Piperlongumine (#SML0221), DMSO, NAC, Paraformaldehyde (PFA), phorbol 12-myristate 13-acetate (PMA), Ionomycin and Brefeldin A were purchased from Sigma-Aldrich (St. Louis, USA). GSH, 5-(and-6) chloromethyl-2′, 7′-dichlorodihydro-fluorescein diacetate, acetyl ester (CM-H₂DCFDA), RPMI 1640 and ThiolTracker^TM Violet were purchased from Thermo Fisher Scientific. Carboxyfluorescein diacetate succinimidyl ester (CFSE) was bought from Invitrogen (Eugene, USA) and fetal bovine serum (FBS) from PAN-Biotech (Aidenbach, Germany). nCounter® GX Human Immunology v2 panel was purchased from NanoString Technologies and Direct-zol^TM RNA MiniPrep kit from QIAGEN. Ficoll-Hypaque (FicoLite H) was obtained from Linarisblue (Wertheim-Bettingen, Germany) and BD FACS^TM lysing solution from BD Bioscience. Recombinant human IL-12, IL-4, IL-1β, IL-23, IL-6, TGF-β, neutralizing antibodies against IFNγ and IL-4, True-Nuclear™ transcription factor buffer set, LEGEND MAX™ Human IL-17A, IL-17F and LEGEND MAX™ Free Active TGF-β ELISA kit were obtained from BioLegend (San Diego, USA).

Antibodies used in this study were specific for the following molecules: CD3 (clone OKT3, mouse mAb), CD28 (clone 28.2, mouse mAb) and isotype control antibodies IgG1, κ and IgG2a, κ (mouse mAb, BD Biosciences, Heidelberg, Germany), GAPDH (clone 6C5, mouse mAb, Invitrogen; Eugene, USA), actin (rabbit polyclonal, Sigma-Aldrich, Hamburg, Germany), GSH (clone D8, mouse mAb, Santa Cruz Biotechnology), HIF-1α (rabbit polyclonal, Cayman chemical, USA). For the secondary antibodies, IRDye® 680CW donkey anti-mouse and IRDye® 800CW donkey anti-rabbit were purchased from LICOR Biosciences (Lincoln, USA). Goat anti-mouse IgG+IgM used as coating antibody was from Jackson ImmunoResearch (West Grove, USA). 7-AAD and all fluorescently-labeled antibodies were obtained from BD Biosciences.

Human peripheral blood mononuclear cells (PBMCs) were obtained from heparinized blood of voluntary healthy donors by density-gradient centrifugation using Ficoll-Hypaque. Pan T cells or naïve CD4⁺ T cells were isolated by negative selection using the pan T cell isolation kit or naïve CD4⁺ T cells isolation kit from Miltenyi Biotec (Bergisch Gladbach, Germany), respectively, according to the manufacturer's instructions. The purity of Pan T cells reached more than 99.5% (Supplementary Figure 1a), of naïve CD4⁺ T cells 99% (Supplementary Figure 1b). The purified T cells were adjusted to 3 x 10⁶ cells/ml in RPMI complete medium (RPMI 1640+10% FBS) and cultured in incubator at 5% CO₂ and 37°C until use. To costimulate human peripheral blood T cells (PBTs), goat anti-mouse IgG+IgM antibody was used to pre-coat microplates (Nunc, Wiesbaden, Germany), followed by blocking with RPMI complete medium and coating with anti-CD3 (20 ng/mL)/anti-CD28 (5 μg/mL) antibodies or the respective isotype controls. T cells were spun down on the antibodies and incubated at 5% CO₂ and 37°C for the indicated time points. This study was approved by the Ethics Committee of the Heidelberg University (S-269/2015). The human T cell leukemia cell line Jurkat ACC282 and Raji cells, which served as antigen-presenting cells (APC) were cultured in RPMI complete medium at 5% CO₂ and 37°C, and cells were split 1:5 every second day by replacing the medium.

Peripheral blood from patients with different diseases was collected into sterile heparinized tubes under aseptic conditions. Informed consent for use of the cells was obtained from all patients included in this study. To test the effects of PL on the lymphocytes in whole blood, 100 μl of blood were diluted 1:1 with RPMI medium, blood was treated with DMSO or with 5 μM PL and activated with 1 μg/ml staphylococcal enterotoxin B (SEB) for 20 hours (h). Afterwards, Brefeldin A (10 μg/ml) was used to prevent the secretion of cytokines (29). After 4 h incubation, red blood cells were lysed with BD FACS Lysis Solution for 10 min at room temperature (RT) and simultaneously, white blood cells were fixed. Samples were then washed with FACS wash buffer (FW, PBS containing 0.5% BSA, 0.5% FBS, and 0.07% NaN₃), permeabilized with FWS (FW containing 0.1% Saponin) and stained with fluorescent-labeled antibodies against CD3, CD69, CD25, IFNγ, and IL-2 in FWS for 20 min at RT. Thereafter, samples were washed and assessed by flow cytometry (LSRII, BD Bioscience, Heidelberg, Germany). Data were analyzed with FlowJo X (FlowJo LLC, Ashland, OR, USA). This study was approved by the Ethics Committee of the Heidelberg University (S-119/2017).

Primary human T cells and Jurkat cells were cultured in 200 μL RPMI complete medium with or without 5 μM PL for the indicated time points (5% CO₂ and 37°C). PL was solved in DMSO, therefore cells were treated with equal concentrations of DMSO as a control. The cells were then washed once with PBS, stained with AnnexinV and 7-AAD in Annexin binding buffer for 20 min at RT. Thereafter, cells were washed once with Annexin binding buffer and analyzed by flow cytometry.

ROS detection reagent CM-H₂DCFDA was used for the measurement of intracellular ROS levels and ThiolTracker^TM violet dye was used for the measurement of intracellular GSH levels as described before (30). Briefly, for the intracellular ROS measurement, T cells (1 × 10⁶ cells/ml) or Jurkat cells (3 × 10⁵ cells/ml) were washed with PBS and stained with CM-H₂DCFDA (5 μM) in PBS for 15 min at 37°C. Cells were then washed with PBS, resuspended in RPMI complete medium and treated with DMSO, PL or H₂O₂. After the indicated time points, the mean fluorescence intensity (MFI) of the CM-H₂DCFDA signal was determined by flow cytometry. For the intracellular GSH measurement, T cells (1 × 10⁶ cells/ml) or Jurkat cells (3 × 10⁵ cells/ml) were incubated with DMSO or PL in RPMI complete medium for 1 h or 24 h. Afterwards, cells were rinsed with PBS and stained with 1 μM ThiolTracker^TM violet dye (diluted in PBS) for 15 min at 37°C. After washing with PBS, cells were resuspended in PBS and measured immediately by flow cytometry.

Western blotting was used for the detection of S-glutathionylated proteins. Control or PL-treated T cells or Jurkat cells were washed with PBS and postnuclear lysates (cytoplasm) were generated. Therefore, T cells were lysed with TKM lysis buffer on ice for 20 min, the nuclei and the cell debris were removed by centrifugation at 14,000 g for 10 min. Proteins in the cytoplasmic fraction were separated by electrophoresis in denaturing SDS polyacrylamide gels (SDS-PAGE), proteins were transferred onto PVDF-membranes and the membranes were blocked in blocking buffer for 1 h. Afterwards, membranes were probed with primary antibodies against GSH (1:200), Actin (1:1,000), GAPDH (1:10,000), and respective IRDye fluorescent-labeled secondary antibodies (1:10,000). The membranes were scanned by a Licor infrared scanner (LI-COR Biosciences, Germany).

The S-glutathionylated protein was further analyzed by mass spectrometry. To this end, postnuclear cell lysates from control or PL-treated T cells (3 × 10⁶ cells/sample), were run on SDS-PAGE and subjected to Coomassie-staining. The bands corresponding to the 42-kDa protein were cut out and analyzed by mass spectrometry. Protein digestion and LC-MS measurement were done as described elsewhere (31). Briefly, gel pieces were washed, dehydrated and incubated with trypsin solution (Thermo-Fisher, Rockford, USA) for 4 h at 37°C. The reaction was quenched by adding of 20 μL of 0.1% trifluoroacetic acid (TFA; Biosolve, Netherlands). The supernatant was dried in a vacuum concentrator before LC-MS analysis. Nanoflow LC-MS analysis was performed with a NanoAcquity UPLC liquid chromatography (Waters, Eschborn, Germany) system coupled to an Orbitrap XL or with an Ultimate 3000 liquid chromatography system coupled to an Orbitrap Q Exactive mass spectrometer (Thermo-Fischer, Bremen, Germany). Detailed information for the LC-MS measurements is described in the Supplementary Material. For the protein identification, raw files were analyzed using Proteome Discoverer with the Sequest version 2.3 (Thermo Fisher Scientific, USA). Sequest was set up to search against Uniprot human databases (retrieved in June, 2017) with trypsin as the digestion enzyme with maximum two missed cleavages. A parent ion mass tolerance was set to 10 ppm and a fragment ion mass tolerance was set to 0.50 Da or 0.02 Da for measurements on the Orbitrap XL or the Q Exactive mass spectrometer, respectively. Proteome Discoverer results were used for a spectral library generation in the Skyline (32) using BiblioSpec algorithm for MS¹ full-scan quantification of glutathionylated peptides and their unmodified counterparts as previously described (33). After raw file data import into Skyline, chromatographic traces for top three isotopic peaks were manually inspected for proper peak picking of MS¹ filtered peptides and peak area determination. The summed area of three isotopic peaks was used to calculate peptide intensities. Target peptide areas were normalized to summed area of selected actin peptides (AVFPSIVGRPR; RGILTLK; VAPEEHPVLLTEAPLNPK; SYELPDGQVITIGNER).

Control or PL-treated T cells (3.5 × 10⁶ cells/sample, 1 × 10⁶ cells/ml) were centrifuged and re-suspended in 250 μL RPMI medium (+2% FBS) with the calcium (Ca²⁺) -binding dye, Indo-1, at a concentration of 2 μg/ml. Cells were then stained at 37°C for 45 min. The dye changes its emission wavelength when binding to Ca²⁺, so a ratio between bound and unbound dye can be calculated. The samples were washed twice with PBS, then resuspended in 500 μL RPMI complete medium. The kinetics of the Ca²⁺ signal was measured by flow cytometry (Fortessa, BD Bioscience, Heidelberg, Germany). To this end, the baseline Ca²⁺ signal was acquired for 1 min, followed by adding anti-CD3 antibody to a final concentration of 20 ng/ml. The Ca²⁺ signal was acquired for another 8 min, until goat-anti-mouse antibody was added (7.2 μg/mL) for crosslinking of the anti-CD3 antibody to induce activation and Ca²⁺ influx to the cytosol. Thereafter, the Ca²⁺ signal was monitored for up to 18 min and then ionomycin was added to induce a maximum Ca²⁺ signal as a positive control and the measurement was stopped after 20 min.

Conjugates were formed between T cells and Raji cells as described below. T cells (1 × 10⁶ cells/ml) were preincubated with or without 5 μM PL for 1 h or overnight and Raji cells were loaded with 5 μg/ml SEB or kept unloaded. Pretreated T cells and Raji cells were then coincubated for 45 min at 37°C at a ratio of 1:1. Afterwards, 1.5% PFA was added while vortexing the sample for 10 s in order to separate non-specifically bound cells. After 10 min fixation, cells were washed twice with FW buffer followed by staining with anti-CD19 PE-Cy5 and anti-CD3 PE conjugated antibodies. 1 × 10⁴ T cells per sample of three independent experiments were acquired to quantify the proportion of T cell/APC conjugates. CD3 and CD19 double positive events were counted as cell couples.

Immune synapse formation was assessed by multispectral imaging flow cytometry (MIFC, IS100, Amnis Corp., Seattle, WA, USA). To this end, T cell/APC conjugates were formed and fixed as described above. The samples were then stained with anti-CD3 PE-Cy7 and anti-CD18 PE. After permeabilization with FWS buffer, samples were further stained with SiR-actin and DAPI. Thereafter, cells were subjected to MIFC and as many as 15,000 images were acquired per sample. The IDEAS 6.0 software (Amnis, Seattle, WA, USA) was used to analyze the subcellular localization of proteins (34).

T cells (1 × 10⁶ cells/ml) were preincubated with DMSO or different concentrations of PL for 30 min and were seeded in 96-well microplates coated with anti-CD3/CD28 antibodies. In parallel, mouse anti-human IgG1, κ, and IgG2a, κ were coated as isotype control. Then the cells were incubated at 37°C for 24 h and the activation state of the T cells was determined by measuring the expression of CD69 and CD25. Briefly, T cells were collected after plate stimulation and washed twice with FW buffer, fixed with 1.5% PFA for 10 min and stained at RT with anti-CD69 PE-Cy7 and anti-CD25 APC antibodies in FW buffer. After 20 min of staining, cells were washed with FW and measured by flow cytometry. For the detection of the total extra- and intracellular CD69 and CD25, T cells were permeabilized by FWS buffer prior to staining with antibodies in FWS buffer.

T cells were washed with PBS, resuspended in PBS containing 1 μM CFSE and incubated at 37°C for 15 min. The unbound CFSE was then removed by washing twice with PBS. T cells were then resuspended in RPMI complete medium and treated with DMSO or PL for 30 min. Afterwards, the cells were seeded on a 96-well microplate coated with anti-CD3/CD28 antibodies and incubated for 72 h. The CFSE signal was determined using flow cytometry and data were analyzed with FlowJo X. The proliferation index was calculated as described previously (30).

For the assessment of costimulation induced intracellular cytokines, Pan T cells were treated with various concentrations of PL in the presence and absence of exogenously added NAC and then costimulated for 48 h. Afterwards, T cells were restimulated with 10 nM PMA and 1 μg/ml ionomycin for 4 h to enhance the cytokine production. Simultaneously, 10 μg/ml Brefeldin A was used to prevent the secretion of cytokines. Then, T cells were collected and washed twice with FW, fixed in 1.5% PFA for 10 min and permeabilized with FWS for 20 min. Afterwards, T cells were stained with anti-IL-2, anti-IFNγ and anti-TNFα (50 μL/sample) antibodies diluted by FWS for 30 min. Florescence Minus One (FMO) staining was included in control samples to detect unspecific staining. Cells were then washed twice with FWS and measured by flow cytometry. Data were analyzed with FlowJo X.

T cells (1 × 10⁶ cells/sample) were pretreated with DMSO or 5 μM PL for 30 min and then stimulated for 4 h in anti-CD3/CD28 antibodies coated plates. Total RNA was extracted from T cells by trizol. The RNA yield from each sample was detected using a NanoDrop^TM 2000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE), RNA integrity was confirmed using a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA). The gene expression profiling was performed with 25 ng total RNA per sample and all RNA samples were analyzed using the Nanostring nCounter GX Human Immunology v2 kit. The gene expression profiling and data analyses were performed as previously described (30). Briefly, the hybridization reaction was carried out at 65°C overnight by mixing RNA samples with nCounter reporter probes and capture probes in hybridization buffer. After hybridization of the probes with the targets of interest, samples were purified and immobilized on a cartridge and data assessed on the nCounter SPRINT Profiler. The data were then collected with an automated fluorescence microscope and nCounter digital analyzer, and subsequently assessed with the help of the nSolver Analysis Software Version 4.0 (NanoString Technologies). For the data analysis, sample counts were initially normalized by scaling all the values by the ratios of geometric mean (GM) of sample controls to the overall GM of control gene counts across all samples. This was done for both spike-in positive controls and housekeeping genes. The final counts of the mRNA transcripts were used for the further evaluation.

Naïve CD4⁺ T cells were isolated as described above. Cells (2 × 10⁵ cells/ml) were treated with DMSO or 1 μM PL and seeded in a 12 well plate coated with anti-CD3/CD28 antibodies. Naïve CD4⁺ T cells differentiation was induced by adding the following cytokines and antibodies. T_H1: IL-12 (2 ng/ml), anti-IL-4 antibody (2 μg/ml); T_H2: IL-4 (2.5 ng/ml), anti-IFNγ antibody (2 μg/ml); T_H17: IL-6 (10 ng/ml), TGF-β (2.5 ng/ml), IL-1β (10 ng/ml), IL-23 (20 ng/ml), anti-IFNγ (2 μg/ml), and anti-IL-4 antibody (2 μg/ml); T_reg: IL-2 (20 U/ml), TGF-β (5 ng/ml), anti-IFNγ (2 μg/ml) and anti-IL-4 antibody (2 μg/ml). After stimulation and differentiation for 6 days, cells were collected and washed twice with FW buffer, then stained with antibodies specific for the surface molecules CXCR3 (CD183), CCR4 (CD194), and CCR6 (CD196) as mentioned above. The T_H subtypes were identified according to the surface staining as T_H1 cells are CXCR3⁺CCR4⁻CCR6⁻, T_H2 cells are CXCR3⁻CCR4⁺CCR6⁻ and T_H17 cells are CXCR3⁻CCR4⁺CCR6⁺ according to standard flow cytometry procedures (35). The T_reg cell population was determined by the expression of transcription factor FOXP3. For the FOXP3 staining, cells (1 × 10⁶ cells/sample) were fixed with 100 μL 1 × True-Nuclear™ fixation buffer for 1 h, followed by three times washing with 200 μL 1 × True-Nuclear™ Perm buffer. Afterwards, cells were resuspended in 50 μL 1 × True-Nuclear™ Perm buffer with anti-FOXP3 (1:20) antibody. After incubation for 30 min at RT, cells were washed three times with 200 μL of 1 × True-Nuclear™ Perm buffer, resuspended in 200 μL PBS and measured with flow cytometry.

For the measurement of cytokine production after 6 days of T cell differentiation, cells were centrifuged, cell free supernatants were collected and immediately aliquoted and stored at −80°C. Secreted cytokines in undiluted cell culture supernatants were detected by ELISA kits specific for IL-17A, IL-17F, and TGF-β as per the manufacturer's instructions.

Western blot was used to detect HIF-1α. Briefly, T cells (1 × 10⁶ cells/ml) were treated with DMSO or PL as indicated, and activated in anti-CD3/CD28 antibodies coated plates for 24 h. Then, the cells were collected, washed with PBS and total lysate was prepared using PBS with 1 × reducing sample buffer. Proteins were detected by western blotting as described above with the following antibodies: HIF-1α (1:200), Actin (1:1,000) and respective IRDye® conjugated secondary antibodies (1:10,000).

Statistical tests were performed with GraphPad Prism 6.0 (GraphPad; San Diego, USA). Two data sets were analyzed using t-test or paired t-test for matched observations. Multiple groups were compared using ANOVA. All data are presented as mean ± standard error of the mean (SEM). P < 0.05 were considered statistically significant.

First, we have determined whether PL is toxic to human T cells. To this end, we prepared freshly isolated human PBTs and assessed T cell viability after treatment with 10 μM PL up to 24 h. Since, it was reported that 1511.9 ng/ml (4.8 μM) of PL could be found in rat plasma after 1 dose of 50 mg/kg PL intraperitoneal injection (36), in our study we considered concentrations of PL up to 10 μM as physiologically relevant. Even the highest PL concentration used in our experiments (10 μM) showed no significant cytotoxicity on primary human T cells (Supplementary Figure 2a). Interestingly, this concentration of PL induced cell death in malignant Jurkat T leukemia cells over the same time period (Supplementary Figure 2b) which is in line with previous studies showing cytotoxic effects of 10 μM PL on myeloid and B cell leukemia cells (37, 38).

PL exhibits potent anticancer activity in multiple cancer types by increasing the cellular ROS levels (39). ROS are also important regulators of T cell activation and function (24, 40, 41). The impact of PL on the redox system of primary human T cells, however, is unknown. We therefore evaluated the intracellular ROS levels in primary human T cells after PL treatment. To this end, the ROS-sensitive probe, CM-H₂DCFDA, was used to stain T cells, followed by T cell treatment with 1–10 μM PL for 1 h prior to flow cytometric analysis. Supplementary Figure 3a shows the gating of lymphocytes and the staining with CM-H₂DCFDA. PL treatment induced a significant increase of ROS in a concentration-dependent manner (Figure 1A). The intracellular ROS levels in primary human T cells after treatment with 5–10 μM PL were comparable to those observed after treatment with 50 μM H₂O₂. Similar to primary human T cells, the ROS levels in Jurkat leukemia cells also increased significantly after PL treatment (Figure 1A).

The response of cells to prooxidative substances is antagonized by antioxidant systems, primarily by GSH. Therefore, we next determined the intracellular levels of GSH, an important ROS-scavenger, using ThiolTracker^TM violet dye in both primary human T cells and Jurkat cells. Supplementary Figure 3b shows the gating of lymphocytes and the staining with ThiolTracker^TM. After 1 h treatment, 10 μM PL was able to significantly diminish GSH levels, whereas already 1 μM PL was sufficient to significantly decrease the amount of GSH in T cells after 24 h treatment (Figure 1B, left). In contrast, the GSH levels in Jurkat cells were even increased after 24 h treatment with 10 μM PL (Figure 1B, right).

GSH can form mixed disulfides with oxidized cysteine residues in proteins in response to mild oxidative stress, which is named S-glutathionylation. This process is considered as a defense mechanism to protect proteins from irreversible oxidative states (42). Since PL shifted the cellular redox equilibrium toward prooxidative conditions, we next investigated whether PL interfered with the S-glutathionylation of cellular proteins. For this purpose, we treated the cells with 5 μM PL for up to 4 h. Subsequently, cell lysates were loaded on SDS-PAGE and immunostained for S-glutathionylation using a monoclonal anti-GSH antibody. We observed a strong glutathionylation signal of a 42-kDa protein which decreased strongly (deglutathionylation) in T cells treated with PL (Figure 1C, lane 1-Ctrl and 2-PL). Notably, treatment of Jurkat cells with PL did not change the S-glutathionylation of the same protein band (Figure 1C, lane 3-Ctrl and 4-PL). This finding is consistent with the strongly diminished intracellular levels of GSH, the essential material for S-glutathionylation, only in PL-treated T cells but not in Jurkat cells.

The 42-kDa protein was further analyzed by liquid chromatography-tandem mass spectrometry (LC-MS²). LC-MS² analysis revealed that the highly S-glutathionylated 42-kDa band was beta-actin. To further elaborate this finding, we inspected which cysteines in actin were S-glutathionylated and whether PL treatment alters the relative levels of this modification. To this end, we searched for unmodified, S-glutathionylated (+305), sulfenylated (SOH), sulfinylated (SO₂H), and sulfonylated (SO₃H) forms of the cysteine-containing peptides. The peptide Ac-DDDIAALVVDNGSGMCK was detected as unmodified, S-glutathionylated (+305), sulfinylated (SO₂H, +32) and sulfonylated (SO₃H, +48) on Cys17 with MS² spectra at 50.4, 49.4, 50, and 50 min, respectively (Figure 1D and Supplementary Figures 4a,c). The MS² spectrum fingerprint for all peptides is comparable. b ions have the same masses and relative intensities in all spectra. Relative intensities of y ions are comparable in all spectra while their masses are (except for y₁) 305.0681, 31.990 or 47.986 m/z higher in the spectrum of the S-glutathionylated, sulfinylated or sulfonylated peptide, respectively (Figure 1E and Supplementary Figures 4b,d). These MS² spectra show that Cys17 is present as unmodified, S-glutathionylated (+305), sulfinylated (SO₂H) and sulfonylated (SO₃H) residue in the peptide Ac-DDDIAALVVDNGSGMCK.

For relative quantification of the peptides, peak areas were calculated with the Skyline software (32) and non-cysteine-containing peptides were used for normalization in each sample. The unmodified peptide Ac-DDDIAALVVDNGSGMCK was detected in the PL-treated sample, whereas it was absent in the control (Figure 1D, left). Cys17 glutathionylated peptide Ac-DDDIAALVVDNGSGMC(+305)K was detected with intense signals in both control and PL-treated samples (Figure 1D, right). Figure 1F shows that the intensity of the S-glutathionylated peptide (Cys17) decreased significantly upon PL treatment compared to the control sample. Moreover, lower levels of S-glutathionylation on Cys217 upon PL treatment were also found in two experiments, while it was not detected in the third experiment in both control and PL-treated samples (data not shown). Reasonably, sulfinylated (SO₂H) and sulfonylated (SO₃H) forms of the Cys17 in peptide Ac-DDDIAALVVDNGSGMCK in control and PL-treated sample were detected, and the signal intensities increased for both peptides with PL treatment (Supplementary Figures 4a,c,e). This indicates oxidation of the cysteine to sulfinic or sulfonic acid.

Taken together, PL exerts prooxidative activity on primary human T cells, based on an increase in ROS levels and a decrease in GSH levels. In Jurkat cells, PL also enhanced the intracellular ROS levels, but the levels of GSH remained stable or rather increased. In line with these observations, PL diminished the S-glutathionylation of actin in T cells but not in Jurkat cells.

Since excessive ROS production or prolonged exposure to high ROS levels impair T cell functions (24), we investigated whether PL influences T cell activation and T cell-dependent immune responses.

In a physiological setting, one of the first steps of T cell activation is the formation of a contact zone with an APC and the maturation of the immunological synapse (IS). To investigate this, control or PL-pretreated T cells (1 h or overnight pretreatment with 5 μM PL) were cultured with SEB-loaded Raji cells which served as APCs. First, T cell/APC conjugates were identified by flow cytometry (Figures 2A,B). As expected, in the absence of SEB, only 2.5 ± 0.25% T cell/APC conjugates were found, whereas this number of conjugates increased to 9.4 ± 0.8% in the presence of SEB. No effect on cell couple formation was observed after 1 h preincubation with 5 μM PL (data not shown), but overnight pretreatment with PL significantly decreased the formation of T cell/APC conjugates in the presence of SEB to 3.0 ± 0.79% (Figure 2B). In order to clarify whether this result is due to the prooxidative capacity of PL described above (compare Figure 1), we made use of the antioxidant NAC. As a synthetic precursor of intracellular cysteine and GSH, NAC can lead to replenishment of GSH stores (43). Intriguingly, we observed that NAC (3 mM) could prevent the PL-induced downregulation of T cell/APC conjugates formation (Figure 2B).

The impact of PL on formation of a mature immune synapse was investigated by multi-spectral imaging flow cytometry (MIFC). By defining regions of interest, MIFC allows the spatial quantification of the accumulation of receptors at the IS. T cell/APC couples were identified according to staining of nuclei with DAPI and detection of CD3 expression in T cells. The mature IS was then defined by accumulation of LFA-1 (green) and CD3 (red) in the T cell/APC contact area. As expected, while most T cells did not show LFA-1 and CD3 enrichment in the contact area without SEB treatment, a clear receptor enrichment -hence IS maturation- was observed in the presence of SEB. In samples pretreated overnight with PL, the enrichment of LFA-1 and CD3 in the contact zone was very low but it occurred normally when T cells were preincubated with PL and NAC (Figures 2C,D). Together, overnight pretreatment of T cells with PL impairs the formation of a mature IS between T cells and APCs, most likely due to depletion of GSH.

IS formation initiates and controls Ca²⁺ signals in T cells (44). We asked whether PL also influences the anti-TCR/CD3-induced Ca²⁺ signaling which occurs independently of T cell/APC conjugate formation. Figure 3A shows the Ca²⁺ flux signal monitored by flow cytometry. The parameter AUC (area under curve) was used to evaluate the Ca²⁺ signal in our experiments, which refers to the total amount of cytosolic Ca²⁺ during the whole recorded period of Ca²⁺ flux. These data revealed that the Ca²⁺ signal in control T cells increased steeply after CD3 cross-linking while overnight pretreatment of T cells with 5 μM PL inhibited the TCR/CD3-mediated Ca²⁺ flux. This inhibition was completely prevented by NAC (Figures 3A,B). Together, these results imply that PL also interferes with Ca²⁺ signaling in T cells by depleting GSH.

To investigate whether pretreatment with PL for 30 mins interferes with later T cell activation events, expression of the T cell activation markers CD69 and CD25 in primary human T cells was assessed after costimulation with anti-CD3/CD28 antibodies or isotype control antibodies. The flow cytometry analysis procedure is shown in Supplementary Figure 5a. T cells seeded in isotype control antibody-coated plates are similar to resting T cells according to the CD69 and CD25 surface staining (Supplementary Figure 5b). After costimulation with anti-CD3/CD28 antibodies for 24 h, CD69 and CD25 were highly expressed on the T cell surface in the absence of PL, while in the presence of PL, their expression was suppressed in a dose-dependent manner (Figure 4A and Supplementary Figure 5c). Intriguingly, while the MFI value of CD69 was decreased after PL treatment, the percentage of CD69-positive cells was not altered under the same conditions. Moreover, not only the surface expression, but also the total amount of CD69 and CD25 within the cells was decreased upon PL treatment (Supplementary Figure 5d). This indicates that PL inhibited the expression of both proteins rather than impairing the receptor transport to the surface. Again, treatment with NAC prevented the PL-induced downregulation of CD69 and CD25 (Figure 4A).

T cell proliferation over time can be monitored via labeling T cells with CFSE. To this end, we pretreated CFSE-labeled T cells with or without PL for 30 mins, and determined T cell proliferation by flow cytometry after 72 h of costimulation. As shown in Figure 4B, compared to untreated control cells, PL treatment significantly inhibited T cell proliferation. Again, NAC abolished the suppressive effect of PL on T cell proliferation. In order to determine whether these immunosuppressive effects are due to the depleted GSH stores, we repeated the experiments with exogenously added GSH. Indeed, inhibition of the expression of CD69 and CD25 (Figure 4C), and T cell proliferation (Figure 4D) were completely prevented in the presence of GSH. All together, these findings demonstrate that PL acts as an immunosuppressive agent on primary human T cells through depletion of intracellular GSH. PL induced immunosuppression can be prevented by the antioxidants GSH or NAC.

Next, we took advantage of the NanoString nCounter® Technology and the GX Human Immunology v2 panel to gain an unbiased view on the expression of differential mRNA profiles. The genes evaluated were related to leukocyte functions including major classes of cytokines and their receptors. The final counts of the mRNA transcripts were visualized by hierarchical clustering (Figure 5A). For these experiments, primary human T cells were again pretreated for 30 min with PL or solvent control and then costimulated with anti-CD3/CD28 antibodies for 4 h. Upon PL treatment, 81 genes were significantly downregulated and 27 genes were significantly upregulated (fold change >2, p < 0.05) compared to the control. Notably, consistent with our finding of diminished CD25 (IL-2RA) at the protein level (compare Figure 4A), the mRNA level of IL-2RA was also substantially decreased by PL treatment (Figure 5B, red dot). Intriguingly, other genes that are important for T cell activation and proliferation were also inhibited upon PL treatment, e.g., CD2 and the TNF receptor superfamily member 9 [TNFRSF9, a TNF receptor that possesses costimulatory activity for activated T cells (45)] and JAK2, JAK3, and STAT5A [an essential mediator of IL-2 signaling in T cells (46)] (Figure 5B, red dots), as well as the proinflammatory cytokines IFNγ and TNFα (Figure 5B, green dots). Moreover, the expression of the key proinflammatory cytokine IL-1B (47), the chemokine (C-C motif) ligands (CCL) 3 and CCL4, which can be produced by T cells following nonspecific activation or T cell receptor engagement (48) were inhibited by PL treatment (Figure 5B, green dots). In addition, PL was able to suppress the expression of cytolytic enzymes like Granzyme (Grz) A, GrzB and Perforin, which are critical mediators of anti-viral and anti-tumor immunity (Figure 5B, blue dots). These results, together with the aforementioned findings, demonstrate that PL is an inhibitor for T cell activation and possesses anti-inflammatory properties. Furthermore, our data imply that PL is a potential suppressor of T cell-mediated cytotoxicity.

Intriguingly, the expression of T_H17-specific cytokines, e.g., IL-17A, IL-17F, and IL-22 as well as expression of B cell activating transcription factor (BATF) (49) and interferon regulatory factor 4 (IRF4) (50), which together initiate a transcriptional program for T_H17 cell development (51), were inhibited by PL treatment (Figure 5C, red dots). Expression of RORC gene was, however, unaffected. We have then further checked genes associated with other T cell subpopulations, i.e., T_H1, T_H2, and T_reg cells. In addition to TNFα, the expression of IL-2 and the T_H1 cytokine IFNγ were inhibited by PL although the differences to untreated control cells were not statistically significant. PL significantly inhibited the T_H1-specific transcription factor TBX21 (Figure 5C, green dots). The expression of T_H2-associated cytokines, e.g., IL-3, IL-4, IL-5, and IL-13 was in part suppressed by PL, but not significantly and to a lower extent compared to T_H1 cytokines (Figure 5C, sky blue dots). The T_H2-specific transcription factor GATA3 was even increased by tendency upon PL treatment (Figure 5C, sky blue dot). Following costimulation for 4 h, the T_reg-related transcription factor FOXP3 remained unaffected by PL, but BATF3, a transcription factor that prevents the differentiation of T_reg cells (52), was inhibited significantly by PL (Figure 5C, yellow dots). These results together provided first hints that PL acts differently on various T cell populations: At the gene expression level, T_H17-associated cytokines were more sensitive to PL treatment compared to T_H1 and T_H2 cytokines (Figure 5C).

To test these findings at the protein level, the cellular content of IL-2, IFNγ, and TNFα was determined by intracellular staining. Consistent with our findings of diminished mRNA levels of IL-2, IFNγ, and TNFα (Figures 5B,C), also the intracellular protein levels of these cytokines were substantially decreased by PL treatment (Figure 5D). Note that FMO samples showed that there is no nonspecific staining for these cytokines (Supplementary Figure 6a). Also, at the protein level, the PL-mediated decrease of IL-2, IFNγ, and TNFα was prevented when the cells were coincubated with NAC, which is consistent with the hypothesis that depletion of GSH plays a crucial role in PL-mediated T cell immunomodulation. Of note, NAC alone did not upregulate expression of IL-2, IFNγ, and TNFα (Supplementary Figure 6b).

The results from the nCounter gene analysis suggested that PL differentially affects T cell subpopulations. To further investigate this finding, we have next measured the effect of PL on the differentiation of naïve CD4⁺ T cells into T_H1, T_H2, T_H17, and T_reg cells. Naïve CD4⁺ T cells were cultured under appropriate polarizing differentiation conditions (+cytokines) in the absence or presence of 1 μM PL (+cytokines+PL) as described in Materials and Methods. After 6 days, the differentiation into T_H1, T_H2, and T_H17 cells was assessed by surface marker staining (Figure 6A and supplementary Figure 7a) and cytokine detection (Figures 6C–E). Nuclear FOXP3 staining was used for identifying T_reg cells (Figure 6B and Supplementary Figure 7b). 7-AAD staining confirmed that cells did not die after 6 days of 1 μM PL treatment (data not shown). Quantification of the T cell subpopulations (Figures 6A,B) confirmed that the percentage of T_H1, T_H2, T_H17, and T_reg cells in the respective polarizing conditions (+cytokines) was much higher than with anti-CD3/28 costimulation alone (none), which proved that the polarizing conditions were optimal. PL inhibited the differentiation of naïve CD4⁺ T cells into T_H17 cells while it had no clear effect on the differentiation into T_H1 and T_H2 cells, as assessed by staining of the respective T cell surface markers (Figure 6A). Interestingly, FOXP3 staining showed that PL significantly promoted the differentiation of naïve CD4⁺ T cells into T_reg cells (Figure 6B). This notion holds true not only for the proportion of FOXP3⁺ T_reg cells, but also for the absolute number of FOXP3⁺ T_reg cells. Importantly, NAC treatment prevented the PL mediated decrease in the T_H17 population and the PL mediated increase in the T_reg population (+cytokines+PL+NAC) (Figures 6A,B).

To get a more detailed picture about the consequences of PL for the differentiation of naïve CD4⁺ T cells into T_H17 and T_reg cells, we analyzed the levels of the anti-inflammatory T_reg cell cytokine TGF-β, and the proinflammatory T_H17 cell cytokines IL-17A and IL-17F. Indeed, in line with the increased proportion of FOXP3⁺ cells (Figure 6B), under T_reg polarizing differentiation conditions the concentration of TGF-β also increased significantly upon treatment with PL (+cytokines+PL) compared to the control without PL (+cytokines). And again, NAC prevented this effect (Figure 6C). As expected, PL treatment (+cytokines+PL) significantly diminished the levels of the proinflammatory T_H17 cell cytokines IL-17A and IL-17F (Figures 6D,E). Surprisingly, these cytokine levels remained unaffected by NAC.

It was reported that HIF-1α, a subunit of the heterodimeric transcription factor HIF-1, enhances T_H17 development and attenuates T_reg development (53). Intriguingly, we could show by western blot analysis that PL strongly inhibits the expression of HIF-1α in a dose-dependent manner starting already at a concentration of 1 μM PL. This inhibition was prevented by NAC treatment (Figure 6F). Thus, the diminished expression of HIF-1α may provide one molecular explanation for the PL-induced downregulation of T_H17-related genes. Taken together, PL affects T cell subpopulations differently. It inhibits the differentiation of naïve CD4⁺ T cells into proinflammatory T_H17 cells, while promoting the differentiation into T_reg cells.

Given the newly discovered immunomodulatory role of PL on primary human T cells, it was tempting to speculate that PL could be beneficial for patients with autoimmune diseases. Therefore, we analyzed the impact of PL on whole blood samples from patients with RA and systemic lupus erythematosus (SLE), which are prototypical autoimmune rheumatic diseases (54), as well as from patients with TNF receptor-associated periodic syndrome (TRAPS), a very rare autoinflammatory disease. For these analyses, 100 μl of heparin blood were pretreated with 5 μM PL and activated with SEB overnight. Erythrocytes were then lysed and leukocytes were permeabilized, stained with fluorescently-labeled antibodies specific for CD3, CD69, CD25, IFNγ, and IL-2 and assessed by flow cytometry. In the presence of PL, the expression of the T cell activation markers CD25 (Figure 7A, and Supplementary Figure 8) and CD69 (Figure 7B, and Supplementary Figure 8) after SEB stimulation was significantly downregulated in the lymphocyte population (CD3⁺). Consistently, a PL-mediated inhibition was also observed on the production of IL-2 (Figure 7C) and the proinflammatory cytokine IFNγ (Figure 7D). Taken together, this experiment confirmed the immunosuppressive activity of PL on human T cells in clinically relevant samples. This implies that PL might be a promising substance for the treatment of autoimmune diseases.