Abstract
CD1d-restricted activation of invariant NKT (iNKT) cells results in the abundant production of various types of cytokines and the subsequent modulation of immune responses. This has been shown to be relevant in several clinical disorders, including cancer, autoimmunity, and graft tolerance. Although it is well known that the suppressive function of regulatory T cells is critically dependent on the FOXP3 gene, FOXP3 can also be expressed by conventional human T cells upon activation, indicating the lack of specificity of FOXP3 as a marker for suppressive cells. In this study, we report that the mammalian target of rapamycin (mTOR) inhibitor rapamycin and IL-10, but not TGF-β, can induce FOXP3 expression in iNKT cell lines. Importantly, however, FOXP3+ iNKT cells only acquired suppressive abilities when cultured in the presence of the mTOR inhibitor rapamycin. Suppression of responder T cell proliferation by FOXP3+ iNKT cells was found to be cell contact–dependent and was accompanied by a reduced capacity of iNKT cells to secrete IFN-γ. Notably, imaging flow cytometry analysis demonstrated predominant nuclear localization of FOXP3 in suppressive FOXP3+ iNKT cells, whereas nonsuppressive FOXP3+ iNKT cells showed a predominance of cytoplasmically localized FOXP3. In conclusion, whereas IL-10 can enhance FOXP3 expression in iNKT cells, mTOR inhibition is solely required for promoting nuclear localization of FOXP3 and the induction of suppressive FOXP3+ iNKT cells.
Introduction
Invariant NKT cells (iNKT) constitute a specific T cell subset that is restricted by the nonpolymorphic CD1d Ag-presenting molecule and is characterized by the expression of an invariant TCR α-chain, Vα24-Jα18 in humans and Vα14-Jα18 in mice, preferentially paired with Vβ11 in humans and Vβ8.2, Vβ7, or Vβ2 in mice. Certain glycolipid Ags, such as α-galactosylceramide (α-GalCer), can be presented by CD1d to iNKT cells, resulting in their activation. Upon activation, iNKT cells rapidly produce a wide variety of pro- and anti-inflammatory cytokines, and it is the balance between these immunostimulatory and immunoregulatory cytokines that appears to determine the outcome of ensuing immune responses, for example, in the setting of autoimmunity, infection, transplantation, and cancer (1–3). Of interest, recent studies demonstrated the induction of FOXP3 expression in iNKT cells upon exposure to TGF-β (4, 5) and the induction of an immunosuppressive phenotype in the additional presence of the mammalian target of rapamycin (mTOR) inhibitor rapamycin in iNKT cells derived from cord blood or PBMC (6). Although expression of FOXP3 is classically associated with CD4+CD25hiFOXP3+ regulatory T cells (Tregs), which represent a functionally distinct lineage of immunoregulatory T cells that is critically dependent on this transcription factor (7, 8), human conventional T cells can transiently express FOXP3 upon activation (9, 10), as can iNKT cells upon exposure to TGF-β, as mentioned above (4–6).
In this study, we undertook a comparative investigation of the influence of the suppressive cytokines IL-10 and TGF-β, alone or combined with the mTOR inhibitor rapamycin, on the expression and localization of FOXP3 in differentiated iNKT cell lines, taking into account phenotypic and functional characteristics of these differentially conditioned iNKT cells to acquire insight into possible uses of these cells for adoptive transfer in transplantation medicine on the one hand and consequences for the treatment of cancer on the other hand. In contrast to an earlier study in which, instead of established cell lines, freshly isolated peripheral blood iNKT cells were used, we found that IL-10 rather than TGF-β induced FOXP3 expression. Moreover, we report that whereas IL-10 can enhance FOXP3 expression in iNKT cells, rapamycin is solely required for nuclear localization of FOXP3 and the induction of suppressive FOXP3+ iNKT cells (iNKTregs).
Materials and Methods
Generation of monocyte-derived dendritic cells
Monocytes were isolated from PBMC by MACS with the use of CD14 MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. After isolation, monocytes were cultured for 5–7 d at 37°C in a humidified atmosphere under 5% CO2 in the presence of 10 ng/ml recombinant human (rh)IL-4 (R&D Systems, Minneapolis, MN) and 100 U/ml rhGM-CSF (Genzyme/Bayer HealthCare Pharmaceuticals, Seattle, WA) in RPMI 1640 (Lonza, Basel, Switzerland) supplemented with 100 IU/ml sodium penicillin (Astellas Pharma, Leiden, the Netherlands), 100 μg/ml streptomycin sulfate (Radiumfarma-Fisiofarma, Napels, Italy), 2.0 nM l-glutamine (Life Technologies, Bleiswijk, the Netherlands), 10% FBS (HyClone, Amsterdam, the Netherlands), and 0.05 mM 2-ME (Merck, Darmstadt, Germany), hereafter referred to as complete medium. Immature monocyte-derived dendritic cells (moDC) were matured with 100 ng/ml LPS (Sigma-Aldrich, St. Louis, MO) for 24–48 h. moDC were harvested with 5 mM EDTA in PBS (Braun Melsungen, Melsungen, Germany), irradiated (5000 rad), and used for surface marker expression by flow cytometry and iNKT cell line cultures.
Generation of iNKT cell lines
As previously described (11), iNKT (defined as Vα24+Vβ11+) cell lines were generated from PBMC by MACS sorting using the 6B11 mAb (a gift of Mark Exley, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA), or the murine anti-human TCR Vα24-chain mAb (Beckman Coulter), combined with a polyclonal goat anti-mouse Ab labeled with magnetic beads (Miltenyi Biotec). iNKT cells were then cocultured in complete medium with mature moDC that were pulsed with 100 ng/ml α-GalCer (Funakoshi, Tokyo, Japan) during maturation and 50–100 U/ml rhIL-2 (Proleukin, Novartis, Arnhem, the Netherlands). Purified cells were cultured by weekly restimulation with α-GalCer–loaded mature moDC in a 1:5–1:10 ratio. Purity of iNKT cell lines used for experiments was >90%.
Effects of IL-10, TGF-β, and rapamycin on iNKT cell lines
Resting iNKT cells (defined as ≥6 d after restimulation) were cultured in complete medium in a 48-well plate with immature moDC loaded for 24 h with α-GalCer (2 × 105 iNKT and 4 × 104 moDC) and 50 U/ml rhIL-2 in the presence of 50 ng/ml rhIL-10 (eBioscience, San Diego, CA), 5 ng/ml rhTGF-β1 (BioVision, Milpitas, CA), and/ or 20 nM rapamycin (Calbiochem, Merck, Darmstadt, Germany). After 7 d, iNKT cells were harvested for further analyses.
Flow cytometry
moDC were analyzed by flow cytometry using FITC- or PE-labeled Abs against IgG1, CD14, and CD1a (BD Biosciences, East Rutherford, NJ). iNKT cells were stained with FITC-labeled Vα24, PE-labeled Vβ11 (both from Beckman Coulter), PE-labeled CD25, PerCP-Cy5.5–labeled CD4, allophycocyanin-labeled CD25 (BD Biosciences), allophycocyanin-labeled latency-associated peptide (LAP; R&D Systems), and FITC-labeled Helios (BioLegend, San Diego, CA). Stainings were performed in PBS supplemented with 0.1% BSA and 0.02% sodium azide for 30 min at 4°C. Intracellular staining was performed after fixation and permeabilization using a fixation/permeabilization kit according to the manufacturer’s protocol (eBioscience). For staining of CTLA-4, a PE-labeled Ab against CTLA-4 (BD Biosciences) was used, and FOXP3 was stained with either PCH101 PE (eBioscience) or 259D Alexa Fluor 488 (BioLegend) anti-FOXP3 mAbs.
For intracellular IFN-γ staining, iNKT cells were stimulated for 4 h with 50 ng/ml PMA and 500 ng/ml ionomycin in the presence of brefeldin A (1:500; GolgiPlug, BD Biosciences) and stained for CD25, CD4, FOXP3, and IFN-γ (BD Biosciences) using the eBioscience fixation/permeabilization kit. Live cells were gated based on forward and side scatter and analyzed on a BD FACSCalibur (BD Biosciences) using CellQuest or Kaluza analysis software (Beckman Coulter).
Suppression assay
The capacity of cultured iNKT cells to suppress proliferation of allogeneic CD8+ T responder cells was determined by labeling responder T cells with 1 μM CFSE (Sigma-Aldrich) cultured in a 96-well round-bottom plate at a concentration of 5 × 104 cells/well in complete medium in the presence of 1 μg/ml anti-CD3 mAb, 1 μg/ml anti-CD28 mAb (clones 16A9 and 15E8, provided by Dr. René van Lier, Sanquin, Amsterdam, the Netherlands), and 20 U/ml rhIL-2 with or without the addition of cultured iNKT cells in an iNKT/T responder ratio of 1:1 and 1:2. After 4 d of coculture, cells were stained with allophycocyanin-labeled CD8 (BD Biosciences), and proliferation of CD8+ responder T cells was analyzed by assessing CFSE dilution. For Transwell assays, 7.5 × 105 CFSE-labeled CD8+ responder T cells were cultured in 24-well plates (lower compartment) with 7.5 × 105 iNKT cells in the Transwell insert (0.4 μm pore size, Costar Corning) for 4 d. After 4 d, cultures were stained with allophycocyanin-labeled CD8, and proliferation of responder T cells was assessed as described above. Relative proliferation was calculated by the equation % proliferation = (% responder T cells that proliferated when cultured in the presence of iNKT cells/% responder T cells that proliferated when cultured alone) × 100 and relative suppression was calculated by the equation % suppression = 100 − [(% responder T cells that proliferated when cultured in the presence of iNKT cells/% responder T cells that proliferated when cultured alone) × 100].
Intracellular FOXP3 localization experiments
To determine the intracellular FOXP3 localization, iNKT cells were cultured with 50 ng/ml rhIL-10, 5 ng/ml rhTGF-β1, and/or 20 nM rapamycin for 4–7 d, stained with Hoechst and FOXP3–Alexa Fluor 488 using the eBioscience fixation/permeabilization kit as previously described (12), and subsequently analyzed using imaging flow cytometry (ImageStream X-100, Amnis-Millipore). A CD4+CD25+ Treg isolation was performed resulting in two fractions, CD4+CD25−/int T cells (conventional T cells), which were used as a negative control, and CD4+CD25hi T cells (Tregs), which were used as a positive control. Both fractions were activated, stained, and analyzed. At least 5000 cells per sample were acquired. For the analysis, a mask to delineate the nucleus was made based on the Hoechst signal denoting the nuclei. The ratio of the amount of FOXP3 in the entire cell versus the nuclear mask was calculated and this ratio was log transformed and termed “nuclear translocation score” (see also Supplemental Fig. 1).
Statistical analysis
One-way or two-way repeated measures ANOVA was used to determine statistical significance of differences between groups with Bonferroni posttests. Findings were considered statistically significant when p values were ≤0.05. Statistical analyses were performed using GraphPad Prism software version 5.02.
Results
Effect of IL-10, TGF-β, and rapamycin on iNKT cell FOXP3 and CTLA-4 expression
As IL-10 and TGF-β are known to be involved in the induction of Tregs (13), we assessed the relative ability of these cytokines to induce FOXP3 expression in iNKT cell lines (Fig. 1), generated as described previously (11). In the evaluated iNKT cell lines, CD4 was expressed by 49% (mean; range, 4.3–96%; n = 10) of the iNKT cells, whereas CD161 was expressed by 42% (mean; range, 15–76%; n = 3) of the iNKT cells. iNKT cell lines were cultured with α-GalCer–loaded immature moDC and IL-2 for 7 d with or without the addition of 50 ng/ml IL-10, 5 ng/ml TGF-β, and/or 20 nM rapamycin. FOXP3+ iNKT cells were defined as CD25hi, excluding CD25int/lowFOXP3+ iNKT cells, which might rather be activated iNKT cells than potentially suppressive iNKT cells (Fig. 1A), in analogy to the established gating strategy used for the identification of bona fide Tregs (14). Compared to iNKT cells cultured in medium (iNKT control), FOXP3 expression increased when iNKT cells were cultured with IL-10, rapamycin, or the combination of IL-10 and rapamycin. Of note, and in contrast to previous reports (4–6), FOXP3 expression did not increase when iNKT cells were cultured with TGF-β. Compared to the iNKT control, a slight increase in FOXP3 expression was observed when iNKT cells were cultured with TGF-β and rapamycin, although this increase was not statistically significant (Fig. 1B). To evaluate whether this observed relative increase also represented an increase in the absolute number of FOXP3+ iNKT cells in the culture, the mean (±SEM) expansion factor of absolute numbers of CD25hiFOXP3+ iNKT cells was calculated for all six conditions: control, 0.6 (±0.2); rapamycin, 1.4 (±0.6); IL-10, 2.0 (±0.6); IL-10 plus rapamycin, 2.2 (±0.9); TGF-β, 0.8 (±0.4); TGF-β plus rapamycin, 0.8 (±0.3). These data show that in the three conditions with a significantly increased frequency of CD25hiFOXP3+ iNKT cells compared with the control condition as shown in Fig. 1B, that is, rapamycin, IL-10, and IL-10 plus rapamycin, there was also an increase in the absolute number of CD25hiFOXP3+ iNKT cells, indicating the true induction and expansion of FOXP3+ iNKT cells and not merely selective survival.
IL-10 and rapamycin upregulate FOXP3 and CTLA-4 expression in iNKT cells. FOXP3 and CTLA-4 expression was determined after 7 d of coculture of iNKT cells with α-GalCer–loaded immature moDC and IL-2 in the presence of medium (iNKT control), rapamycin, IL-10, IL-10 and rapamycin, TGF-β, or TGF-β and rapamycin. (A) Representative dot plots illustrating the gating of CD25hiFOXP3+ iNKT cells after culture of iNKT cells in medium (iNKT control, left dot plot) or in the presence of rapamycin (right dot plot). (B and C) The percentages of CD25hiFOXP3+ and CD25hiCTLA-4+ iNKT cells were assessed by flow cytometry according to the gating strategy shown in (A). Means ± SEM are shown. (B) n = 10, (C) n = 4; one-way repeated measures ANOVA with Bonferroni posttest. *p ≤ 0.05, **p < 0.01, *** p < 0.001.
IL-10 and rapamycin upregulate FOXP3 and CTLA-4 expression in iNKT cells. FOXP3 and CTLA-4 expression was determined after 7 d of coculture of iNKT cells with α-GalCer–loaded immature moDC and IL-2 in the presence of medium (iNKT control), rapamycin, IL-10, IL-10 and rapamycin, TGF-β, or TGF-β and rapamycin. (A) Representative dot plots illustrating the gating of CD25hiFOXP3+ iNKT cells after culture of iNKT cells in medium (iNKT control, left dot plot) or in the presence of rapamycin (right dot plot). (B and C) The percentages of CD25hiFOXP3+ and CD25hiCTLA-4+ iNKT cells were assessed by flow cytometry according to the gating strategy shown in (A). Means ± SEM are shown. (B) n = 10, (C) n = 4; one-way repeated measures ANOVA with Bonferroni posttest. *p ≤ 0.05, **p < 0.01, *** p < 0.001.
Additionally, when determining the expression of another marker constitutively expressed on Tregs, that is, the key costimulatory molecule CTLA-4 (15), we observed a similar expression pattern as for FOXP3, with a significant increase of CTLA-4 expression in iNKT cells cultured with IL-10 and rapamycin (Fig. 1C). FOXP3 induction was assessed both in CD4low (n = 4) and in CD4hi (n = 6) iNKT cell lines. CD4low iNKT cell lines had a mean (±SEM) expression of CD4 of 7.3 (±1.2%) and CD4hi iNKT cell lines of 79.4 (±5.6%). Because both CD4hi and CD4low iNKT cell lines responded to IL-10, TGF-β, and rapamycin in a similar pattern, their results are collectively shown in Fig. 1.
Expression of LAP and Helios in FOXP3− and CD25hiFOXP3+ iNKT cells
To investigate whether the iNKT cells cultured in the presence of IL-10, TGF-β, and/ or rapamycin shared further subset-defining markers with Tregs (16), we analyzed additional markers associated with regulatory function, that is, LAP and Helios. Because a previous report showed no effect of GITR blockade on the suppressive capacity of iNKT cells (6), GITR expression was not analyzed.
No major differences were observed in expression of LAP on FOXP3− versus CD25hiFOXP3+ iNKT cells; however, CD25hiFOXP3+ iNKT cells showed a higher LAP expression compared with FOXP3− iNKT cells when TGF-β was added to the culture (Fig. 2A). Overall LAP levels were low. Expression of Helios on CD25hiFOXP3+ iNKT cells was significantly increased compared with FOXP3− iNKT cells when rapamycin or TGF-β and rapamycin were added (Fig. 2B). Moreover, addition of IL-10 and rapamycin resulted in a significant decrease in Helios expression in CD25hiFOXP3+ iNKT cells compared with CD25hiFOXP3+ iNKT cells cultured in the presence of rapamycin. Altogether, cytokine- or rapamycin-mediated FOXP3 induction in iNKT cells was not associated with a clear induction of Treg-defining marker expression levels.
Expression of LAP and Helios in FOXP3− and FOXP3+ iNKT cells in response to IL-10, TGF-β, and/or rapamycin. Expression of LAP (A) and Helios (B) was determined after 7 d of culture of iNKT cells with α-GalCer–loaded immature moDC and IL-2 in the presence of IL-10, TGF-β, and/or rapamycin. Percentages of LAP and Helios expression were assessed on FOXP3− and CD25hiFOXP3+ iNKT cells by flow cytometry. Means ± SEM are shown. n = 4; one-way and two-way repeated measures ANOVA with Bonferroni posttests. *p ≤ 0.05.
Expression of LAP and Helios in FOXP3− and FOXP3+ iNKT cells in response to IL-10, TGF-β, and/or rapamycin. Expression of LAP (A) and Helios (B) was determined after 7 d of culture of iNKT cells with α-GalCer–loaded immature moDC and IL-2 in the presence of IL-10, TGF-β, and/or rapamycin. Percentages of LAP and Helios expression were assessed on FOXP3− and CD25hiFOXP3+ iNKT cells by flow cytometry. Means ± SEM are shown. n = 4; one-way and two-way repeated measures ANOVA with Bonferroni posttests. *p ≤ 0.05.
Rapamycin reduces iNKT cell IFN-γ production
iNKT cells are well known for their capacity to produce high levels of IFN-γ upon stimulation, resulting in enhanced cell-mediated immunity (1–3). Because FOXP3 expression is rather associated with suppressive functions, we examined IFN-γ production by iNKT cells cultured in the presence of IL-10, TGF-β, and/or rapamycin. As shown in Fig. 3, the addition of rapamycin resulted in a decreased capacity of iNKT cells to produce IFN-γ in response to a 4-h stimulation with PMA and ionomycin. This effect was irrespective of the presence of IL-10 or TGF-β, indicating that rapamycin skewed iNKT cells into a less proinflammatory phenotype consistent with a conversion to a more Treg-like state. Because the activation induced by the 4-h PMA/ionomycin stimulation resulted in FOXP3 expression in all iNKT cells, it was impossible to perform specific gating on FOXP3− and CD25hiFOXP3+ iNKT subsets in these experiments. Interestingly, however, when comparing IFN-γ production by CD25int and CD25hi iNKT cells, we noted that especially the CD25hi iNKT cell subset expressed reduced levels of IFN-γ (data not shown), suggesting that especially this subset had acquired a Treg-like state.
Rapamycin reduces iNKT cell IFN-γ production. (A) Bar graph showing the mean fluorescence intensity (MFI) of IFN-γ relative to the production of IFN-γ in the control condition. Means ± SEM are shown. n = 4; one-way repeated measures ANOVA with Bonferroni posttest. *p ≤ 0.05. (B) Representative histograms of IFN-γ production in iNKT cells cultured with IL-10 without a 4-h PMA/ionomycin stimulation (left histogram), iNKT cells cultured with IL-10 (middle histogram), or IL-10 and rapamycin (right histogram) after a 4-h stimulation with PMA/ionomycin. MFI of IFN-γ expression is indicated.
Rapamycin reduces iNKT cell IFN-γ production. (A) Bar graph showing the mean fluorescence intensity (MFI) of IFN-γ relative to the production of IFN-γ in the control condition. Means ± SEM are shown. n = 4; one-way repeated measures ANOVA with Bonferroni posttest. *p ≤ 0.05. (B) Representative histograms of IFN-γ production in iNKT cells cultured with IL-10 without a 4-h PMA/ionomycin stimulation (left histogram), iNKT cells cultured with IL-10 (middle histogram), or IL-10 and rapamycin (right histogram) after a 4-h stimulation with PMA/ionomycin. MFI of IFN-γ expression is indicated.
In contrast to a previous report, we found no evidence that exposure of iNKT cells to TGF-β alone resulted in decreased iNKT cell IFN-γ production (6). Therefore, although culture of iNKT cells in the presence of IL-10, rapamycin, or the combination of IL-10 and rapamycin resulted in an increase in iNKT cell FOXP3 expression, this expression of FOXP3 was not directly correlated with a decrease in IFN-γ production, as FOXP3 was similarly induced in iNKT cells cultured with just IL-10 (Fig. 1), whereas no reduction in IFN-γ production was observed in this condition.
Rapamycin is required for the induction of suppressive function of iNKT cells
To assess whether the iNKT cells cultured in the presence of IL-10, TGF-β, and/or rapamycin were able to suppress proliferation of conventional effector T cells, iNKT cells were cultured with α-GalCer–loaded immature moDC and IL-2 in the presence of IL-10, TGF-β, and/or rapamycin for 7 d, and subsequently harvested and cocultured with CFSE-labeled CD8+ responder T cells. Of note, for suppression assays CD4hi iNKT cell lines were used with a mean (±SEM) CD4 expression of 76.9 ± 11.6%. Cell division of the responder T cells was assessed after a 4-d culture period of responder T cells with different ratios of iNKT cells in the presence of 1 μg/ml anti-CD3 mAb, 1 μg/ml anti-CD28 mAb, and 20 U/ml IL-2.
iNKT cells cultured with α-GalCer–loaded immature moDC and IL-2 (iNKT control) were not able to suppress proliferation of responder T cells, and neither were iNKT cells cultured in the presence of IL-10 or TGF-β. In contrast, iNKT cells cultured in the presence of rapamycin, with or without the addition of IL-10 or TGF-β, were able to suppress the proliferation of responder T cells with statistically significant suppression observed for all rapamycin conditions as compared with the control, or compared with the corresponding cytokine conditions (Fig. 4). Furthermore, a significant correlation was observed between the percentages of CD25hiFOXP3+ iNKT cells within the differentially conditioned bulk cultures and their suppressive activity (Supplemental Fig. 2), underscoring that CD25hiFOXP3+ iNKT cells are most likely responsible for the observed suppression. Because a previous report studying iNKT cell conditioning directly ex vivo did not analyze the induction of FOXP3 expression or the induction of suppressive functionality caused by rapamycin alone (6), we additionally studied the effects of rapamycin on freshly isolated iNKT cells from peripheral blood and found a similar dominant effect of rapamycin on ex vivo iNKT cells in terms of induction of FOXP3 expression and suppressive activity as observed in iNKT cell lines (Supplemental Fig. 3).
Rapamycin induces suppressive functionality of iNKT cells. iNKT cells were cultured with α-GalCer–loaded immature moDC and IL-2 in the presence of IL-10, TGF-β, and/or rapamycin. Their capacity to suppress T cell proliferation was tested by measuring CFSE dilution of stimulated CD8+ responder T cells using anti-CD3 mAb, anti-CD28 mAb, and IL-2. (A) Representative histograms showing CFSE dilution of responder T cells alone (left histogram) or in the presence of medium-cultured (middle histogram) or rapamycin-cultured iNKT cells (right histogram). Percentages of responder T cell proliferation are indicated. (B) Bar graph showing the proliferation of responder T cells cultured with iNKT cells in the presence of medium (iNKT control), IL-10, TGF-β, and/or rapamycin relative to responder T cells alone. Means ± SEM are shown. n = 3; one-way repeated measures ANOVA with Bonferroni posttest. **p < 0.01, ***p < 0.001.
Rapamycin induces suppressive functionality of iNKT cells. iNKT cells were cultured with α-GalCer–loaded immature moDC and IL-2 in the presence of IL-10, TGF-β, and/or rapamycin. Their capacity to suppress T cell proliferation was tested by measuring CFSE dilution of stimulated CD8+ responder T cells using anti-CD3 mAb, anti-CD28 mAb, and IL-2. (A) Representative histograms showing CFSE dilution of responder T cells alone (left histogram) or in the presence of medium-cultured (middle histogram) or rapamycin-cultured iNKT cells (right histogram). Percentages of responder T cell proliferation are indicated. (B) Bar graph showing the proliferation of responder T cells cultured with iNKT cells in the presence of medium (iNKT control), IL-10, TGF-β, and/or rapamycin relative to responder T cells alone. Means ± SEM are shown. n = 3; one-way repeated measures ANOVA with Bonferroni posttest. **p < 0.01, ***p < 0.001.
Therefore, rapamycin is responsible for the induction of iNKT cell suppressive functions, irrespective of the addition of the other putative immunosuppressive cytokines. To formally exclude a possible effect of serum- or iNKT cell–derived TGF-β on the induction of FOXP3 expression, we additionally performed experiments in which a neutralizing TGF-β mAb was added to rapamycin-conditioned iNKT cultures and found that neutralization of TGF-β did not interfere with either the induction of FOXP3 in iNKT cells or their suppressive potential (data not shown). Furthermore, it is clear that expression of FOXP3 per se does not reflect the suppressive potential of iNKT cells, because FOXP3 was induced in the IL-10 only condition (Fig. 1) without the apparent acquisition of suppressive capacity or phenotype.
iNKT cells predominantly suppress T cell proliferation in a cell contact–dependent mechanism
Several mechanisms have been shown to be involved in Treg-mediated suppression, and despite intensive research, the exact mechanism still remains unclear (17). Tregs have been shown to be able to exert suppression in a contact-dependent and contact-independent manner, with the involvement of cell surface molecules, for example, CTLA-4 or LAP (18–20), or secreted cytokines, for example, TGF-β or IL-10 (reviewed in Ref. 21), respectively. To investigate the mechanism of suppression by iNKT cells, Transwell assays were performed. CFSE-labeled responder T cells were cultured in a 24-well plate, whereas iNKT cells were cultured in the Transwell insert, thereby only allowing exchange of soluble factors. Compared to conditions in which iNKT cells were in direct contact with responder T cells, Transwell experiments showed a significant reduction in the capacity of iNKT cells to inhibit proliferation of responder T cells (Fig. 5). These data indicate that suppression of proliferation by iNKT cells is primarily mediated by cell contact–dependent mechanisms.
iNKT cells predominantly suppress T cell proliferation via a cell contact–dependent mechanism. iNKT cells cultured with α-GalCer–loaded immature moDC and IL-2 for 7 d in the presence of rapamycin, IL-10, and rapamycin or TGF-β and rapamycin in coculture (●) with CD8+ responder T cells or cultured in the Transwell insert (▪) with responder T cells in the well of a 24-well plate (lower compartment). After 4 d, CFSE dilution of responder T cells was evaluated by flow cytometry. Each symbol represents one condition; horizontal bars indicate mean. n = 3; one-way repeated measures ANOVA with Bonferroni posttest. *p ≤ 0.05, **p < 0.01.
iNKT cells predominantly suppress T cell proliferation via a cell contact–dependent mechanism. iNKT cells cultured with α-GalCer–loaded immature moDC and IL-2 for 7 d in the presence of rapamycin, IL-10, and rapamycin or TGF-β and rapamycin in coculture (●) with CD8+ responder T cells or cultured in the Transwell insert (▪) with responder T cells in the well of a 24-well plate (lower compartment). After 4 d, CFSE dilution of responder T cells was evaluated by flow cytometry. Each symbol represents one condition; horizontal bars indicate mean. n = 3; one-way repeated measures ANOVA with Bonferroni posttest. *p ≤ 0.05, **p < 0.01.
IL-10 and rapamycin differentially affect intracellular FOXP3 localization in iNKT cells
Our data showed that whereas IL-10 and rapamycin both promote FOXP3 expression in iNKT cells, rapamycin was required for the induction of the contact-dependent suppressive capacity of iNKT cells and was also responsible for a decreased capacity of iNKT cells to produce IFN-γ in response to PMA/ionomycin stimulation. This implies that expression levels of FOXP3 alone do not reflect the suppressive potential of iNKT cells. Because it was recently reported that activated conventional T cells mainly expressed FOXP3 in the cytoplasm, whereas in Tregs FOXP3 was predominantly expressed in the nucleus (12), we evaluated whether this difference in intracellular FOXP3 localization in iNKT cells could be responsible for the observed functional differences in FOXP3-expressing iNKT cells.
Localization of FOXP3 in activated conventional T cells and Tregs obtained after CD4- and CD25-based Treg isolation was determined by imaging flow cytometry. As described in 2Materials and Methods, a nuclear translocation score was calculated per cell (see Supplemental Fig. 1) based on the amount of FOXP3 in the entire cell versus the Hoechst signal. As shown in Fig. 6A, a low nuclear translocation score corresponds to cytoplasmic localization of FOXP3 and a high score corresponds to colocalization of FOXP3 and the nuclear Hoechst staining, as demonstrated in activated conventional T cells (i.e., CD4+CD25−/int T cells) and Tregs (CD4+CD25hi T cells), respectively. To evaluate the localization of FOXP3 in iNKT cells, cells were stained with FOXP3 and Hoechst and the nuclear translocation score was calculated. Our results indicate that iNKT cells cultured with rapamycin, IL-10 and rapamycin, or TGF-β and rapamycin have a higher nuclear translocation score compared with iNKT cells cultured with medium (control), IL-10, or TGF-β, respectively, demonstrating increased nuclear localization of FOXP3 in the presence of rapamycin (Fig. 6B). Because iNKT cells cultured with rapamycin (either with or without the addition of IL-10 or TGF-β) also showed suppressive functionality, these data strongly suggest that rapamycin induces suppressive capacity in iNKT cells by promoting nuclear localization of FOXP3.
FOXP3 localization in conventional CD4+ T cells, Tregs, and iNKT cells by imaging flow cytometry. Based on the Hoechst signal, a mask delineating the nucleus was made. The nuclear translocation score was estimated based on the log-transformed ratio of the amount of FOXP3 in the entire cell versus the nuclear mask. (A) FOXP3 localization was determined in activated conventional T cells (Tconv) and Tregs, showing a differential subcellular localization in the two cell subtypes. A low nuclear translocation score corresponds to cytoplasmic localization of FOXP3 (activated Tconv cells) and a high score to nuclear localization (Tregs). (B) IL-10 and rapamycin differentially affect intracellular FOXP3 localization in iNKT cells. Culture of iNKT cells in the presence of rapamycin, but not of IL-10 or TGF-β alone, results in nuclear localization of FOXP3. Means ± SEM are shown. Data are representative of two independent experiments.
FOXP3 localization in conventional CD4+ T cells, Tregs, and iNKT cells by imaging flow cytometry. Based on the Hoechst signal, a mask delineating the nucleus was made. The nuclear translocation score was estimated based on the log-transformed ratio of the amount of FOXP3 in the entire cell versus the nuclear mask. (A) FOXP3 localization was determined in activated conventional T cells (Tconv) and Tregs, showing a differential subcellular localization in the two cell subtypes. A low nuclear translocation score corresponds to cytoplasmic localization of FOXP3 (activated Tconv cells) and a high score to nuclear localization (Tregs). (B) IL-10 and rapamycin differentially affect intracellular FOXP3 localization in iNKT cells. Culture of iNKT cells in the presence of rapamycin, but not of IL-10 or TGF-β alone, results in nuclear localization of FOXP3. Means ± SEM are shown. Data are representative of two independent experiments.
Discussion
In this study, we show that both IL-10 and rapamycin promote FOXP3 expression in iNKT cell lines, whereas in contrast to previous reports using ex vivo cells (4–6), stimulation of these lines in the presence of TGF-β did not result in increased levels of FOXP3 in iNKT cells. Complementing recent data (6), we show that a suppressive phenotype was only induced when iNKT cell lines were cultured in the presence of rapamycin, irrespective of the additional presence of IL-10 or TGF-β. This suppressive phenotype was accompanied by a reduced IFN-γ–producing capacity in iNKT cells and nuclear translocation of FOXP3. Of note, rapamycin alone was also sufficient to induce FOXP3 expression and accompanying suppressive activity in ex vivo iNKT cells (Supplemental Fig. 3).
In agreement with a previous study on iNKT cells (6) and several other reports describing the expression of FOXP3 in activated T cells without the acquisition of suppressive functions (9, 10), our data confirm that expression of FOXP3 alone does not reflect the suppressive potential of iNKT cells, but appears to be more indicative of the iNKT cell activation status. We showed that FOXP3 could be induced in iNKT cells upon culture with IL-10 and rapamycin but not upon culture with TGF-β as previously described (6). There are several reasons that may account for this observed discrepancy. First, Moreira-Teixeira et al. (6) used iNKT cells derived from cord blood and freshly isolated PBMC, whereas we used iNKT cell lines. Freshly isolated PBMC, and cord blood cells in particular, have been suggested to be more plastic and perhaps easier to convert into a Treg phenotype with TGF-β compared with iNKT cell lines (6, 22). In this study, we focused on the use of iNKT cell lines, as these are more reflective of the iNKT cell populations eventually used for future in vivo adoptive transfer studies. Second, results might be different due to an apparent difference in the flow cytometric gating strategy that was used to define FOXP3+ iNKT cells. We defined the FOXP3+ iNKT cells as CD25hi and thereby excluded CD25int/lowFOXP3+ iNKT cells, which might be, in analogy to the established gating strategy used for the identification of Tregs (14), activated iNKT cells rather than potentially suppressive iNKT cells. Our present data are in line with results previously reported by our group, showing upregulation of the activation marker CD25 on iNKT cells cultured in the presence of IL-10 and downregulation in the presence of TGF-β (23). Furthermore, only in conditions where rapamycin was present was an effect on the IFN-γ production seen, whereas neither IL-10 nor TGF-β affected IFN-γ production by iNKT cells. The reduced capacity of rapamycin-exposed iNKT cells to produce IFN-γ could contribute to a less proinflammatory activity of these iNKT cells. Additionally, Monteiro et al. (4) also cultured iNKT cells in the presence of anti–IFN-γ mAbs, which may explain the possible induction of FOXP3. However, it was not shown that these FOXP3+ iNKT cells were able to suppress other T cell subsets, so it could rather reflect an activation-related induction of FOXP3.
Tregs have been shown to be able to suppress other immune subsets by both contact-dependent as well as contact-independent mechanisms, exerted by natural and induced Tregs (nTregs and iTregs), respectively (24). Our data show a predominant contact-dependent suppression mechanism of iNKT cells, because they were not able to exert their suppressive effect when cultured with responder T cells in a Transwell system. The cell-bound factor responsible for the observed contact-dependent suppression as yet remains unidentified. Because our results (Fig. 1C) showed comparable CTLA-4 expression in both suppressive and nonsuppressive iNKT cells and no major differences in the expression of LAP between FOXP3− and CD25hiFOXP3+ iNKT cells (Fig. 2A), both markers were not considered to be responsible for the observed cell contact–dependent suppression. Of note, we found granzyme B levels to be significantly increased in iNKT cells cultured with IL-10 (data not shown), whereas levels were decreased in the additional presence of rapamycin. Therefore, because we did not find a correlation between granzyme B expression and suppressive capacities, we consider it also unlikely that granzyme B is responsible for the observed cell contact–dependent suppression. The combined absence of IFN-γ production and the ability to exert contact-dependent T cell suppression points to the acquisition of an nTreg-like state by the mTOR-inhibited iNKT cells rather than an iTreg-like state, which is characterized by simultaneous IL-10 and IFN-γ production and contact-independent T cell suppression (25).
To investigate whether iNKT cells share other phenotypic characteristics with nTregs, we determined the expression of several markers associated with conventional Tregs. We show that the expression of CTLA-4 was induced in iNKT cells upon culture with IL-10 and rapamycin, as seen for FOXP3, similar to the results previously described in nTregs (9). Because LAP is a marker found on the surface of activated nTregs, associated with latent TGF-β, and possibly involved in cell contact–dependent immunosuppression (20, 26), its expression on iNKT cells was determined. We found low percentages of LAP, which could be associated with a more resting status of the iNKT cells, and no major differences between FOXP3− and CD25hiFOXP3+ iNKT cells were observed. However, addition of TGF-β did lead to a significant increase of LAP expression on CD25hiFOXP3+ iNKT cells compared with FOXP3− cells. Nevertheless, iNKT cells cultured in the presence of TGF-β did not acquire suppressive capacity. Helios is a marker shown to be expressed in all immature thymocytes whereas only small numbers of mature T cells still express Helios in the periphery (27). However, nTregs were shown to maintain Helios expression in the periphery (28), and therefore absence of Helios could be used as an indication of the generation of iTregs (29). Our data indicate that addition of rapamycin and the combination of TGF-β and rapamycin result in a significant increase of Helios expression in CD25hiFOXP3+ iNKT cells compared with FOXP3− iNKT cells, which may indicate a selective survival of a specific subset of iNKT cells with possibly preexisting suppressive capacities, consistent with nTregs. Furthermore, Helios expression was significantly decreased in CD25hiFOXP3+ iNKT cells upon culture with the combination of IL-10 and rapamycin compared with CD25hiFOXP3+ iNKT cells cultured in the presence of rapamycin only, indicative of an iTreg conversion rather than nTreg survival. However, these differences were not reflected in the mechanism of suppression because all rapamycin conditions, with or without the addition of IL-10 and TGF-β, showed suppression in a cell contact–dependent mechanism.
Strikingly, we did observe a difference in localization of FOXP3 between activated FOXP3+ iNKT cells and suppressive FOXP3+ iNKT cells by cytoplasmic and nuclear FOXP3 expression, respectively, as previously described in Tregs (12). As nuclear localization was only observed when iNKT cells were exposed to rapamycin, we conclude that mTOR inhibition by rapamycin is responsible for the induction of suppressive capacities in iNKT cells, regardless of the presence of suppressive cytokines such as IL-10 or TGF-β, resulting in the induction of iNKTregs.
iNKT cells can play a diverse role in the tissue microenvironment by rapidly producing stimulatory or inhibitory cytokines (1–3). Additionally, in the tumor microenvironment iNKT cells can play an important role by modulating myeloid cells, thus achieving the abolishment of suppressor activity of myeloid-derived suppressor cells by CD1d- and CD40-dependent interactions (30). Alternatively, as our data also indicate, iNKT cells can have direct suppressive capacities similar to Tregs and myeloid-derived suppressor cells. Highly proliferative tumors can cause amino acid deprivation that can lead to inhibition of mTOR (31), resulting in a possible contribution to a more suppressive tumor microenvironment and leading to conversion of iNKT cells into suppressor cells. This should be taken into consideration when thinking about the use of iNKT cells in the treatment of cancer. Furthermore, mTOR inhibitors are frequently used in the treatment of various types of cancer, for example, metastatic renal cell cancer, advanced breast cancer, and others (32, 33), and it has been shown to result in the expansion of Tregs both in vitro and in vivo (34–36). Because our data indicate that iNKT cells can also be converted into suppressor cells, the putative detrimental immune effects that mTOR inhibitors may have in light of cancer treatment (37) could be more extensive and include iNKT cells. In keeping with this notion, regulatory γδ T cells were also recently shown to be induced upon culture with rapamycin (38).
Taking the immunosuppressive properties of iNKTregs into account, it could be possible to apply iNKTregs in the treatment of immune-mediated inflammatory diseases. The most efficacious therapy for hematological malignancies is allogeneic hematopoietic stem cell transplantation (HSCT). Unfortunately, a major complication of HSCT is graft-versus-host disease (GVHD), induced by the donor-derived T cells. Because it has been shown that iNKT cell numbers can predict the incidence of GVHD after HSCT (39), our data suggest that mTOR inhibitors could have additional value during HSCT or in the treatment of GVHD by inducing a more suppressive microenvironment, possibly in part by the induction of suppressive iNKT cells.
In conclusion, we show that rapamycin induces a contact-dependent suppressive capacity in iNKT cells, regardless of the addition of suppressive cytokines. We show that in this setting FOXP3 is primarily localized in the nuclear compartment of iNKT cells whereas in other cases, where iNKT cells are activated or cultured in the presence of IL-10, cytoplasmic FOXP3 expression predominates. This effect of rapamycin should be taken into account as possibly detrimental in the treatment of cancer patients (e.g., with mTOR inhibitors) and vice versa as possibly beneficial for adoptive transfer of regulatory iNKT cells in the treatment of inflammatory disorders.
Acknowledgements
We thank Prof. Yvette van Kooyk for the use of the Amnis-Millipore ImageStream facility.
Footnotes
This work was supported by The Netherlands Organization for Health Research and Development Grant 90700309 and by Dutch Cancer Society Grant VU 2010-4728.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- α-GalCer
α-galactosylceramide
- GVHD
graft-versus-host disease
- HSCT
hematopoietic stem cell transplantation
- iNKT
invariant NKT
- iNKTreg
regulatory iNKT
- iTreg
induced Treg
- LAP
latency-associated peptide
- moDC
monocyte-derived dendritic cell
- mTOR
mammalian target of rapamycin
- nTreg
natural Treg
- rh
recombinant human
- Treg
regulatory T cell.
References
Disclosures
The authors have no financial conflicts of interest.