Streptococcal Pyrogenic Exotoxin A-Stimulated Monocytes Mediate Regulatory T-Cell Accumulation through PD-L1 and Kynurenine

Bacterial superantigens (SAgs) are exotoxins that promote a fulminant activation of the immune system. The subsequent intense release of inflammatory cytokines often results in hypotension, shock, and organ failure with high mortality rates. In the current paradigm, the direct and simultaneous binding of SAgs with T-cell receptor (TCR)-bearing Vβ regions and conserved structures on major histocompatibility complex class II (MHC class II) on antigen-presenting cells (APCs) induces the activation of both cell types. However, by crosslinking MHC class II molecules, APCs can be activated by SAgs independently of T lymphocytes. Recently, we showed that streptococcal pyrogenic exotoxin A (SPEA) of Streptococcus pyogenes stimulates an immunogenic APC phenotype with upregulated costimulatory molecules and inflammatory cytokines. Additionally, we revealed that SPEA triggers immunosuppressive programs in monocytes that facilitate the accumulation of regulatory T cells (Tregs) in in vitro monocyte/CD4+ T-cell cocultures. Immunosuppressive factors include anti-inflammatory interleukin 10 (IL-10), co-inhibitory surface molecule programmed cell death 1 ligand 1 (PD-L1), and the inhibitory indoleamine 2,3-dioxygenase (IDO)/kynurenine effector system. In the present study, we investigated the underlying mechanism of SPEA-stimulated monocyte-mediated accumulation of Tregs. Blood-derived monocytes from healthy donors were stimulated with SPEA for 48 h (SPEA-monocytes). For the evaluation of SPEA-monocyte-mediated modulation of CD4+ T lymphocytes, SPEA was removed from the culture through extensive washing of cells before adding allogeneic CD3/CD28-activated T cells. Results: In coculture with allogeneic CD4+ T cells, SPEA-monocytes mediate apoptosis of CD4+Foxp3− lymphocytes and accumulation of CD4+Foxp3+ Tregs. PD-L1 and kynurenine are critically involved in the mediated cell death because blocking both factors diminished apoptosis and decreased the proportion of the CD25+/Foxp3+ Treg subpopulation significantly. Upregulation of PD-L1 and kynurenine as well as SPEA-monocyte-mediated effects on T cells depend on inflammatory IL-1β. Our study shows that monocytes activated by SPEA mediate apoptosis of CD4+Foxp3− T effector cells through PD-L1 and kynurenine. CD4+Foxp3+ T cells are resistant to apoptosis and accumulate in SPEA-monocyte/CD4+ T-cell coculture.


Introduction
Streptococcus pyogenes (also known as Group A streptococcus (GAS)) is a Gram-positive coccus and possibly part of the microbiota of our skin and upper respiratory tract. In humans, GAS can PD-L1 binds to PD-1 (programmed cell death protein 1), a negative co-stimulator receptor on activated T cells [41,42]. The PD-l/PD-L1 binding promotes development and function of regulatory T cells (Tregs) by induction and maintenance of the Treg-specific transcription factor forkhead box protein P3 (Foxp3) [40,43]. During primary T-cell activation, PD-l/PD-L1 interaction mediates blockage of T-cell proliferation and cytokine production and inhibits cytotoxic activity and cell survival [44,45]. Additionally, effector T-cell reactivation and function is negatively modulated by the PD-1/PD-L1 interaction [46]. Altogether, PD-L1 fulfills a major role in suppressing the adaptive immune system during infection.
IDO is also known for its role in immune suppression. The enzyme is strongly induced in APCs in response to inflammatory signals, IFNγ, interleukin 1 (IL-1), and IL-6, as well as in response to CTLA-4-mediated signaling, and it depends on signal transducer and activator of transcription 1 (STAT1) and STAT3 transcription factors [47,48]. It catabolizes the degradation of tryptophan (Trp) into derivates such as kynurenine. Although the depletion of Trp from the microenvironment itself is immunosuppressive [49], kynurenine additionally mediates immune modulatory effects. As ligand for the aryl hydrocarbon transcription (AHR) factor complex, it promotes the differentiation of activated CD4 + T cells into Tregs [50][51][52][53][54].
In SAg-mediated modulation of T cells, the role of PD-L1 as well as IDO-generated kynurenine has not yet been clarified.
Our previous published data reveal that SPEA-stimulated monocytes (SPEA-monocytes) inhibit proliferation of CD3/CD28-stimulated allogeneic T lymphocytes. Furthermore SPEA-monocytes promote accumulation of Tregs [39]. In the present study, we investigated the mechanism underlying T-cell inhibition and Treg accumulation mediated by SPEA-monocytes.

Results
To investigate the influence of SPEA on antigen-presenting cells (APCs), blood-derived monocytes from healthy donors were stimulated with SPEA for 48 h (SPEA-monocytes). For the evaluation of SPEA-monocyte-mediated effects on CD4 + T lymphocytes, SPEA was removed from the culture after two days by washing cells three times with media. Afterward, freshly isolated allogeneic CD4 + T cells were added.
First, we evaluated the best SPEA concentration to activate monocytes. Cells were stimulated with 1 ng/mL, 10 ng/mL, 100 ng/mL, or 1000 ng/mL SPEA for 48 h. After washing the monocytes, cells were cocultured with isolated carboxyfluorescein succinimidyl ester (CFSE)-stained and CD3/CD28-activated T cells for five days. CFSE is a cell permeable fluorescent dye covalently binding molecules intracellularly (lysine residues and other amino groups) via its succinimidyl group. During a cell division, molecules and bound CFSE are shared between daughter cells. By determining halving of the CFSE (FITC) signal at a flow cytometer, cell divisions can be analyzed. The data obtained showed that activated T cells cultured with monocytes had a decreased CFSE signal and thus proliferated, as expected. Furthermore, 1 ng/mL and 10 ng/mL SPEA-stimulated monocytes had the same stimulatory effect on T-cell proliferation. However, 100 ng/mL-stimulated APCs suppressed CD3/CD28-stimulated lowering of the CFSE signal. Thus, less divided cells were monitored. Finally, T cells cultured with 1000 ng/mL-stimulated monocytes had a significantly higher CFSE signal than activated T cells. This means that a significantly lower number of divided T cells was determined (Figure 1a). We subsequently used 1000 ng/mL SPEA to stimulate monocytes prior to T-cell coculture.
Next we performed lymphocyte growth kinetics. CFSE-labeled T cells were cocultured with SPEA-monocytes for one, three and five days and analyzed on a flow cytometer. During the first day of coculture, SPEA-monocytes seemed to increase proliferation of activated lymphocytes (Figure 1b). Nevertheless, the associated quantification of three experiments (Figure 1c) yielded no significant difference in T-cell divisions induced by monocytes and SPEA-monocytes after one and three days of culture. After five days, T-cell numbers in SPEA-monocyte coculture were significantly lower than those of unstimulated monocyte/T-cell cocultures (Figure 1b,c). (a) CD14 + cells were isolated from blood, stimulated with 1 ng/mL, 10 ng/mL, 100 ng/mL, or 1000 ng/mL SPEA for 48 h (SPEA-monocytes) or left unstimulated. After washing the monocytes three times with media, cocultures with allogeneic CFSE-labeled, CD3/CD28activated CD4 + T cells were started (ratio of 1:2). After five days, cell divisions of T cells were analyzed by determining the CFSE (FITC) signal using a FACSCanto. Shown is the quantification of geometric mean fluorescence of the CFSE signal (y-axes). The color depicts whether T cells were cultured alone (white) or together with unstimulated monocytes (black) or SPEA-monocytes (grey). (b) Monocytes were stimulated with 1000 ng/mL SPEA for 48 h. After washing, CFSE-T cells were added and cultured for one, three, or five days (as labeled in the graph). Depicted are FACS histograms of the CFSE (FITC) signal of T cells. (c) Quantification of geometric mean fluorescence of the CFSE signal of T cells at the indicated timepoints. The color depicts whether T cells were cultured alone (white) or together with unstimulated monocytes (black) or SPEA-monocytes (grey). (a,c) Columns: mean of three different donors/experiments (n = 3) + standard deviation (SD) as error bars. Statistical analysis: the comparison of two data groups was analyzed by Mann-Whitney U test with * p ≤ 0.05. The line above the columns depicts the compared groups. Kruskal-Wallis: (a) number of groups: 7; p-value 0.0187; sum value, * the medians vary significantly (p < 0.05). (c) number of groups: 7; p-value 0.0021; sum value, ** the medians vary significantly (p < 0.05).
As the analyzed CFSE signal conveys that many divided T cells disappeared after five days of SPEA-monocyte coculture, apoptosis of lymphocytes was analyzed. We performed annexin V staining to detect cell surface changes associated with early apoptosis events. Annexin-positive (apoptotic) T cells in the SPEA-monocyte coculture remarkably increased from three (around 10%) to five (around 50%) days. Significantly fewer (around 10% at day five) lymphocytes cultured with (a) CD14 + cells were isolated from blood, stimulated with 1 ng/mL, 10 ng/mL, 100 ng/mL, or 1000 ng/mL SPEA for 48 h (SPEA-monocytes) or left unstimulated. After washing the monocytes three times with media, cocultures with allogeneic CFSE-labeled, CD3/CD28-activated CD4 + T cells were started (ratio of 1:2). After five days, cell divisions of T cells were analyzed by determining the CFSE (FITC) signal using a FACSCanto. Shown is the quantification of geometric mean fluorescence of the CFSE signal (y-axes). The color depicts whether T cells were cultured alone (white) or together with unstimulated monocytes (black) or SPEA-monocytes (grey). (b) Monocytes were stimulated with 1000 ng/mL SPEA for 48 h. After washing, CFSE-T cells were added and cultured for one, three, or five days (as labeled in the graph). Depicted are FACS histograms of the CFSE (FITC) signal of T cells. (c) Quantification of geometric mean fluorescence of the CFSE signal of T cells at the indicated timepoints. The color depicts whether T cells were cultured alone (white) or together with unstimulated monocytes (black) or SPEA-monocytes (grey). (a,c) Columns: mean of three different donors/experiments (n = 3) + standard deviation (SD) as error bars. Statistical analysis: the comparison of two data groups was analyzed by Mann-Whitney U test with * p ≤ 0.05. The line above the columns depicts the compared groups. Kruskal-Wallis: (a) number of groups: 7; p-value 0.0187; sum value, * the medians vary significantly (p < 0.05). (c) number of groups: 7; p-value 0.0021; sum value, ** the medians vary significantly (p < 0.05).
As the analyzed CFSE signal conveys that many divided T cells disappeared after five days of SPEA-monocyte coculture, apoptosis of lymphocytes was analyzed. We performed annexin V staining to detect cell surface changes associated with early apoptosis events. Annexin-positive (apoptotic) T cells in the SPEA-monocyte coculture remarkably increased from three (around 10%) to five (around 50%) days. Significantly fewer (around 10% at day five) lymphocytes cultured with unstimulated monocytes were apoptotic (Figure 2a,b). To further strengthen that result, we stained the SPEA-monocyte/T-cell coculture after five days with propidiumiodide (PI) in addition to annexin V. PI is a fluorescent membrane impermeable DNA intercalating agent that stains only necrotic, dead cells. Analysis of the gated double-stained lymphocytes confirmed the increase of apoptotic T cells during coculture. From day 3 to day 5 annexin-positive T cells increased PE (PI) signal and therefore became necrotic (Figure 2c). unstimulated monocytes were apoptotic (Figure 2a,b). To further strengthen that result, we stained the SPEA-monocyte/T-cell coculture after five days with propidiumiodide (PI) in addition to annexin V. PI is a fluorescent membrane impermeable DNA intercalating agent that stains only necrotic, dead cells. Analysis of the gated double-stained lymphocytes confirmed the increase of apoptotic T cells during coculture. From day 3 to day 5 annexin-positive T cells increased PE (PI) signal and therefore became necrotic (Figure 2c). This raised the question whether the previously observed accumulation of Tregs in SPEAmonocyte coculture [39] could be due to a higher resistance of Tregs toward apoptosis. At first, we confirmed the accumulation of Tregs in cocultures of SPEA-monocytes/CD4 + T cells (Figure 3a). After five days, no difference in Treg numbers of cultured T cells and monocyte cocultured T cells was found, as expected. However, in T-cell/SPEA-monocyte coculture a significant increase in Tregs could be observed (Figure 3a).
Next, cells of the SPEA-monocyte/T-cell coculture were stained with annexin, anti-Foxp3 antibody, and anti-CD4 antibody. The staining with anti-Foxp3 antibody enabled discrimination This raised the question whether the previously observed accumulation of Tregs in SPEA-monocyte coculture [39] could be due to a higher resistance of Tregs toward apoptosis. At first, we confirmed the accumulation of Tregs in cocultures of SPEA-monocytes/CD4 + T cells (Figure 3a). After five days, no difference in Treg numbers of cultured T cells and monocyte cocultured T cells was found, as expected. However, in T-cell/SPEA-monocyte coculture a significant increase in Tregs could be observed ( Figure 3a).
Next, cells of the SPEA-monocyte/T-cell coculture were stained with annexin, anti-Foxp3 antibody, and anti-CD4 antibody. The staining with anti-Foxp3 antibody enabled discrimination between Tregs and non-Tregs at the flow cytometer. The data again confirmed the accumulation of  Figure 3c). After three days, around 40% of Foxp3-negative CD4 + T cells bound annexin, and two days later around 70% were annexin-positive (Figure 3 b,c). In contrast to Foxp3 -cells, very few Foxp3 + T cells bound annexin (day 3 around 2%, day 5 around 7 %, Figure 3 b,c). Next, we aimed to clarify the mechanisms responsible for apoptosis in Foxp3 − T cells. From our former studies, it is known that SPEA-monocytes release high amounts of pro-inflammatory cytokines, upregulate T-cell co-inhibitory PD-L1 on their surface, and generate immunosuppressive kynurenine [39]. Immunosuppressive myeloid cells are known to inhibit T effector cells via PD-L1 Next, we aimed to clarify the mechanisms responsible for apoptosis in Foxp3 − T cells. From our former studies, it is known that SPEA-monocytes release high amounts of pro-inflammatory cytokines, upregulate T-cell co-inhibitory PD-L1 on their surface, and generate immunosuppressive kynurenine [39]. Immunosuppressive myeloid cells are known to inhibit T effector cells via PD-L1 and kynurenine [54]. After stimulation with bacterial components, TLR-signaling-induced IL-1β is responsible for the induction of both factors and the reprogramming of APCs toward an immunosuppressive phenotype that inhibits T-cell activation [54]. To test whether IL-1β could also play a role in the SPEA-monocyte-mediated effects, we checked the concentration of IL-1β after SPEA stimulation. We observed that 1 ng/mL and 10 ng/mL SPEA did not induce a noteworthy release of IL-1β. However, 100 ng/mL SPEA induced a significant increase in cytokine expression, and 1000 ng/mL SPEA-stimulated monocytes released the highest amount of IL-1β and significant more cytokine than unstimulated and 100 ng/mL SPEA-stimulated cells (Figure 4a). Then, we blocked IL-1β-signaling. During SPEA stimulation, monocytes were treated with anti-IL-1β neutralizing antibody. After two days, CFSE-labeled CD4 + T cells were added to washed monocytes for five days. To exclude any unspecific effect of the antibody, we included the anti-IL1β isotype antibody in the experiments. T cells cultured for five days with untreated monocytes proliferated, as expected. In SPEA-monocyte coculture, very few divided lymphocytes were detected. Treatment with the isotype antibody had no effect on T-cell proliferation. However, blocking IL-1β-signaling decreased the CFSE signal of SPEA-monocyte cocultured T cells significantly (Figure 4b,c). This result implies an impact of IL-1β-signaling in the SPEA-monocyte-mediated effect on T-cell activation. and kynurenine [54]. After stimulation with bacterial components, TLR-signaling-induced IL-1β is responsible for the induction of both factors and the reprogramming of APCs toward an immunosuppressive phenotype that inhibits T-cell activation [54]. To test whether IL-1β could also play a role in the SPEA-monocyte-mediated effects, we checked the concentration of IL-1β after SPEA stimulation. We observed that 1 ng/mL and 10 ng/mL SPEA did not induce a noteworthy release of IL-1β. However, 100 ng/mL SPEA induced a significant increase in cytokine expression, and 1000 ng/mL SPEA-stimulated monocytes released the highest amount of IL-1β and significant more cytokine than unstimulated and 100 ng/mL SPEA-stimulated cells (Figure 4a). Then, we blocked IL-1β-signaling. During SPEA stimulation, monocytes were treated with anti-IL-1β neutralizing antibody. After two days, CFSE-labeled CD4 + T cells were added to washed monocytes for five days.
To exclude any unspecific effect of the antibody, we included the anti-IL1β isotype antibody in the experiments. T cells cultured for five days with untreated monocytes proliferated, as expected. In SPEA-monocyte coculture, very few divided lymphocytes were detected. Treatment with the isotype antibody had no effect on T-cell proliferation. However, blocking IL-1β-signaling decreased the CFSE signal of SPEA-monocyte cocultured T cells significantly (Figure 4b,c). This result implies an impact of IL-1β-signaling in the SPEA-monocyte-mediated effect on T-cell activation.  Then, we investigated the link between SPEA-stimulated IL-1β and SPEA-monocyte-mediated effects on T cells. We evaluated how blocking IL-1β during SPEA-stimulation modulates the previously demonstrated SPEA-monocyte phenotype, including upregulation of costimulatory CD80, CD86, inhibitory PD-L1, inflammatory TNFα, IL-6 and anti-inflammatory IL-10, and kynurenine [39]. Flow cytometry data (Figure 5a,b) confirmed that 1000 ng/mL SPEA stimulated upregulation of CD80, CD86, HLA-DR (human MHC class II), and PD-L1. SPEA-stimulated induction of costimulatory CD80, CD86, and HLA-Dr surface expression was not significantly modulated after anti-IL-1β antibody treatment. However, PD-L1 expression was significantly repressed by neutralizing IL-1β (Figure 5a,b).
Additional ELISA analysis confirmed that 1000 ng/mL SPEA induced the production of TNFα, IL-6, and IL-10 ( Figure 5c). TNFα and IL-10 release was not changed significantly after blocking the IL-1 receptor. However, IL-6 production was reduced significantly. Furthermore, the generation of SPEA-induced kynurenine, the product of the IDO-catalyzed degradation of tryptophan, was potently and significantly inhibited through IL-1β-neutralization ( Figure 5c). As expected from the decreased kynurenine generation, SPEA-stimulated IDO expression, shown by Western blot analysis after 48 h of stimulation, was also decreased in monocytes after treatment with anti-IL1β antibody (Figure 5d).
Additional ELISA analysis confirmed that 1000 ng/mL SPEA induced the production of TNFα, IL-6, and IL-10 ( Figure 5c). TNFα and IL-10 release was not changed significantly after blocking the IL-1 receptor. However, IL-6 production was reduced significantly. Furthermore, the generation of SPEA-induced kynurenine, the product of the IDO-catalyzed degradation of tryptophan, was potently and significantly inhibited through IL-1β-neutralization ( Figure 5c). As expected from the decreased kynurenine generation, SPEA-stimulated IDO expression, shown by Western blot analysis after 48 h of stimulation, was also decreased in monocytes after treatment with anti-IL1β antibody (Figure 5d).  These results educed the hypothesis that neutralization of IL-1β in SPEA-monocytes reversed Tcell inhibition through modulation of monocyte phenotype and abrogation of PD-L1 and IDOgenerated kynurenine. To review this hypothesis, we first investigated whether PD-L1 surface expression and kynurenine concentration in the cell supernatant were stable during the coculture of monocytes and T cells. We stimulated monocytes with SPEA and analyzed PD-L1 surface expression via flow cytometry (Figure 6a) and kynurenine concentration (Figure 6b) after one and two days. Then, cells were washed and T cells were added as described before. Three and five days later, PD-L1 expression was analyzed on a flow cytometer in the gated monocyte population. Kynurenine concentration in the coculture was determined. PD-L1 expression (Figure 6a) remained stable, and kynurenine was still produced after the start of coculture (Figure 6b). Then, we blocked the generation of kynurenine via a competitive IDO inhibitor (1MT) prior to coculture and prevented These results educed the hypothesis that neutralization of IL-1β in SPEA-monocytes reversed T-cell inhibition through modulation of monocyte phenotype and abrogation of PD-L1 and IDO-generated kynurenine. To review this hypothesis, we first investigated whether PD-L1 surface expression and kynurenine concentration in the cell supernatant were stable during the coculture of monocytes and T cells. We stimulated monocytes with SPEA and analyzed PD-L1 surface expression via flow cytometry ( Figure 6a) and kynurenine concentration (Figure 6b) after one and two days. Then, cells were washed and T cells were added as described before. Three and five days later, PD-L1 expression was analyzed on a flow cytometer in the gated monocyte population. Kynurenine concentration in the coculture was determined. PD-L1 expression (Figure 6a) remained stable, and kynurenine was still produced after the start of coculture (Figure 6b). Then, we blocked the generation of kynurenine via a competitive IDO inhibitor (1MT) prior to coculture and prevented PD-1/PD-1 binding in SPEA-monocyte/CD4 + T-cell coculture by an anti-PD-L1 antibody. Afterward, apoptosis of T cells was determined by analyzing annexin staining. SPEA-monocytes mediated apoptosis of lymphocytes, as expected (Figure 6c). Blocking PD-1 signaling in T cells induced a significant decrease in annexin binding and thereby an increase in survival of SPEA-monocyte cocultured T cells. Inhibition of kynurenine generation (by 1MT) resulted in a similar decrease of apoptosis. Treatment with 1MT and anti-PD-L1 did not further increase survival significantly.  Finally, the accumulation of Tregs was analyzed after five days of coculture. SPEA-monocytes mediated the accumulation of Tregs (38-55%), as expected (Figure 7). Administration of anti-PD-L1 antibody diminished the number of CD25/Foxp3 + Tregs significantly (14-28%), just as inhibition of kynurenine generation via 1MT (12-24%). Consequently, treatment with the anti-IL-1β antibody during monocyte stimulation that suppressed expression of PD-L1, expression of IDO, and generation of kynurenine also potently inhibited Treg accumulation (Figure 7). Finally, the accumulation of Tregs was analyzed after five days of coculture. SPEA-monocytes mediated the accumulation of Tregs (38-55%), as expected (Figure 7). Administration of anti-PD-L1 antibody diminished the number of CD25/Foxp3 + Tregs significantly (14-28%), just as inhibition of kynurenine generation via 1MT (12-24%). Consequently, treatment with the anti-IL-1β antibody during monocyte stimulation that suppressed expression of PD-L1, expression of IDO, and generation of kynurenine also potently inhibited Treg accumulation (Figure 7).

Discussion
A direct and simultaneous binding of SAg with T-cell receptors (TCRs) and MHC class II on APCs is the current accepted paradigm for T-cell activation by SAgs. In this scenario, SAgs must reach a site in the body with both cell types. During infection, the draining lymph node is the most likely site for SAg to interact with V beta T cells as well as APCs. The idea of SAgs binding T cells in peripheral tissues or in the blood is rather unlikely due to the limited numbers of T cells. More likely, SAgs can reach the lymph node near the infection as soluble antigen in lymphatic fluid or presented on APCs that previously inserted the SAg in the tissue at the site of infection [55]. APCs are present in all tissues and migrate after activation through an antigen to the next lymph node to encounter circulating T cells. Therefore, it seems feasible that SAgs in tissue or blood activate APCs to mediate inflammation, to migrate to the nearby lymph node and activate T cells without the need of Sag-TCR binding. Several studies show that SAgs activate APCs, independently of T cells [31][32][33]39]. SAg binding and crosslinking of MHC class II delivers a signal to the APC that accounts for the intracellular increase of calcium and eventually leads to the upregulation of pro-inflammatory cytokines. Thus, Hopkins et al. showed that in the human system SAg-induced ligation of MHC class II on monocytes upregulates TLR4 and enhanced pro-inflammatory responses to TLR ligands, independently of T cells and INFγ [56]. Other studies reveal that SAg-stimulated APCs release proinflammatory cytokines and upregulate costimulatory surface molecules in the absence of a TLR ligand [31][32][33][34]57]. Previously, we observed an induction of the immunosuppressive factors PD-L1 and kynurenine after SPEA stimulation of monocytes [39]. This finding let us hypothesize that activating as well as inactivating effects on T cells could be mediated by SAg-stimulated APCs.
In the present study, we showed that SPEA-monocytes induce apoptosis of Foxp3-negative lymphocytes and accumulation of Tregs through PD-L1 and kynurenine. In our approach, we washed SPEA-stimulated monocytes several times before adding allogeneic CD4 + T cells. We cannot exclude the possibility that potentially ingested SPEA was presented to T cells and that a direct binding of monocytes and T cells occurred in coculture. However, stimulation of monocytes with

Discussion
A direct and simultaneous binding of SAg with T-cell receptors (TCRs) and MHC class II on APCs is the current accepted paradigm for T-cell activation by SAgs. In this scenario, SAgs must reach a site in the body with both cell types. During infection, the draining lymph node is the most likely site for SAg to interact with Vβ T cells as well as APCs. The idea of SAgs binding T cells in peripheral tissues or in the blood is rather unlikely due to the limited numbers of T cells. More likely, SAgs can reach the lymph node near the infection as soluble antigen in lymphatic fluid or presented on APCs that previously inserted the SAg in the tissue at the site of infection [55]. APCs are present in all tissues and migrate after activation through an antigen to the next lymph node to encounter circulating T cells. Therefore, it seems feasible that SAgs in tissue or blood activate APCs to mediate inflammation, to migrate to the nearby lymph node and activate T cells without the need of Sag-TCR binding. Several studies show that SAgs activate APCs, independently of T cells [31][32][33]39]. SAg binding and crosslinking of MHC class II delivers a signal to the APC that accounts for the intracellular increase of calcium and eventually leads to the upregulation of pro-inflammatory cytokines. Thus, Hopkins et al. showed that in the human system SAg-induced ligation of MHC class II on monocytes upregulates TLR4 and enhanced pro-inflammatory responses to TLR ligands, independently of T cells and INFγ [56]. Other studies reveal that SAg-stimulated APCs release pro-inflammatory cytokines and upregulate costimulatory surface molecules in the absence of a TLR ligand [31][32][33][34]57]. Previously, we observed an induction of the immunosuppressive factors PD-L1 and kynurenine after SPEA stimulation of monocytes [39]. This finding let us hypothesize that activating as well as inactivating effects on T cells could be mediated by SAg-stimulated APCs.
In the present study, we showed that SPEA-monocytes induce apoptosis of Foxp3-negative lymphocytes and accumulation of Tregs through PD-L1 and kynurenine. In our approach, we washed SPEA-stimulated monocytes several times before adding allogeneic CD4 + T cells. We cannot exclude the possibility that potentially ingested SPEA was presented to T cells and that a direct binding of monocytes and T cells occurred in coculture. However, stimulation of monocytes with SPEA highly upregulated PD-L1 and kynurenine. Inhibition of both factors diminished the mediated apoptosis of cocultured T cells significantly.
Recently, we showed that TLR-activated immunogenic APCs reprogram themselves toward an immunosuppressive phenotype with high surface expression of PD-L1 and high release of kynurenine. As a trigger for this remodeling, we identified IL-1β that augments IL-6 production and a STAT3-dependent induction of immunosuppressive factors [54]. SPEA-stimulated monocytes seem to undergo a similar remodeling, as neutralization of the cytokine during SPEA stimulation diminished PD-L1 surface expression and kynurenine release. Additionally, preventing PD-L1 binding to PD-1 on T cells as well as suppression of Trp degradation into kynurenine diminished SPEA-monocyte-mediated apoptosis of effector T cells and decreased the shift of T-cell populations toward Tregs.
It is well known that the simultaneous binding of SAgs of MCH class II and TCR results in apoptosis of roughly 50% of the initial numbers of Vβ-bearing T cells after the initial activation of a phase of clonal T-cell expansion [21][22][23]58]. The surviving cells show an anergic phenotype [24,59]. Additionally, the proportion of CD4 + CD25 + Foxp3 + Tregs is strongly augmented [25][26][27]. The underlying mechanism of T-cell apoptosis is not entirely clarified. Several reports have suggested that Fas governs T-cell apoptosis after repeated administration of SAg [60][61][62]. Several other studies have shown that after a single dose of Sag, T cells died in the absence of Fas or Fas ligand [63][64][65][66][67]. The second proposed SAg-mediated way of T-cell death is the regulation of Bcl-2 family members [64]. Bcl-2 is the prototype member of a large family of related proteins having pro-or anti-apoptotic function by regulation of the mitochondria pathway of apoptosis [68,69]. At the end of a T-cell response, before cells die in vivo, the majority of the activated T cells exhibit decreased levels of pro-survival Bcl-2 [70]. Interestingly, SEB-induced apoptosis of T cells can be prevented by retroviral restoration of Bcl-2 [64].
According to the data of the present study, PD-L1 and kynurenine are critically involved in SPEA-monocyte-mediated T-cell apoptosis. From the literature, it is known that PD-L1/PD-1 binding induces signaling events that reduce effector T-cell survival. Inhibiting PI3K activation through the recruitment of phosphatases PD-1 suppresses CD28-mediated induction of the pro-survival Bcl-2 family member Bcl-xL [71]. Additionally, a publication on HIV infection illustrates a kynurenine-dependent mechanism through IL-2 signaling for reduced CD4 + T-cell survival that involves reactive oxygen species [72].
In summary, we show that SPEA-stimulated monocytes induce apoptosis of Foxp3 − lymphocytes through PD-L1 and kynurenine. Foxp3 + lymphocytes are resistant to apoptosis and accumulate. The question of whether the mechanism presented here applies to other SAgs, such as the enterotoxins of S. aureus, cannot be answered by our work and should be clarified in future studies.
Statistical analysis: The comparison of two data groups was analyzed by Mann-Whitney U test (one-tailed, confidence intervals 95%) with * p ≤ 0.05. Additionally, Kruskal-Wallis Test (one-way ANOVA on ranks) was performed. Software: GraphPad Prism Version 5.0 (GraphPad Software Inc., San Diego, CA, USA).
Ethical statement: This study (taking of blood samples from healthy donors and treatment of blood leukocytes with microbial stimuli) was carried out in accordance with the recommendations of the ethics committee of the Medizinische Fakultät Heidelberg with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The animal experiments were approved by the governmental animal ethics committee (Regierungspraesidium Karlsruhe, file number: 35-9185.81/G-132/15) and conducted according to international FELASA recommendations. The study was reviewed and approved by the ethics committee of Medizinische Fakultät Heidelberg.