The programmed death (PD)-1 molecule and its ligands (PD-L1 and PD-L2), negative regulatory members of the B7 family, play an important role in peripheral tolerance. Previous studies have demonstrated that PD-1 is up-regulated on T cells following TCR-mediated activation; however, little is known regarding PD-1 and Ag-independent, cytokine-induced T cell activation. The common γ-chain (γc) cytokines IL-2, IL-7, IL-15, and IL-21, which play an important role in peripheral T cell expansion and survival, were found to up-regulate PD-1 and, with the exception of IL-21, PD-L1 on purified T cells in vitro. This effect was most prominent on memory T cells. Furthermore, these cytokines induced, indirectly, the expression of PD-L1 and PD-L2 on monocytes/macrophages in PBMC. The in vivo correlate of these observations was confirmed on PBMC isolated from HIV-infected individuals receiving IL-2 immunotherapy. Exposure of γc cytokine pretreated T cells to PD-1 ligand-IgG had no effect on STAT5 activation, T cell proliferation, or survival driven by γc cytokines. However, PD-1 ligand-IgG dramatically inhibited anti-CD3/CD28-driven proliferation and Lck activation. Furthermore, following restimulation with anti-CD3/CD28, cytokine secretion by both γc cytokine and anti-CD3/CD28 pretreated T cells was suppressed. These data suggest that γc cytokine-induced PD-1 does not interfere with cytokine-driven peripheral T cell expansion/survival, but may act to suppress certain effector functions of cytokine-stimulated cells upon TCR engagement, thereby minimizing immune-mediated damage to the host.
Programmed death (PD)-12 and its ligands, PD-L1 (B7-H1) and PD-L2 (B7-DC), have been shown to play an important role in peripheral tolerance (1, 2, 3). Additionally, the PD-1 axis has been implicated in the down-regulation of virus-specific T cell responses in the context of certain chronic viral infections (4, 5). PD-1 ligation has been shown to inhibit TCR-mediated events (1, 6, 7) and has been associated with apoptosis in some (8, 9, 10), but not all, studies (6, 11). PD-1 is up-regulated on lymphocytes following TCR-mediated activation and remains elevated in the context of persistent Ag-specific immune stimulation, whether the immune responses are directed at self or foreign Ags (2, 3, 12, 13). However, little is known regarding PD-1 expression in the context of Ag nonspecific T cell stimulation by cytokines.
The common γ-chain (γc) cytokines IL-2, IL-7, IL-15, and, to a lesser extent, IL-21 play a critical role in the support of T cell proliferation, survival, and function during both immune responses and homeostatic peripheral expansion (HPE). Additionally, several of these cytokines are currently being administered in vivo to improve T cell counts and function in certain individuals with lymphopenia and/or immune dysfunction or in conjunction with vaccines (14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25). A potential complication of this therapeutic modality is the fact that the administration of γc cytokines can be associated with prolonged or excessive peripheral T cell expansion/activation, increasing the possibility for immune-mediated damage to the host (26, 27, 28, 29, 30, 31). In this regard, little is known concerning the mechanisms whereby the host avoids immune-mediated damage during prolonged periods of cytokine-driven T cell activation and expansion.
In the present study, we investigated whether cytokine-mediated immune activation can induce PD-1 and PD-1 ligand expression. Of numerous cytokines tested, only IL-2, IL-7, IL-15, and IL-21 were found to directly induce the expression of PD-1 and PD-L1 on purified T cells as well as PD-1 ligands on APCs. These in vitro data were supported by data obtained on PBMC of HIV-infected individuals receiving in vivo IL-2 immunotherapy. Data regarding the effects of PD-1 ligation in functional assays suggest that PD-1 induced by IL-2, IL-7, and IL-15 does not compromise cytokine-driven events (signaling, proliferation, or survival), but may serve to limit certain T cell effector functions upon TCR engagement and thus interfere with immune-mediated damage.
Materials and Methods
Isolation of PBMC and T cell subsets
Lymphophereses were performed on HIV-negative and HIV-positive subjects under National Institutes of Health-approved protocols (National Institutes of Health 02-I-0202 and 81-I-0164), and PBMC were obtained by Hypaque-Ficoll density centrifugation. Total T cells were isolated using negative immunomagnetic bead selection with an Ab cocktail (Stem Cell Technologies), and in certain experiments CD45RA-negative (memory) subsets were isolated using immunomagnetic beads (Invitrogen/Dynal). PBMC from HIV-positive individuals receiving IL-2 (3–7.5 M IU s.c./bid; 5-day cycle) were collected/frozen on day 1 before the first IL-2 administration and day 5 before the last two doses. All patients received IL-2 while also receiving effective antiretroviral therapy, leading to a suppression of viral load to <50 copies/ml.
Freshly isolated PBMC were cultured at 106/ml in media (RPMI 1640 plus 1 mM HEPES buffer and penicillin/streptomycin with 10% human AB serum) alone or with IL-1β (5–100 ng/ml), IL-2 (5–200 IU/ml), IL-4 (5–100 ng/ml), IL-6 (5–100 ng/ml), IL-7 (5–100 ng/ml), IL-8 (5–100 ng/ml), IL-10 (5–100 ng/ml), IL-12 (25–200 ng/ml), IL-15 (5–100 ng/ml), IL-18 (5–100 ng/ml), IL-21 (5–200 ng/ml), TGF-β (5–100 ng/ml), IFN-γ (5–100 ng/ml), TNF-α (2–50 ng/ml), or anti-human CD3 (5 μg/ml) plus anti-human CD28 (2 μg/ml) (BD Pharmingen). All cytokines were purchased from the National Cancer Institute Biological Resource Branch Preclinical Repository (Frederick, MD) or R&D Systems. Cells were monitored for PD-1 expression during a 10–16-day period. For all subsequent analyses, unless otherwise indicated, PBMC, purified total T cells, or CD45RA− (memory) T cells were cultured in media plus IL-2 (200 IU/ml), IL-7 (25 ng/ml), IL-15 (100 ng/ml), IL-21 (100 ng/ml), or anti-CD3/CD28 (as above).
PD-1 and PD-L expression and naive/memory subset phenotypes were assessed using anti-human PD-1 (clone MIH4), PD-L1 (clone MIH1; eBioscience), or PD-L2 (clone MIH18; eBioscience), CD45RA, CD62L, CD3, CD4, and CD8. PBMC were stained for CD3, CD19, CD14, and PD-L1 or PD-L2. Frozen PBMC obtained from HIV-positive subjects receiving IL-2 in vivo were thawed (all time points per patient were assayed at the same time) and stained for PD-1, PD-L1, PD-L2, CD3, CD4, CD8, CD19, CD14, Ki67, and HLA-DR. All mAbs were purchased from BD Pharmingen unless otherwise indicated.
STAT5 and Lck activation
Purified human T cells were prestimulated with IL-2 or anti-CD3/CD28 for 6 days. Harvested cells were restimulated with IL-2 plus tosylactivated beads (Invitrogen/Dynal) coated with hIgG or PD-L2 hIgG (2 μg/107 beads) or with beads coated with anti-CD3 (4 μg/107 beads)/anti-CD28 (2 μg/107 beads) or isotype plus PD-L2 or hIgG (3:1 bead-to-cell ratio). Cells were stimulated for 15 (Lck) or 45 (STAT5) min at 37°C and then stained with phospho-specific Y505 Lck (BD Pharmingen, clone 4/LCK-Y505), phospho-specific Y694 STAT-5 (BD Pharmingen clone 47), PD-1 (BD Pharmingen, clone MIH4), and CD8 (BD Pharmingen) using eBioscience’s Fix/Perm solution per the manufacturer’s instructions.
Freshly isolated, purified T cells, or PBMC (candida) were plated (105/well) in 96-well plates either 1) coated with isotype hIgG or hPD-L1-IgG or hPD-L2-IgG chimeric proteins (5 μg/ml, R&D Systems) or 2) in the presence of neutralizing anti-PD-L1 Ab (5 μg/ml; eBioscience). Cells were then stimulated with plate-bound anti-CD3 (0.2 μg/ml)/anti-CD28 (0.4 μg/ml), Candida albicans (5 μg/ml), or with γc cytokines. Proliferation was assessed by 16-h [3H]thymidine uptake at day 6 poststimulation.
Cytokine production and apoptosis.
Purified T cells received an initial stimulation with γc cytokines or anti-CD3/CD28 (5/2 μg/ml) for 5–8 days. Cells were washed, restimulated with plate-bound anti-CD3/CD28 (0.2/0.4 μg/ml) plus hIgG, PD-L1-hIgG, or PD-L2-hIgG (5 μg/ml) and tested in parallel for annexin V or cytokine production. For cytokine production, brefeldin A (GolgiPlug; BD Pharmingen) was added 2 h after restimulation, and 4 h later cells were stained for PD-1, CD8, and intracellular IL-2 and IFN-γ as per the manufacturer’s recommendations. Apoptosis was assessed at 6 h after restimulation by staining for PD-1, annexin V (BD Pharmingen), CD4, and CD8. In certain apoptosis experiments, prestimulated T cells were exposed to PD-L1 or PD-L2 IgG in the absence of anti-CD3; however, results are not shown but were similar to those reported below.
Paired Student’s t test and Spearman’s nonparametric correlation were used for statistical analyses.
In vitro cytokine-mediated induction of PD-1 expression
Induction of PD-1 expression has been previously described only under conditions of TCR stimulation. In the present study, we investigated whether cytokine-mediated immune activation can contribute to PD-1 expression. PBMC were stimulated with a variety of cytokines and T cells were monitored for PD-1 expression. Of those cytokines tested, IL-2, IL-7, IL-12, IL-15, and IL-21 induced significant levels of PD-1 expression above background (untreated cells), whereas proinflammatory (IL-1β and TNF-α; IL-6, IL-8 data not shown), immunosuppressive (TGF-β; IL-10, data not shown), and certain immunoregulatory (IL-4 and IFN-γ; IL-18, data not shown) cytokines had no significant effect (Fig. 1,A). While IL-12 induced PD-1 on CD8+ T cells in the context of PBMC (Fig. 1,A and supplemental Fig. 1),3 only the γc cytokines IL-2, IL-7, IL-15, and IL-21 were found to directly induce PD-1 expression on purified T cells (Fig. 1, B–E). γc cytokines increased the frequency of PD-1+ cells (Fig. 1,C) and modestly increased the PD-1 mean fluorescence intensity (MFI) (Fig. 1,D) on both CD4+ (IL-21 had no effect on MFI) and CD8+ T cells. A significant difference in cytokine-induced PD-1 expression between CD4+ and CD8+ T cells was seen only with IL-21 treatment (Fig. 1, C and D). γc cytokines induced PD-1 expression on purified T cells more slowly than did anti-CD3/CD28 stimulation, but effects were long lasting and increased over time, with the exception of IL-21 (cells did not survive beyond 6–8 days) (Fig. 1,E, similar results were obtained for CD4+ T cells). γc cytokine treatment of purified T cells increased the frequency of PD-1+ CD8+ and CD8− (CD4+) cells in the central (TCM, CD45RA−CD62L+) and effector (TEM, C45RA−CD62L−) memory subsets; PD-1+ effector cell (TEff, CD45RA+CD62L−) frequencies were also elevated, but few events were present in this gate (Fig. 2, A and B). While there were significant differences in the percentages of PD-1+ CD8+ vs CD4+ T cells, there were no significant differences in the response (fold induction vs untreated) to any stimulation condition in any subset. The frequency of PD-1high TEM cells was increased by IL-2, IL-15, and IL-21, whereas anti-CD3/CD28 stimulation induced PD-1high expression in all T cell subsets (supplemental Fig. 2A). Of note, the TEff population had a higher frequency of PD-1high cells than did other subsets under both untreated and cytokine-treated conditions (supplemental Fig. 2A). Moderate cytokine concentrations (6 ng/ml) were effective in up-regulating PD-1 on purified memory (CD45RA−) T cells; however, IL-2 and IL-15 were most effective at high concentrations (Fig. 2 C).
In vitro cytokine-mediated induction of PD-1 ligand expression
The functional consequences of PD-1 expression on T cells clearly rely on the interaction of these cells with PD-1 ligand-expressing cells. PD-L2 (B7-DC) expression is largely restricted to dendritic cells and activated macrophages, whereas PD-L1 (B7-H1) is expressed more promiscuously and can be found on a variety of hematopoietic and nonhematopoietic cells, particularly in the tissue (32, 33). While PD-L1 and PD-L2 have been shown to be up-regulated on both professional and nonprofessional APC upon stimulation with TLR ligands and certain cytokines (32, 33), such as IFN-γ, the effects of γc cytokines on PD-1 ligand expression have not been reported. In the context of total PBMC, γc cytokines increased the frequency and, with the exception of IL-21, the MFI of PD-L1+ monocytes/macrophages (M/M) (Fig. 3, B and C). CD3+ PBMC (T cell) expression of PD-L1 (MFI; Fig. 3,C) and frequency (data not shown) were modestly, but significantly, elevated by all γc cytokines except IL-21. However, PD-L1 expression (MFI) on T cells remained significantly lower than on M/M under all conditions (Fig. 3,C). The frequency, but not the MFI (data not shown), of PD-L2+ M/M was increased by IL-2, IL-15, and IL-21 (Fig. 3,B); no PD-L2 was observed on T or B cells (data not shown). All γc cytokines, with the exception of IL-21, induced PD-L1 expression on purified T cells, and IL-21 was the only γc cytokine to up-regulate PD-L1 on B (CD19+) cells in PBMC (Fig. 3 D). Of interest, treatment of purified T cells with IL-2, IL-7, and IL-15 increased the frequency of PD-L1 on all memory/effector and naive T cell subsets (supplemental Fig. 2B).
In vivo IL-2-mediated induction of PD-1 and PD-1 ligand expression
To determine whether cytokines can modulate the PD-1 axis in vivo, PD-1 and PD-L expression was assessed on PBMC isolated from HIV-positive subjects (n = 5; HIV viral load <50) before and during (day 5) administration of IL-2 immunotherapy. Following IL-2 administration, the frequency of activated (HLA-DR+) PD-1+ CD4+ and CD8+ T cells, and the MFI of PD-1+ CD8+ T cells, was significantly elevated (Fig. 4, A and B). Furthermore, the frequency of M/M and T cells expressing PD-L1 (Fig. 4,C, left) and of M/M expressing PD-L2 (Fig. 4,C, right) was also increased. Of note, IL-2 therapy was associated with elevated frequencies of PD-1+ T cells expressing the proliferation Ag Ki67 (Fig. 4 D), suggesting that IL-2-mediated induction of PD-1 or its ligands did not interfere with the ability of PD-1+ cells to proliferate in vivo in response to this cytokine.
Functional consequences of γc cytokine-mediated modulation of the PD-1 axis
Several of these γc cytokines are currently being used, or are being considered, as modalities for immunotherapy (16, 23, 25) or as adjuvants in the context of vaccination (34, 35, 36). It is therefore important to determine the potential role of the PD-1 axis in modulating the activity and/or survival of γc cytokine-stimulated lymphocytes. Engagement of PD-1 has been shown to inhibit TCR-stimulated T cell proliferation and effector function (1, 6, 37, 38) and has also been associated with increased apoptosis (8, 9, 10); any of these effects could have negative consequences on cytokine-mediated T cell expansion and survival.
Proliferation was assessed in purified T cells stimulated with γc cytokines or TCR stimuli (anti-CD3/CD28 and recall Ag (C. albicans in PBMC)) in the presence of hIgG control, PD-L1-hIgG, or PD-L2 hIgG (Fig. 5,A) or in the presence of neutralizing anti-PD-L1 Ab (Fig. 5,B). IL-2, IL-7, and IL-15 induced significant proliferation of purified T cells; IL-21-treated cells did not survive well at day 6 and thus did not incorporate thymidine (Fig. 5,A). The presence of either PD-L1 IgG or PD-L2 IgG at the time of initial stimulation significantly suppressed anti-CD3/CD28, and the presence of PD-L1 IgG suppressed Candida-stimulated proliferation; however, neither PD-L1 nor PD-L2 IgG significantly reduced γc cytokine-induced proliferation (Fig. 5,A). In addition, we investigated the role of endogenous T cell-associated PD-L1 in proliferation by conducting assays using purified T cells in the presence of neutralizing anti-PD-L1 Ab. Anti-PD-L1 Ab had no significant effect on the ability of purified T cells to proliferate in response to either immobilized anti-CD3/CD28 or γc cytokines (Fig. 5 B), suggesting that endogenous T cell-associated PD-L1 does not significantly inhibit proliferation under these in vitro conditions.
To verify that PD-1 engagement does not inhibit γc cytokine-induced intracellular signaling events, STAT5 activation (γc cytokine signaling; IL-2) and Lck activation (TCR complex-signaling; anti-CD3-CD28) was determined in the presence PD-L2 IgG. Purified T cells were pre-stimulated with IL-2 or anti-CD3/CD28 (6 days) to induce PD-1 expression. Cells were then exposed to hIgG or PD-L2 IgG at the time of secondary stimulation with IL-2 or anti-CD3/CD28 and stained for active (phosphorylated) Lck or STAT5. Whether PD-1 was initially induced by IL-2 or by anti-CD3/CD28 pre-treatment, engagement of PD-1 by PD-L2 IgG at the time of secondary stimulation inhibited signaling mediated by the TCR complex (Fig. 5,C, left; Lck) but not by IL-2 (Fig. 5 C, right; STAT5).
Having determined that PD-1 engagement does not directly interfere with γc cytokine-mediated proliferation or signaling events, the effects of PD-1 ligation on the survival and function γc cytokine and anti-CD3/CD28 pretreated T cells following restimulation via TCR was evaluated. Purified γc cytokine and anti-CD3/CD28 prestimulated T cells were harvested at time points of strong PD-1 expression (6–8 days poststimulation), washed, and then restimulated with anti-CD3/CD28 in the presence of hIgG (control), PD-L1 IgG, or PD-L2 IgG. Cells were then tested in parallel for cytokine production (intracellular cytokine (ICC)) and apoptosis (annexin V) after 5–6 h.
γc cytokine pretreated T cells produced high levels of IL-2 and IFN-γ following restimulation with anti-CD3/CD28 in the absence of PD-1 ligands (hIgG control) (Fig. 6). The frequency of IL-2+, IFN-γ+, and IL-2+/IFN-γ+ T cells following restimulation in the presence of hIgG was significantly higher in γc cytokine pretreated cells compared with cells not receiving prior stimulation and with anti-CD3/CD28 prestimulated cells (p = 0.04–0.0004; not shown) (Fig. 6,B). However, the presence of either PD-L1 or PD-L2 IgG during restimulation dramatically reduced the frequency of IFN-γ+ (Fig. 6,B, left), IL-2+/IFN-γ+ (Fig. 6,B, middle), and IL-2+ (Fig. 6 B, right) populations in T cells prestimulated with either γc cytokines or anti-CD3/CD28. There were no significant differences between PD-L1 and PD-L2 IgG for this effect.
PD-1 engagement has been described to induce apoptosis in certain studies, while other studies find PD-1 to simply be a marker for susceptibility to apoptosis (6, 8, 9, 10, 11). To investigate the impact of cytokine-induced PD-1 on apoptosis, pretreated (as above) T cells were washed, restimulated with anti-CD3/CD28 in the presence of hIgG, PD-L1, or PD-L2 IgG, and apoptosis (surface annexin V+) was assessed in the PD-1high (MFI > 100), PD-1mid (80 > MFI > 20) and PD-1− (MFI ≤ 10) T cell populations (Fig. 7,A). When restimulated with anti-CD3/CD28 in the absence of the PD-1 ligand (control hIgG), the frequency of apoptosis differed significantly based on the level of PD-1 expression (PD-1high > PD-1mid > PD-1− subsets) under all conditions tested (Fig. 7,B); similar results were observed in cells not receiving a secondary TCR stimulation (data not shown). γc cytokine pretreatment resulted in a small, but significant, reduction in apoptosis within the PD-1mid population (Fig. 7,C, left) compared with anti-CD3/CD28 pretreatment or no prior treatment; only IL-7 reduced apoptosis in the PD-1high (Fig. 7,C, right) T cell population. Of note, the reduction of apoptosis in PD-1+ populations by cytokine pretreatment was maintained even in the presence of PD-L1 or PD-L2 IgG (Fig. 7 C).
γc cytokines play an important role in peripheral T cell expansion, function, and survival, and therefore several of these cytokines are currently being administered in vivo as immunotherapy in certain diseases (16, 25), as well as adjuvants in vaccination. However, γc cytokine administration and HPE under lymphopenic conditions can be associated with prolonged or excessive peripheral T cell expansion/activation, increasing the possibility for immune-mediated damage to the host (26, 27, 28, 29, 30, 31). The PD-1 axis has been shown to play an important role in controlling the potentially harmful activity of peripheral T cells that are reactive to self Ags and persistent foreign Ags (2, 6, 13); however, the role of PD-1 in maintaining tolerance during cytokine-driven T cell activation/expansion has not been investigated. The present study demonstrates that the γc cytokines IL-2, IL-7, IL-15, and IL-21 directly induce the PD-1 axis in purified T cells and PBMC of healthy subjects in vitro and in PBMC of HIV-infected subjects receiving IL-2 immunotherapy in vivo. The results of in vitro functional assays suggest that engagement of γc cytokine-induced PD-1 does not suppress γc cytokine-mediated signal transduction, T cell expansion, or enhancement of function and survival, but does significantly inhibit certain effector functions (cytokine production) of cytokine-stimulated T cells upon subsequent TCR triggering.
This is the first report, to our knowledge, of γc cytokine-mediated up-regulation of the PD-1 axis. Although the list of cytokines screened in the present study for the ability to modulate PD-1 was not exhaustive, the ability to induce PD-1 expression on purified T cells was limited to certain γc cytokines (IL-2, IL-7, IL-15, and IL-21; Fig. 1). While all these cytokines play a role in promoting T cell expansion, survival, or function during normal immune responses, IL-7 and IL-15 play a critical role in HPE (20, 22, 27, 39, 40, 41, 42, 43, 44, 45, 46, 47). The induction of PD-1 expression in vitro by γc cytokines was restricted to memory (TEM and TCM) and effector T cells (Fig. 2). In this regard, it has been reported that memory, but not naive, T cell populations can undergo expansion that is cytokine driven and TCR-MHC-independent, at least in mice (39, 40, 45). Of interest, naive T cell subsets did respond to IL-2, IL-7, and IL-15, as did memory/effector subsets, by up-regulating PD-L1 (supplemental Fig. 2B). IL-2, IL-7, and IL-15 increased both the frequency and, to a lesser extent, the MFI of PD-L1+ T cells; however, PD-L1 remained at much lower levels (MFI) than those seen on M/M (Fig. 3). We found no significant inhibitory role of T cell-associated PD-L1 in purified T cell cultures (Fig. 5); however, in our in vitro conditions PD-L1+ T cells are not presenting Ag. In studies demonstrating suppression via PD-L1-B7.1 (48, 49) or PD-L1-PD-1 (1, 6, 7), TCR is co-engaged either by beads or alloreactive T cells. Only IL-21 significantly induced PD-L1 on B cells (Fig. 3), consistent with its effect on B cell activation/differentiation (44, 50). PD-L1 and, to a lesser extent, PD-L2 expression on M/M in PBMC in vitro (Fig. 3) and in vivo (HIV-positive subjects receiving IL-2 immunotherapy; Fig. 4) was dramatically up-regulated by γc cytokines. γc cytokines did not induce PD-1 ligand expression on purified M/M in vitro, and the effect on M/M in PBMC was found to be largely IFN-γ-dependent (data not shown).
The lack of an inhibitory effect of PD-1 engagement on γc cytokine-mediated signaling or downstream functional events in vitro (Fig. 5) is not surprising based on our understanding of PD-1-mediated immunosuppression. Our data are consistent with previous studies demonstrating that the inhibitory effects of PD-1 engagement may require TCR-triggered SHP-2 recruitment to the cytoplasmic tail of PD-1 and that PD-1 ligation by itself does not generate a suppressive signal (7). Furthermore, PD-1 inhibitory effects appear to require that the TCR and PD-1 signals be delivered in both temporal and physical proximity (7). Our observation that PD-L1 Ig or anti-PD-L1 neutralizing Ab does not alter cytokine-driven proliferation data is supported by the observation of increased proliferation (Ki67 expression) in PD-1+ T cells following in vivo administration of IL-2 to HIV-infected subjects (Fig. 4,D). Of interest, IL-7 and IL-15 have previously been found to rescue proliferation of T cells stimulated with anti-CD3 in the presence of PD-L1 (51), suggesting that PD-1 engagement does not inhibit proliferation driven by these cytokines even if TCR triggering is contributing to the expansion of T cells in vivo. In contrast, γc cytokine pretreatment did render TCR-triggered effector function (measured by cytokine (IL-2/IFN-γ) secretion; Fig. 6) of T cells sensitive to PD-1-mediated suppression. The ability of PD-1 engagement to induce apoptosis remains controversial (6, 8, 9, 11). Our data suggest that PD-1 does not mediate apoptosis, even in the context of TCR signaling, but rather that the level of PD-1 expression (MFI) correlates with susceptibility to apoptosis (Fig. 7). Consistent with the pro-survival effects of these γc cytokines (21, 39, 40, 41, 52), pretreatment of T cells with γc cytokines, particularly IL-7, moderately reduced apoptosis within the PD-1+ cell populations and this effect was maintained even in the presence of PD-1 ligands. Of note, the protective effects of cytokines on cell death seen in this study were less than those reported by other studies in which cytokines were present during stimulation/apoptosis assessment.
Whether γc cytokine-induced PD-1 influences T cell function upon subsequent TCR triggering is particularly relevant in the setting of cytokines administered as immunotherapy or as vaccine adjuvants (14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25). While modest (6 ng/ml) concentrations of γc cytokines induced PD-1 expression on memory T cells in vitro (Fig. 2), normal in vivo levels of these cytokines may be insufficient, except perhaps in tissue microenvironments, to induce the PD-1 axis at a functionally relevant level. Furthermore, note that T cell restimulations were conducted using low concentrations of anti-CD3 and anti-CD28 in the presence of considerable concentrations of PD-1 ligands. These in vitro conditions may reflect conditions existing during the interaction between T cells and nonprofessional APC expressing high levels of PD-1 ligands. Thus, our data need not conflict with the use of γc cytokines as adjuvants in the context of vaccination or as immunotherapy to enhance Ag-specific T cell responses (14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25). However, these data may provide some insight into the optimal timing of cytokine administration (and thus PD-1 axis induction) relative to TCR triggering (vaccines or persistent pathogens whose replication can be modulated by treatment) (53, 54). Furthermore, the inability to detect an improvement of T cell function following γc cytokine administration may be influenced by the timing, relative to cytokine administration, of when cells are reexposed to Ag and assayed for function (53, 54).
In summary, the present study provides evidence that the γc cytokines IL-2, IL-7, IL-15, and IL-21, known to play an important role in peripheral T cell activation and/or expansion and survival, up-regulate both PD-1 and its ligands in vitro and in vivo. Of note, the expression of PD-1/PD-1 ligand does not appear to negatively impact the ability of T cells to expand, function, or survive in response to further γc cytokine exposure, but does render them susceptible to PD-1 ligand-mediated suppression of TCR-triggered function. It remains to be established whether induction of the PD-1 axis by any of these γc cytokines plays a role in maintaining tolerance in vivo under conditions of significant cytokine-driven T cell activation/expansion, such as during immunotherapy and, in some cases, HPE in lymphopenic individuals.
The authors have no financial conflicts of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Abbreviations used in this paper: PD, programmed death; γc, γ-chain; HPE, homeostatic peripheral expansion; MFI, mean fluorescence intensity; M/M, monocytes/macrophages; PD-L, ligand of PD; TCM, central memory T; TEff, effector T; TEM, effector memory T.
The online version of this article contains supplemental material.