Induction of contact-dependent CD8⁺ regulatory T cells through stimulation with staphylococcal and streptococcal superantigens

Abstract

The bacterial superantigen exotoxins of Staphylococcus aureus and Streptococcus pyogenes are potent stimulators of polyclonal T-cell proliferation. They are the causes of toxic shock syndrome but also induce CD25⁺ FOXP3⁺ regulatory cells in the CD4 compartment. Several studies have recently described different forms of antigen-induced regulatory CD8⁺ T cells in the context of inflammatory diseases and chronic viral infections. In this paper we show that bacterial superantigens are potent inducers of human regulatory CD8⁺ T cells. We used four prototypic superantigens of S. aureus (toxic shock syndrome toxin-1 and staphylococcal enterotoxin A) and Str. pyogenes (streptococcal pyrogenic exotoxins A and K/L). At concentrations below 1 ng/ml each toxin triggers concentration-dependent T-cell receptor Vβ-specific expression of CD25 and FOXP3 on CD8⁺ T cells. This effect is independent of CD4⁺ T-cell help but requires antigen-presenting cells for maximum effect. The cells also express the activation/regulatory markers cytotoxic T-lymphocyte antigen-4 and glucocorticoid-induced tumour necrosis factor receptor-related protein and skin homing adhesins CD103 and cutaneous lymphocyte-associated antigen. Superantigen-induced CD25⁺ FOXP3⁺ CD8⁺ T cells were as potent as freshly prepared naturally occurring CD4⁺ regulatory T cells in suppressing proliferation of CD4⁺ CD25⁻ T cells in response to anti-CD3 stimulation. Although superantigen-induced CD8⁺ CD25⁺ FOXP3⁺ express interleukin-10 and interferon-γ their suppressive function is cell contact dependent. Our findings indicate that regulatory CD8⁺ T cells may be a feature of acute bacterial infections contributing to immune evasion by the microbe and disease pathogenesis. The presence and magnitude of regulatory CD8⁺ T-cell responses may represent a novel biomarker in such infections. Superantigen-induced regulatory CD8⁺ T cells also have therapeutic potential.

Introduction

The superantigenic exotoxins of Staphylococcus aureus and Streptococcus pyogenes are among the most potent T-cell mitogens known, stimulating human lymphocytes at concentrations down to 10⁻⁹ m.¹ Over 30 such toxins have now been described including staphylococcal toxic shock syndrome toxin-1 (TSST-1) and staphylococcal enterotoxins (SE) A to R, the streptococcal pyrogenic exotoxins (SPE) A, C, G–M and streptococcal mitogenic exotoxin Z.¹

Superantigens trigger polyclonal activation of a substantial proportion of CD4⁺ and CD8⁺ T cells by binding the MHC class II molecule and the T-cell receptor (TCR) simultaneously at sites not involved in conventional antigen recognition. Superantigens may be classified by their interactions with MHC class II. One group bind the α-chain with (e.g. TSST-1) or without (e.g. SPEA) contact with antigenic peptide. Another group bind the β-chain (e.g. SPEC and SPE-K/L). A third group bind at both sites to cross-link MHC class II (e.g. SEA).¹ Binding at the TCR is, in most cases, through the TCR Vβ region, although some superantigens such as SEH, interact with the TCR Vα region.² Superantigens vary in their TCR Vβ specificity and this is determined primarily by interactions with the TCR Vβ CDR2 loop. Some superantigens are more TCR Vβ-specific than others as a result of interactions with other hyper-variable regions of the TCR Vβ region; CDR1, CDR3 and HV4. For example, TSST-1 is highly specific for TCR Vβ2 whereas SEB and SPEA each activate several structurally related TCR Vβ types (TCR Vβ1, 5.1, 8, 9, 22 and TCR Vβ3, 12, 13.1, 14 respectively) particularly at higher concentrations.³

The clinical syndrome of infection most clearly linked to superantigen production is toxic shock syndrome (TSS), in which superantigen-triggered polyclonal T-cell activation and systemic cytokinaemia occur. TSS is a rare complication of forms of infection where large quantities of toxin are produced, for example within materials such as wound packs or tampons, or in the context of necrotizing deep infections with associated toxaemia.^4,5 However, most infections by S. aureus and Str. pyogenes are mild or asymptomatic and the fact that most healthy adults have specific antibodies to many superantigens indicates that they are present in the course of clinically trivial episodes of infection.^6,7 Superantigen immunology research has tended to focus on the dramatic inflammatory responses to superantigens associated with TSS but have not addressed the question of how polyclonal T-cell activation could be advantageous to these organisms.

During infection, inflammatory responses are crucial to the control and clearance of the pathogen. Over the last 15 years it has become clear that regulatory arms of both the innate and adaptive immune systems serve to limit the extent and duration of inflammatory responses to prevent tissue damage. Among the best studied are regulatory T (Treg) cells.⁸ Treg cells can be broadly categorized as either naturally occurring or activation-induced. Naturally occurring Treg (nTreg) cells, characterized by expression of the fork-head transcription factor FOXP3 and the interleukin-2 receptor (IL-2R) α-chain CD25, make up around 2–5% of peripheral blood CD4⁺ T cells but are not found as a distinct population among CD8⁺ T cells in humans.^9–11 Activation-induced Treg (iTreg) cells develop when either CD4⁺ or CD8⁺ T cells encounter antigen in the periphery.^12,13 Analysis of iTreg cells is hampered by the fact that ‘regulatory’ markers including CD25, FOXP3, CD152 (CTLA-4) and CD357 [glucocorticoid-induced tumour necrosis factor receptor-related protein (GITR)] are also transiently expressed by activated non-regulatory cells. For these reasons, assessment of cells expressing regulatory markers after activation needs to include functional evaluation of regulatory activity.

Although much of the recent interest in iTreg cells has focused on CD4⁺ cells, antigen-specific CD8⁺ iTreg cells have been reported in the context of viral infections such as HIV and hepatitis C virus.^14,15 Joosten et al. described antigen-specific CD8⁺ iTreg cells in the context of mycobacterial infection.^16,17 Two studies have sought to determine the conditions under which polyclonal CD8⁺ iTreg cells could be generated and demonstrated that continuous stimulation with SEB, at a concentration 10 000 times more than that required to achieve half maximal proliferation of T cells, or with anti-CD3 monoclonal antibodies, is capable of inducing functionally regulatory CD8⁺ T cells expressing FOXP3.^18,19

We have recently reported that stimulation of CD4⁺ CD25⁻ T cells by bacterial superantigens results in TCR Vβ-specific expansion of CD4⁺ CD25⁺ FOXP3⁺ cells, which produce IL-10 and have functional regulatory activity.²⁰ This represents a potential mechanism by which S. aureus and Str. pyogenes may delay or prevent the adaptive immune responses that would otherwise clear these organisms from the nose or throat. Given that CD8⁺ T cells are also stimulated by bacterial superantigens we sought to determine whether superantigen-induced iTreg cells also develop in the CD8 compartment. We have used four superantigens (two of S. aureus and two of Str. pyogenes) chosen to exemplify the main groups based on interaction with MHC class II; TSST-1 (staphylococcal, α-chain and peptide), SEA (staphylococcal, β-chain), SPE-K/L (streptococcal, two β-chain interaction sites) and SPEA (streptococcal, α-chain).¹ In each case we demonstrate a TCR Vβ-specific, dose-dependent expansion of CD8⁺ CD25⁺ FOXP3⁺ cells which is independent of changes in the CD4⁺ compartment. Superantigen-induced CD8⁺ CD25⁺ FOXP3⁺ also express cell surface markers typical of Treg cells, including CD152, CD357 and CD103 and the skin homing receptor cutaneous lymphocyte-associated antigen (CLA). These cells are as potent as CD4⁺ nTreg cells in suppressing CD4⁺ T-cell proliferation and, although they produce IL-10 and interferon-γ (IFN-γ), suppression of CD4⁺ T-cell proliferation is independent of soluble factors and dependent on cell contact.

Methods

Antibodies

Aqua Live/Dead fixable cell stain was purchased from Invitrogen (Paisley, UK). Anti-human CD3-AlexaFluor®700, CD8-phycoerythrin/cychrome 5 (PE/Cy5), CD8-PE/Cy7, CD25-allophycocyanin/H7, CD103-FITC, FOXP3-AlexaFluor®647, FOXP3-AlexaFluor®488, HLA-DR-FITC, HLA-DP-FITC, HLA-DQ-FITC, IFN-γ-PE/Cy7 and IL-10-PE were purchased from BD Biosciences (Oxford, UK). Anti-human CD4-ECD and TCR Vβ1-PE, TCR Vβ1-FITC, TCR Vβ2-PE, TCR Vβ14-PE, TCR Vβ22-PE were obtained from Beckman Coulter (High Wycombe, UK). BioLegend® anti-human CD3-Pacific Blue™, CLA-Pacific Blue™, CD152-allophycocyanin and CD357-PE/Cy5 were purchased from Cambridge Bioscience (Cambridge, UK). CD127-eFluor®450 and CD152-PE were purchased from eBioscience Ltd. (Hatfield, UK).

Superantigens

The recombinant SPE-K/L (rSPE-K/L) was prepared as previously described.²⁰ The spe-k/l vector was constructed by Thomas Proft (University of Auckland, New Zealand). In brief, SPE-K/L was expressed as a thioredoxin-SPE-K/L fusion protein in Escherichia coli Novagen® OrigamiB(DE3)pLysS (Merck, Nottingham, UK). Thioredoxin was cleaved using r3C protease. The rTSST-1 was purchased from Sigma-Aldrich (Dorset, UK) and rSEA and rSPEA were obtained from Toxin Technology (Sarasota, FL).

Cell isolation

Peripheral blood mononuclear cells (PBMCs) were purified from healthy human donor whole blood using Ficoll-Plaque PLUS (GE Healthcare, Little Chalfont, UK) gradient centrifugation. The PBMCs were washed twice in PBS and resuspended in complete RPMI medium [RPMI-1640 medium containing 10% heat-inactivated fetal calf serum, 2 mm l-glutamine and 1% penicillin-streptomycin (Invitrogen)] at 1 × 10⁶ cells/ml.

Untouched CD4⁺ T cells were isolated from PBMCs using the CD4⁺ T-cell biotin-antibody cocktail and anti-biotin micro-beads from the CD4⁺ CD25⁺ Regulatory T Cell Isolation Kit (Miltenyi Biotec, Bisley, UK) as per the manufacturer's instructions. Negative selection was performed to isolate CD8⁺ T cells from PBMCs using the human CD8⁺ T Cell Isolation Kit (Miltenyi Biotec). Antigen-presenting cells (APCs) were prepared by irradiating (50 Gy) the CD4⁺-depleted PBMCs. Both CD4⁺ CD25⁺ and CD8⁺ CD25⁺ T cells were isolated by positive selection using the CD25 micro-beads supplied in the CD4⁺ CD25⁺ Regulatory T Cell Isolation Kit. The purity of isolated cells was determined as > 90% by flow cytometry (data not shown). The isolated cells were resuspended at 1 × 10⁶ cells/ml in complete RPMI medium.

Stimulation of cells

The PBMCs (1 × 10⁵ cells per well) were incubated with superantigen in complete RPMI medium in 96-well round-bottom plates for 3 days in a humidified incubator at 37° with 5% CO₂. For stimulation of isolated cells, 5 × 10⁴ CD4⁺ or CD8⁺ isolated T cells were incubated with superantigen in the presence of 5 × 10⁴ irradiated APCs as described above for stimulation of PBMCs. Intracellular cytokine experiments had brefeldin A (Sigma-Aldrich) added at a final concentration of 10 μg/ml to the stimulated cells 14 hr before antibody staining.

Surface and intracellular staining

Following stimulation with superantigen, the PBMCs were washed with FACS buffer (PBS, 0·5% BSA and 0·1% sodium azide) and stained for surface markers at 4° for 30 min. Cells were washed using FACS buffer and the Human FOXP3 Buffer Set (BD Biosciences) used to fix and permeabilize the cells. Permeabilized cells were stained with intracellular markers for 30 min at 4°. CD3 was stained following permeabilization to ensure all CD3⁺ cells were included. Isolated CD4⁺ and CD8⁺ cells following superantigen stimulation were stained as described above.

Suppression assays

The suppressive ability of cells was determined by assessing their ability to inhibit the proliferative responses of fresh donor T cells by dye dilution. The PBMCs were stimulated with SPE-K/L or TSST-1 (1 ng/ml) for 3 days. Stimulated PBMCs were washed twice and the superantigen-induced CD4⁺ CD25⁺ T cells and CD8⁺ CD25⁺ T cells were isolated as described above. CD4⁺ CD25⁻ and CD4⁺ CD25⁺ T cells were freshly isolated from the same donor. Responder cells were prepared by labelling the CD4⁺ CD25⁻ T cells with 2 × 10⁶m PKH67 Fluorescent Cell Linker (Sigma-Aldrich) according to the manufacturer's instructions. Isolated CD25⁺ cells (5 × 10⁴), responder cells (5 × 10⁴), irradiated APCs (5 × 10⁴) and anti-human CD3 OKT3 (0·5 μg/ml) (eBioscience Ltd) were added to 96-well round-bottom plates and incubated for 4 days. Assays were performed in triplicate. Cells were washed in FACS buffer before staining as described above. As FOXP3 staining was not performed, the cells were fixed and permeabilized with FACS Lysing Solution (BD Biosciences) and FACS Permeabilizing Solution 2 (BD Biosciences).

Transwell experiments were performed as above except that responder cells (2 × 10⁵), isolated CD25⁺ cells (2 × 10⁵), irradiated APCs (5 × 10⁴) and anti-human CD3 OKT3 (0·5 μg/ml) were incubated in a 24-well plate in the presence or absence of 0·4 μm transwell inserts (Corning, Amsterdam, the Netherlands).

Flow cytometric analysis

Stained cells were washed with FACS buffer and acquired with a three-laser LSR II bench top flow cytometer (BD Biosciences). FlowJo software (Tree Star Inc., Ashland, OR) was used to analyse the acquired data. To ensure the results were not due to antibody effects fluorescence minus one controls were performed for all antibody panels (data not shown).

Statistical analysis

Data were expressed as the mean ± SEM. Statistical analysis of the suppression data was performed using a paired Student's t-test. Significance is indicated as follows: *P < 0·05; **P < 0·01; ***P < 0·001.

Results

CD8⁺ T cells dose-dependently increase expression of CD25⁺ FOXP3⁺ following superantigen stimulation

Previously we showed that in vitro stimulation of CD4⁺ T cells with superantigens results in a dose-dependent increase in expression of CD25 and FOXP3.²⁰ To examine whether superantigen stimulation of CD8⁺ T cells similarly results in increased CD25 and FOXP3 expression, PBMCs from healthy donors were incubated with TSST-1, SEA, SPE-K/L or SPEA at a range of concentrations for 3 days and expression of CD25 and FOXP3 was examined by flow cytometry. For each toxin we observed a dose-dependent increase in the percentage of CD8⁺ T cells expressing CD25 and FOXP3 at concentrations above 100 pg for TSST-1 and SEA, above 10 pg for SPE-K/L and above 1 ng for SPEA, reflecting differences in the potency of these toxins¹ (Fig. 1).

Figure 1.

Superantigen stimulated CD8⁺ T cells have increased expression of the regulatory markers CD25 and FOXP3. Peripheral blood mononuclear cells (PBMCs) were stimulated with toxic shock syndrome toxin-1 (TSST-1), staphylococcal enterotoxin A (SEA), streptococcal pyrogenic exotoxin K/L (SPE-K/L) and SPEA for 3 days. The expressions of CD25 and FOXP3 in CD8⁺ and CD4⁺ T cells were measured by flow cytometry. Cells were gated on live/CD3⁺/CD8⁺ or live/CD3⁺/CD4⁺ expression. Data represent the mean ± SEM of three donors.

TCR Vβ-specific expression of CD25⁺ FOXP3⁺ in CD8⁺ T cells

To confirm that CD25 and FOXP3 expression in CD8⁺ T cells is a superantigen response we determined, for each superantigen, whether CD25 and FOXP3 expression was TCR Vβ-specific. To allow direct comparison of CD8⁺ responses with CD4⁺ responses and to determine whether CD8⁺ responses are independent of CD4⁺ responses, CD8⁺ and CD4⁺ T cells were isolated from PBMCs and equal numbers were stimulated at toxin concentrations from 10 fg/ml to 10 ng/ml in the presence of irradiated APCs. For each superantigen we compared CD25 and FOXP3 expression of the main TCR Vβ type targeted by that superantigen (TCR Vβ2 for TSST-1, TCR Vβ22 for SEA, TCR Vβ1 for SPE-KL and TCR Vβ14 for SPEA). For each superantigen, CD8⁺ T-cell expression of CD25 and FOXP3 was TCR Vβ-specific appropriately for the superantigen tested (Fig. 2a). At the higher superantigen concentrations, particularly for SEA, we observed some CD25⁺ FOXP3⁺ expression among T cells not of the non-targeted TCR Vβ and this is in keeping with the ability of superantigens such as SEA to target other TCR Vβ types at higher concentrations.³ In keeping with its lower potency as a superantigen, and the involvement of other TCR Vβ types, the highest concentration of SPEA used only resulted in CD25 and FOXP3 expression in 47% of TCR Vβ14⁺ CD8⁺ T cells. Superantigen-induced expression of CD25 and FOXP3 by CD8⁺ T cells was independent of the CD4⁺ T-cell response (Fig. 2a) and showed the same concentration–response relationship as CD4⁺ T cells (Fig. 2a,b).

Figure 2.

CD25⁺ FOXP3⁺ expression in CD8⁺ T cells is T-cell receptor (TCR) Vβ-specific. Isolated cells were stimulated with superantigen in the presence of irradiated antigen-presenting cells for 3 days and analysed by flow cytometry for TCR Vβ, CD25 and FOXP3 expression. Data represent the mean ± SEM of three donors. (a) CD8⁺ T cells from peripheral blood mononuclear cells stimulated with toxic shock syndrome toxin-1 (TSST-1), staphylococcal enterotoxin A (SEA), streptococcal pyrogenic exotoxin K/L (SPE-K/L) and SPEA express CD25⁺ FOXP3⁺ in a TCR Vβ-specific manner in the absence of CD4⁺ T cells. Events were gated for live/CD3⁺/CD8⁺ cells. (b) Isolated CD4⁺ T cells stimulated with TSST-1, SEA, SPE-K/L and SPEA have TCR Vβ-specific expression of CD25⁺ FOXP3⁺. Cells were gated on live/CD3⁺/CD4⁺ T cells.

CD8⁺ T-cell expression of CD25⁺ FOXP3⁺ following superantigen stimulation requires APCs for maximal expression

A defining feature of superantigenicity is that T-cell activation by superantigens is MHC class II dependent.²¹ To determine if the TCR Vβ-specific increased CD25⁺ FOXP3⁺ expression observed following superantigen stimulation was dependent on the presence of MHC class II, isolated CD8⁺ T cells were stimulated with superantigen in the presence and absence of irradiated APCs. For these and subsequent experiments, to allow unambiguous comparisons between superantigen-responding and non-responding cells, we used TSST-1 and SPE-K/L because of their specificity for TCR Vβ2 and TCR Vβ1, respectively.

Expression of CD25⁺ FOXP3⁺ was markedly dependent on the presence of APC (Fig. 3). Although some expression of CD25⁺ FOXP3⁺ was observed in the absence of APCs this is likely to be the result of presentation by MHC class II expressed on activated T cells (see Supplementary material, Fig. S1).

Figure 3.

Antigen-presenting cells (APCs) are required for optimal CD25⁺ FOXP3⁺ expression in isolated CD8⁺ T cells. Isolated CD8⁺ T cells were stimulated for 3 days with toxic shock syndrome toxin-1 (TSST-1) or streptococcal pyrogenic exotoxin K/L (SPE-K/L) in the presence or absence of irradiated APCs and analysed by flow cytometry. Data represent mean ± SEM of three donors. Live/CD3⁺/CD8⁺ cells were analysed for CD25⁺ FOXP3⁺ expression.

Superantigen-induced expression of regulatory and activation T-cell markers

Expression of regulatory and activation markers on CD8⁺ T cells was examined following TSST-1 or SPEK-L stimulation. A concentration of 1 ng/ml was chosen because at this concentration 100% of the superantigen-responsive TCR Vβ CD8⁺ T cells are CD25⁺ FOXP3⁺ (Fig. 2). The activation and regulatory markers examined were as follows. Expression of CD152 (CTLA-4) and CD357 (GITR) was examined because both are well-established Treg cell markers.^22–25 Expression of CD103 (αE integrin) has been described as a Treg activation marker because of its role in Treg cell migration and retention in the skin.^26–29 Although low expression of the IL-7 receptor, CD127, is characteristic of naturally occurring Treg cells,^30,31 Simonetta et al.³² recently described increased CD127 expression on activated Treg cells. The skin homing receptor CLA has previously been shown to be increased on T cells following superantigen stimulation.^33,34

Changes in these markers on TCR Vβ⁺-specific CD8⁺ T cells following stimulation with TSST-1 and SPE-K/L are shown in Fig. 4(a). Superantigen stimulated CD8⁺ T cells showed markedly increased expression of CD152, CD357, CD103 and CLA but expression of CD127 was unaltered. The nature of these changes in expression is illustrated for a single representative donor following TSST-1 stimulation in Fig. 4(b).

Figure 4.

Expression of regulatory and activation T-cell markers increases following superantigen stimulation. Peripheral blood mononuclear cells (PBMCs) were stimulated for 3 days with 1 ng/ml toxic shock syndrome toxin-1 (TSST-1) (a) and streptococcal pyrogenic exotoxin K/L (SPE-K/L) (b). Expression was measured by flow cytometry. Data represent the mean ± SEM of three donors after unstimulated expression has been subtracted from the respective groups. Expression of CD152, CD357, CD103 and CLA were all significantly increased on CD8⁺ T cells of the T-cell receptor (TCR) Vβ targeted by the superantigen used compared with other TCR Vβ types. Expression of CD127 remained low following stimulation. (c) Differences in regulation and activation markers between TCR Vβ2⁺ CD8⁺ T cells and TCR Vβ2⁻ following stimulation with TSST-1. Events displayed were gated on TCR Vβ2⁻ and TCR Vβ2⁺ for live/CD3⁺/CD8⁺ T cells. One representative experiment of three is shown.

CD8⁺ CD25⁺ FOXP3⁺ T cells produce the regulatory cytokine IL-10

To further characterize superantigen-induced CD25⁺ FOXP3⁺ cells in terms of potential suppressive function we assessed the production of protypic pro-inflammatory and anti-inflammatory cytokines IFN-γ and IL-10. We used intracellular staining for these cytokines to allow their production to be assessed at an individual cell level.

Intracellular cytokine staining relies on the use of brefeldin A to delay trafficking of cytokines from the cell. This also impairs trafficking of TCR and reduces TCR expression. As a consequence, expression of IL-10 and IFN-γ was measured in total CD8⁺ T cells following superantigen stimulation at a range of concentrations up to 1 ng/ml. A greater percentage of cells produced IL-10 than IFN-γ at lower concentrations of superantigen for both TSST-1 and SPE-K/L (Fig. 5a). To confirm that the IL-10 production was specific to the superantigen-induced CD25⁺ FOXP3⁺ cells, the percentage of CD25⁺ FOXP3⁺ cells expressing IL-10 was assessed. Interleukin-10 production was observed in a total of 70% and 80% of CD25⁺ FOXP3⁺ cells following stimulation with TSST-1 and SPE-K/L, respectively (Fig. 5b).

Figure 5.

Interleukin-10 (IL-10) is produced by CD8⁺ CD25⁺ FOXP3⁺ T cells following superantigen stimulation. Peripheral blood mononuclear cells (PBMCs) were stimulated with superantigen for 3 days. Brefeldin A was added 14 hr before staining for intracellular expression of IL-10 and interferon-γ (IFN-γ) and flow cytometry analysis. Graphs display the mean ± SEM of three donors. (a) CD8⁺ T cells dose-dependently produce IL-10 following simulation with toxic shock syndrome toxin-1 (TSST-1; top) and streptococcal pyrogenic exotoxin K/L (SPE-K/L; bottom). Cells were gated for live/CD3⁺/CD8⁺. (b) Superantigen-induced CD8⁺ CD25⁺ FOXP3⁺ produce IL-10. CD8⁺ CD25⁺ FOXP3⁺ T cells were analysed for IL-10 and IFN-γ production following gating on live cells and CD3⁺ cells.

CD8⁺ T cells following superantigen stimulation are functionally suppressive and suppression is contact dependent

To confirm that superantigen-stimulated CD8⁺ T cells function as suppressor T cells, PBMCs were stimulated with superantigen for 3 days, CD8⁺ CD25⁺ T cells were isolated and incubated with fresh labelled CD4⁺ CD25⁻ T cells, irradiated APCs and anti-CD3 OKT3 for 4 days. The ability of the CD8⁺ CD25⁺ T cells to inhibit proliferation of CD4⁺ CD25⁻ T cells was assessed and compared with freshly isolated CD4⁺ CD25⁺ (nTreg cells) and CD4⁺ CD25⁺ T cells isolated following superantigen stimulation. Proliferation of the CD4⁺ CD25⁻ T cells was normalized to allow comparison between separate experiments with different donors. CD8⁺ CD25⁺ T cells isolated following stimulation with either SPE-K/L or TSST-1 displayed the ability to inhibit the proliferation of fresh CD4⁺ CD25⁻ cells from the same donor stimulated with irradiated APCs and OKT3 (Fig. 6a). This suppression of CD4⁺ CD25⁻ proliferation was comparable to that seen with nTreg cells isolated on the same day as OKT3 stimulation and CD4⁺ CD25⁺ T cells isolated following 3 days stimulation with superantigen. There was no significant difference between the degrees of suppressive capability between the CD25⁺ groups.

Figure 6.

Superantigen-induced CD8⁺ CD25⁺ T cells are functionally suppressive by a contact-dependent mechanism. (a) CD8⁺ CD25⁺ T cells induced following stimulation with superantigens inhibit the proliferation of CD4⁺ CD25⁻ T cells. Peripheral blood mononuclear cells (PBMCs) were stimulated with either streptococcal pyrogenic exotoxin K/L (SPE-K/L) or toxic shock syndrome toxin-1 (TSST-1) for 3 days before isolation of superantigen-induced CD4⁺ CD25⁺ and CD8⁺ CD25⁺ T cells. Fresh responder CD4⁺ CD25⁻ from the same donor were labelled and stimulated with irradiated antigen-presenting cells (APCs) and OKT3 for 4 days. CD25⁺ cells were added 1 : 1 to stimulated responder cells and the ability to inhibit responder cell proliferation was measured by flow cytometry. Data represent the mean ± SEM of three donors. (b) The suppressive capability of superantigen-induced CD8⁺ CD25⁺ cells is contact dependent. Isolated CD25⁺ cells were separated from responder cells and irradiated APCs by a transwell insert and the ability to inhibit responder cell proliferation was assessed by flow cytometry. Data represent the mean ± SEM of three donors.

Superantigen-stimulated CD8⁺ T cells produce the anti-inflammatory cytokine IL-10 and are functionally suppressive (Figs 5 and 6a). To determine whether the suppressive capability of the superantigen-induced CD8⁺ CD25⁺ T cells is dependent on soluble factors or contact-dependent, transwell inserts were used to separate the isolated CD25⁺ cells from the stimulated responder cells (Fig. 6b). Superantigen-induced CD8⁺ CD25⁺ T cells were unable to inhibit the proliferation of labelled CD4⁺ CD25⁻ T cells when physically separated from the labelled CD4⁺ CD25⁻ T cells by the transwell insert.

Discussion

Bacterial superantigens are highly potent toxins. They induce half-maximal T-cell proliferation at concentrations between 0·02 and 10 pg/ml and have broad specificity for CD4⁺ and CD8⁺ T cells restricted only by TCR Vβ type.¹ These properties underlie TSS, a rare form of sepsis that occurs when microgram quantities of superantigens are produced at a focus of infection and high levels of pro-inflammatory cytokines are produced.^4,5 However, most adults have detectable superantigen antibodies, implying that these toxins are also active during common, minor streptococcal and staphylococcal infections.^6,7

Regulatory immune responses limit and terminate inflammatory immune responses during infection but could also result in failure of pathogen clearance by the immune system.³⁵ Antigen-specific CD4⁺ iTreg cells expressing FOXP3 and CD25 have been widely studied in this context and we recently proposed that superantigen-induced CD4⁺ iTreg cells with specificity limited only by TCR Vβ type could result in failure of the immune system to clear S. aureus and Str. pyogenes infection.²⁰

The phenomenon of activation-induced regulatory function has been explored predominantly in CD4⁺ T cells and a role for human CD8⁺ iTreg cells has only been explored in a limited range of settings. CD8⁺ T cells expressing CD103 and showing contact-dependent suppression have been described following allostimulation in the presence of anti-inflammatory cytokines and stimulation of virus-specific CD8⁺ cells with hepatitis C virus and influenza virus peptides.^13,15,29 Such cells have also been identified in patients with chronic viral infections or autoimmune diseases.^15,36 CD8⁺ T cells that suppress T-cell responses through contact-independent secretion of cytokines such as IL-10 and transforming growth factor-β have been generated in vitro and identified in patients with chronic viral infections such as HIV.^14,37,38 A third mechanism described for suppression by CD8⁺ CD28⁻ cells is APC-dependent through down-regulation of co-stimulatory ligands such as CD80 and CD86.³⁹

The work presented here extends these observations by showing that superantigen toxins produced in the course of an acute bacterial infection are potent inducers of CD8⁺ iTreg cells. Although we have only worked with four toxins these are broadly representative of the main structural families of bacterial superantigens and we conclude that the induction of CD8⁺ iTreg cells is likely to be a general effect of bacterial superantigens. We have shown that superantigen-induced expression of the regulatory markers CD25, FOXP3, CD152 (CTLA-4) and CD357 (GITR) is TCR Vβ-specific. This demonstrates that induction of CD8⁺ iTreg cells is a superantigen response rather than related to any other property of these toxins.

Superantigen-induced CD8⁺ iTreg cells are phenotypically similar to antigen-specific CD4⁺ iTreg cells and previously described CD8⁺ iTreg cells, which express CD25, FOXP3, CD152 (CTLA-4), CD357 (GITR), the adhesin CD103 and suppress via a contact-dependent mechanism.^13,16–19 Using mouse models to explore T-cell responses to the superantigens produced by human pathogens is problematic because the responses mounted by human and murine T cells are qualitatively and quantitatively different through differences in both MHC class II and TCR. The fact that antigen-specific CD8⁺ iTreg cells expressing CD103 have been demonstrated in several human disease states indicates that the effect we have observed in vitro is likely to occur in the course of infection.^28,29 Previously clinical studies of CD8⁺ iTreg cells have focused on chronic inflammatory processes and intracellular pathogens but our findings indicate that the nature of CD8⁺ iTreg cells should be studied systemically in TSS and locally during streptococcal and staphylococcal infection.

At the highest concentrations of toxin we have used we observe expression of FOXP3 and CD25 on 100% of superantigen-responsive TCR Vβ-specific CD8⁺ T cells. However, not all these cells express the other markers we assessed such as CD152 (CTLA-4) and CD357 (GITR). It is noteworthy therefore that although the totality of CD8⁺ CD25⁺ FOXP3⁺ T cells induced by superantigen stimulation are functionally suppressive this does not mean that all these cells are suppressive.

As far as we are aware this study provides the first evidence that an acute bacterial infection could be associated with induction of regulatory CD8⁺ T cells. Our findings are in keeping though with those of Mahic et al.,¹⁹ which describe conditions under which CD8⁺ iTreg cells could be induced including continuous stimulation with SEB, at a concentration of 1 μg/ml. However, this concentration is 10 000 times the half maximal proliferation concentration of this toxin and around 1000 times the equivalent concentrations used in our work. Furthermore no characterization of the superantigenicity of the response was made.

In summary we have described the ability of bacterial superantigens to induce potent, contact-dependent CD8⁺ Treg cells. This suggests for the first time that CD8⁺ iTreg cells may play a role in acute bacterial infection both as an immune evasion strategy for the microbe and in the pathogenesis of sepsis. The presence and magnitude of regulatory CD8⁺ T-cell responses should be explored in sepsis, including TSS and during other forms of streptococcal and staphylococcal infection. The CD8⁺ iTreg cells may represent a novel diagnostic or prognostic marker in such infections. The therapeutic potential of superantigen-induced CD8 iTreg cells also warrants further study.

Disclosures

The authors declare no conflicts of interest.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Figure S1. HLA surface expression is increasedin CD8⁺ T cells following superantigen stimulation.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.