pairs signed rank test.

Discussion

To begin to define potential impairments in Mtb-specific CD4 T cell function in HIV-infected individuals, we assessed the frequency, phenotype, and functional capacity of Mtb-specific CD4 T cells in a cohort of HIV-infected and HIV-uninfected South Africa adults with LTBI. We found that HIV infection is associated with an overall lower frequency of Mtb-specific CD4 T cells, with significant depletion of a discrete subset of Mtb-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells. While HIV infection did not modify the memory differentiation profile of Mtb-specific CD4 T cells, HIV-infected individuals had a significantly higher proportion of activated, cycling Mtb-specific CD4 T cells, compared with HIV-uninfected individuals, thus suggesting recent exposure to mycobacterial Ags in vivo. Lastly, the proliferative capacity of Mtb-specific CD4 T cells was significantly impaired in HIV-infected individuals, which was associated with increased CD4 T cell death following in vitro mycobacterial Ag stimulation of PBMCs from HIV-infected individuals.

Previous studies have suggested that frequencies of IFN-γ⁺ Mtb-specific CD4 T cells are reduced in HIV-infected individuals, compared with HIV-uninfected individuals (28, 40), and are not fully restored following suppression of HIV viremia by antiretroviral therapy (41, 42). Here, we extend these findings using multi-parameter flow cytometry to further define populations of Mtb-specific CD4 T cells producing IFN-γ in combination with other two additional Th1 cytokines, TNF-α and IL-2, in individuals with LTBI. Polyfunctional Ag-specific T cells capable of producing multiple cytokines simultaneously have previously been associated with superior functional capacity (43), and enhanced control of viral and bacterial infections (44–46), thus we initially hypothesized that HIV-infected individuals with LTBI would have a lower frequency of polyfunctional Mtb-specific CD4 T cells than HIV-uninfected individuals. Contrary to this, we found that the frequency of polyfunctional Mtb-specific CD4 T cells was not significantly different between HIV-infected and HIV-uninfected individuals with LTBI, consistent with recent reports (30, 47). Instead, we found that depletion of Mtb-specific CD4 T cells in HIV-infected individuals was most striking in the IFN-γ⁺IL-2⁻TNF-®⁺ subset, a finding that was consistent across multiple mycobacterial Ags (PPD, and pooled peptides spanning CFP-10, ESAT-6, and TB10.4). Depletion of this particular subset of cytokine-producing CD4 T cells was observed following Mtb Ag stimulation, although there was no evidence of depletion of this subset of CD4 T cells in the same individuals following stimulation with HCMV pp65 or SEB. Further studies of the phenotypic profile of Mtb-specific IFN-γ⁺IL-2⁻TNF-®⁺ CD4 T cells are warranted to identify factors contributing to depletion of this subset in HIV-infected individuals.

By analysis of the ex vivo frequency of cytokine⁺ subsets of Mtb-specific CD4 T cells and proliferative capacity within the same individuals, we determined that both the frequency and proportion of Mtb-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells were significant predictors of proliferative capacity, thus providing further evidence of the functional significance of the depletion of this particular subset of Mtb-specific CD4 T cells in HIV-infected individuals. A recent report by Riou et al comparing the frequency of cytokine producing CD4 T cell subsets in HIV-infected and HIV-uninfected individuals with LTBI reported selective depletion of Mtb-specific CD4 T cells producing IFN-γ only, in the absence of IL-2 and TNF-α (30). The difference in the particular cytokine⁺ Mtb-specific CD4 T cell subset that is preferentially depleted in HIV-infected individuals in our study versus Riou et al could be due to several factors, including different participant characteristics, as well as technical differences in ICS assays used to measure Mtb-specific T cell responses (incubation time, Ag preparations used, and the addition of co-stimulatory Abs during the Ag stimulation). Nevertheless, both studies are consistent in reporting that distinct subsets of cytokine⁺ Mtb-specific CD4 T cells are preferentially depleted in HIV-infected individuals, thus suggesting there may be characteristics of some Mtb-specific CD4 T cells that may predispose them to preferential depletion in the setting of HIV/Mtb co-infection.

In addition to cytokine production capacity, Ag-specific T cells subsets can be characterized by combinatorial expression of surface molecules that further define their memory T cell differentiation profiles, activation status, and functional programs (48, 49). To address the hypothesis that HIV co-infection skews memory CD4 T cell profiles in individuals with LTBI, we evaluated the memory phenotype of Mtb-specific CD4 T cells by expression of CD45RA and CCR7, two receptors that identify T cells subsets with different lymphoid homing potential and effector functions (50). We found that Mtb-specific CD4 T cells are predominately effector memory (CD45RA⁻CCR7⁻) in both HIV-infected and HIV-uninfected individuals with LTBI. These results are consistent with previous reports, using the differentiation markers CD45RO and CD27, that Mtb-specific CD4 T cells in HIV-infected and HIV-uninfected individuals with LTBI are largely early differentiated (CD45RO⁺CD27⁺) and late differentiated (CD45RO⁺CD27⁻) memory cells (30, 47). Taken together, these data suggest that HIV-associated dysregulation of memory T cell differentiation (51) is not pervasive to circulating Mtb-specific CD4 T cells at the stage of LTBI.

Although we did not see modulation of Mtb-specific CD4 T cell memory differentiation profiles in HIV-infected individuals, we did see evidence of increased cell turnover in vivo, as evidenced by increased Ki67 expression, a nuclear protein expressed by cycling cells but absent in resting cells (38, 39). Ki67 expression by Mtb-specific CD4 T cells was approximately 10-fold higher in HIV-infected individuals, compared with HIV-uninfected individuals. These data suggest that mycobacterial Ag load may be higher in HIV-infected individuals with LTBI, compared with HIV-uninfected individuals, and that recent exposure to mycobacterial Ags in vivo may be driving Mtb-specific CD4 T cell entry into the cell cycle. Interestingly, Ki67 expression by HCMV pp65-specific CD4 T cells was not different between HIV-infected and HIV-uninfected individuals, thus providing further evidence that increased turnover of CD4 T cells in HIV-infected individuals with LTBI is pathogen specific. In another study utilizing cryopreserved PBMCs, increased level of the activation marker HLA-DR was found in Mtb-specific CD4 T cells form HIV-infected individuals with LTBI, compared with HIV-uninfected individuals (52). Taken together, these data indicate increased immune activation of Mtb-specific CD4 T cells in HIV-infected individuals with LTBI. Importantly, T cell activation has been described as a correlate of risk of active TB disease in BCG-vaccinated infants and adolescents (53). Future, prospective studies will be required to determine whether increased activation of Mtb-specific CD4 T cells in individuals with LTBI is indicative of increased mycobacterial Ag exposure in vivo and predictive of progression to active TB disease.

An important feature of Ag-specific memory T cells is their ability to proliferate upon secondary exposure to Ag. Using a sensitive flow cytometry-based dye-dilution proliferation assay with freshly isolated PBMCs, we established that the proliferative capacity of Mtb-specific CD4 T cells is markedly impaired in HIV-infected individuals with LTBI. Despite the presence of cytokine⁺ Mtb-specific CD4 T cells in HIV-infected individuals detected ex vivo in the whole blood ICS assay, these cells were not capable of robust proliferation in the 6-day proliferation assay. The reduced Mtb-specific T cell proliferative capacity observed in HIV-infected individuals with LTBI is consistent with a previous study reporting reduced lymphoproliferation, as detected by [³H] thymidine incorporation, in heat-killed Mtb-stimulated PBMC cultures from HIV-infected active TB patients (29). In HIV-uninfected individuals, impaired Mtb-specific T cell proliferative capacity has been associated with active TB disease, while HIV-uninfected individuals with LTBI maintain robust proliferative capacity (37, 54). Our data now provide compelling evidence that, unlike HIV-uninfected individuals, the proliferative capacity of Mtb-specific CD4 T cells in HIV-infected individuals is impaired at an earlier stage in Mtb infection, during latency and prior to development of active TB disease. Additional studies are required to determine whether specific interventions, such as the addition of exogenous cytokines (such as IL-2) or blockade of immunoregulatory receptors (such as PD-1 or CTLA-4), can restore TB-specific T cell proliferative capacity, as has been demonstrated previously for restoring HIV-specific T cell proliferation (18, 55, 56).

The presence of detectable cytokine⁺ Mtb-specific CD4 T cell responses ex vivo in HIV-infected individuals, combined with the lack of proliferative capacity, suggested the possibility that Mtb-specific CD4 T cells from HIV-infected individuals were undergoing cell death during the 6-day stimulation period. To test this hypothesis, we analyzed CD4 T cell labeling of the amine-reactive viability dye used at the end of the proliferation assay. While the percentage of non-viable CD4 T cells did not change in HIV-uninfected individuals following stimulation with Mtb Ags, the proportion of non-viable CD4 T cells increased significantly in HIV-infected individuals. The increase in non-viable CD4 T cells in HIV-infected individuals was specific for mycobacterial Ag stimulation (including both PPD and CFP-10/ESAT-6 pooled peptides), as mitogenic stimulation of PBMCs with SEB did not lead to increased CD4 T cell death. A higher predisposition of CD4 T cells to Mtb-induced apoptosis has been reported in HIV-uninfected patients with active TB disease, and has been associated with increased levels of TFG-β in supernatants and increased FasL expression by CD4 T cells, following Mtb stimulation in vitro of PBMCs from active TB patients (57–59). Thus, the increased frequency of non-viable CD4 T cells in Mtb Ag-stimulated PBMCs that we observed in HIV-infected individuals with LTBI is similar to that reported in HIV-uninfected active TB patients, and further identify parameters of CD4 T cell immunity that are dysregulated in HIV-infected individuals, which may contribute to their increased risk of developing active TB disease. One possible mechanism contributing to Mtb-induced CD4 T cell death in HIV-infected individuals comes from reports that Mtb-specific CD4 T cells are preferentially infected by HIV (31), thus perhaps increasing susceptibility of Mtb-specific CD4 T cells to death and subsequent depletion. Moreover, in the proliferation assays from HIV-infected individuals, there may be some viral replication and infection of CD4 T cells during the 6-day culture, thus further contributing to CD4 T cell death. Future studies utilizing Mtb MHC class II tetramers to sort Mtb-specific CD4 T cell populations and further analyze by RNA sequencing will facilitate delineation of mechanisms involved in Mtb-induced CD4 T cell death in HIV-infected individuals.

Some limitations should be taken into consideration when interpreting the results of our study. This is a cross-sectional study of HIV-infected and HIV-uninfected individuals with LTBI, and a relatively small sample size within each group. We do not have longitudinal follow-up data of the HIV-infected individuals to determine whether the Mtb-specific CD4 T cell profiles we observed, including depletion of cytokine-producing Mtb-specific CD4 T cells and lack of proliferative capacity, are predictors of progression to active TB disease. Moreover, we profiled Mb-specific CD4 T cells circulating in peripheral blood, and not at the site of Mtb infection in the lung. The observed depletion of Mtb-specific IFN-γ⁺IL-2⁻TNF-α⁺ CD4 T cells circulating in peripheral blood of HIV-infected individuals could be reflective of increased trafficking of these cells to the site of Mtb infection in the lung. Lastly, our ability to detect the presence of Mtb-specific CD4 T cells was based on the ability of these cells to produce Th1 cytokines, thus Mtb-specific CD4 T cells that do not produce these cytokines would be missed in our evaluation of cell populations that are preferentially depleted in HIV-infected individuals. The use of MHC Class II tetramers specific for Mtb epitopes is vital in future studies to perform detailed characterization of the phenotypic and functional signatures of Mtb-specific CD4 T cells in HIV-infected individuals, without relying on a particular functional readout for detection.

In conclusion, we provide new insights into how HIV infection dysregulates Mtb-specific CD4 T cell immunity, including preferential depletion of distinct subsets of Mtb-specific CD4 T cells, induction of CD4 T cell death, and diminished Mtb-specific CD4 T cell proliferation. While these impairments of Mtb-specific CD4 T cell immunity have previously been observed in HIV-uninfected patients with active TB disease, we now report that these impairments are evident in HIV-infected individuals earlier in Mtb infection, prior to the onset of active TB disease. Identification of the mechanisms whereby HIV infection impairs protective T cell immunity to Mtb will be of crucial importance to facilitate development of effective TB vaccines and targeted immunotherapeutic interventions and treatment of individuals co-infected with HIV and Mtb to promote sustained control of Mtb infection and prevent progression to active TB disease.