Objective: Interaction of ICOS with its ligand (ICOSL, B7-H2) promotes T cell responses. As CD4+CD25highFoxp3+ naturally occurring T regulatory cells in melanoma patients express ICOS, we investigated the impact of ICOS on naturally occurring T regulatory cell function. Methods: Expression of ICOS and T regulatory (Treg) cell markers was determined on CD4+CD25high T cells in PBMC and tumor-infiltrating lymphocytes from melanoma patients (n = 10) and PBMC of normal controls (n = 10) by multicolor flow cytometry. Suppression mediated by sorted ICOShigh and ICOSlow Treg was assessed in CFSE-based suppression assays with autologous CD4+CD25 responder cells (RC). Transwell inserts separating Treg from RC were used to evaluate suppression mechanisms used by Treg. ICOShigh or ICOSlow Treg were coincubated with RC ± TCR and IL-2 stimulation. ICOShigh and ICOS Treg were also expanded under conditions previously shown to induce Tr1 from RC. Results: Treg in tumor-infiltrating lymphocytes expressed ICOS (mean fluorescence intensity = 70 ± 10), while Treg in PBMC had low ICOS expression (mean fluorescence intensity = 3.5 ± 2.5, p ≤ 0.001). ICOShigh Treg up-regulated Treg markers (p ≤ 0.0016) and mediated stronger suppression (p ≤ 0.001) relative to ICOSlow Treg. ICOShigh Treg induced Tr1 cells in nonactivated RC and Th2 cells in preactivated RC. ICOShigh Treg exposed to Tr1 cytokines expressed IL-10 and suppressed RC (92 ± 12%) in contrast to ICOSlow Treg, which mediated low suppression (21 ± 15%; p ≤ 0.0028). Conclusion: ICOShigh Treg can induce diverse immune responses in RC, depending on activation signals and cytokines present. ICOShigh Treg induce Tr1 or Th2 cells depending on the activation state of RC. In a “Tr1” cytokine milieu, ICOShigh Treg transit to Tr1.

The activation and expansion of immune cells is tightly regulated by positive and negative signals that allow for the fine tuning of immune responses. The B7 family ligands and their receptors play a critical role in activation, expansion, as well as contraction of T cell populations. During T cell activation, interaction of B7-1 (CD80) and B7-2 (CD86) with CD28 promotes proliferation, cytokine production, and survival of T cells (1). Additionally, interactions of the ICOS ligand (ICOSL, B7-H2)4 with the TCR-inducible costimulatory receptor (ICOS) enhances effector T cell responses (2).

CD4+CD25highFoxp3+ naturally occurring T regulatory (nTreg) cells have emerged in recent years as critical players in the contraction of activated CD4+ and CD8+ T effector and memory T cells. The absence of nTreg induces severe autoimmunity in mice (3, 4). In contrast, evidence has accumulated demonstrating that the frequency and function of nTreg are increased in the peripheral circulation and in the tumor of patients with cancer (5, 6, 7). nTreg are capable of suppressing proliferation of autologous tumor Ag (TA)-specific as well as Ag-nonspecific CD4+ and CD8+ T cell responses (8). Importantly, nTreg isolated from PBMC or tumors of patients with cancer express higher levels of surface-associated activation molecules and display stronger suppressor function compared with nTreg from normal donors (7, 9). It appears that Treg in tumor-infiltrating lymphocytes (TIL) can mediate suppression of autologous immune cells via distinct mechanisms involving either cell-to-cell contact or inhibitory cytokines, TGF-β and IL-10 (7). In contrast, suppression mediated by Treg in PBMC is exclusively regulated via direct cell-to-cell contact (9). These results suggest that nTreg in patients with cancer comprise highly activated, differentiated, and cytokine-secreting T cell subsets, whereas nTreg in normal donors are largely nonactivated precursor cells. Furthermore, we and others have shown that signals delivered via the TCR receptor and IL-2R are necessary for nTreg to become active suppressor cells (10, 11). Also, we have previously shown that only “hyperactivated” nTreg obtained from cancer patients induce apoptosis in CD4+ and CD8+ responder cells (RC) via Fas/FasL and Granzyme B-mediated pathways (12).

These earlier observations on the activation state of Treg raised the question about the role of ICOS expressed on nTreg. This receptor is exclusively expressed on TCR-activated T cells and is known to induce secretion of IL-10 and TGF-β (Th2 responses) in human non-Treg cells (2). In cancer patients, nTreg are continuously exposed to strong activation-induced signals originating in the tumor microenvironment or directly mediated by TA. Hence, these cells might be expected to express ICOS. In fact, Miller and colleagues (13) have recently shown that tumor-infiltrating CD4+CD25+ Treg accumulating in prostate tumors express higher levels of ICOS than autologous Treg in the peripheral blood. Nevertheless, the role of ICOS expressed on Treg in patients with cancer remains undefined, as are mechanisms ICOS uses to modulate Treg function. The possibility exists that ICOShigh Treg could cross-link CD4+ effector cells, but the impact of such an interaction on anti-tumor immune responses has not been analyzed so far. Furthermore, a relationship between the ICOShigh Treg subset and the Tr1 regulatory T cell subset comprising cells that express high levels of IL-10 has not been studied. We have recently demonstrated by immunohistochemistry that in human tumor sections, CD4+CD25IL-10+ T cells (Tr1) cluster around CD4+CD25highFoxp3+ nTreg, suggesting that nTreg might attract, activate, or induce Tr1 cells (7). However, the possible role of ICOS+ nTreg in inducing IL-10 or otherwise triggering differentiation of Tr1 cells has not been investigated.

In this study, we demonstrate that ICOS is selectively expressed on strongly activated CD4+CD25high nTreg expressing Foxp3, IL-10, and TGF-β and endowed with potent suppressor function. This “hyperactivated” nTreg subset is exclusively present in the tumor and is not detected in the peripheral circulation of melanoma patients or normal controls (NC), indicating that the tumor attracts or induces nTreg precursor cells and promotes their differentiation to ICOShigh active suppressor cells. We also show for the first time that ICOShigh nTreg are able to differentiate into Tr1 cells, when they are exposed to a cytokine milieu favorable to Tr1. ICOShigh nTreg not only have the capacity to “transit” to Tr1 cells, but they also can induce Tr1 cells from autologous CD4+CD25 RC. We further demonstrate that ICOShigh nTreg are only capable of inducing Tr1 cells in RC that are not fully activated. In contrast, CD4+CD25 T cells that are prestimulated via the TCR and IL-2 express a Th2 cytokine profile after interacting with ICOShigh nTreg and do not mediate suppression.

Ten patients diagnosed with metastatic melanoma, who were seen at the University of Pittsburgh Cancer Institute outpatient clinic between November 2001 and January 2006, were included in this study. Tumor tissues and autologous PBMC were obtained from patients with stage III-IV melanoma with in-transit or distant metastases who underwent surgery at the University of Pittsburgh Medical Center. Also, PBMC from 10 healthy volunteers were used in this study as NC. In all experiments, each patient was always tested in parallel with a NC. Patient and control specimens were processed and handled in the same way. All subjects signed an informed consent approved by the Institutional Review Board at the University of Pittsburgh.

TIL were isolated at the University of Pittsburgh Cancer Institute Tissue Procurement Facility according to a standard operating procedure, as previously described (14). Briefly, after fat, blood, or necrotic areas were removed, primary solid human tumor tissue were washed in RPMI 1640, cut into 1 mm3 pieces in a petri dish covered with RPMI 1640, washed again with RPMI 1640, and, upon transfer to flasks, dissociated using 0.05% collagenase (type IV) and 0.02% DNase (type I) (Sigma-Aldrich) in RPMI 1640 supplemented with 5% (v/v) FCS and antibiotics (Invitrogen Life Technologies). Tissues were dissociated for up to 4 h using a magnetic stirrer at 37°C. The digest was then passed through 90-μm and 50-μm nylon mesh to remove clumps, and the filtrate was washed 2–3 times in medium followed by centrifugation at 350 × g for 10 min. To separate tumor cells from lymphocytes, the cell suspension was layered onto a discontinuous gradient of 75% over 100% Ficoll-Hypaque (GE Healthcare Bio-Sciences) in medium and centrifuged for 800 × g for 20 min at room temperature. Lymphocytes were collected from the interphase between 75% and 100% Ficoll-Hypaque and washed twice before further use.

Venous blood was obtained from patients and controls (20 ml) in the morning and collected in heparinized tubes. Blood samples were hand-carried to the laboratory and immediately processed by Ficoll-Hypaque gradient centrifugation. PBMC were recovered from the gradient interface, washed in AIM-V medium (Invitrogen Life Technologies), counted in the presence of a trypan blue dye to evaluate viability, and either immediately used for experiments or cryopreserved using a rate-control process (Cryo-Med). Cell vials were stored in liquid N2 vapors at −80°C.

PBMC or TIL before and after culture were phenotyped as previously described (13). Briefly, cell surface molecules were detected using anti-CD3-ECD, anti-CD4-PC5, anti-glucocorticoid-induced TNF receptor (GITR)-FITC (clone FAB689F; R&D Systems), anti-Fas-FITC, anti-CD127-FITC, anti-CCR7-FITC, anti-CD62L-FITC, anti-CD45RO-FITC, anti-CD45RA-FITC, anti-CD25-FITC and -PE, anti-CD122-FITC, anti-CD132-PE, anti-FasL-PE (NOK-1,42 kDa; BD Pharmingen), and anti-ICOS-PE Abs. Isotype controls IgG1, IgG2a, and IgG2b were included in all experiments. For intracellular staining, samples were first incubated with mAbs against surface markers CD3, CD4, and CD25. After extensive washing, cells were fixed with 4% (v/v) formaldehyde in PBS for 20 min at room temperature, washed once with PBS containing 0.5% (v/v), BSA (w/v), and 2 nM EDTA, permeabilized with PBS containing 0.5% BSA and 0.1% (v/v) saponin, and stained with anti-Perforin-FITC, anti-CD152-PE, anti-GranzymeA-PE, anti-GranzymeB-PE, anti-IL-10-PE or its isotype control PE rat IgG2a,κ (BD Pharmingen), anti-Foxp3-FITC (clone PCH101), and anti-IL-4-FITC from eBioscience and anti-TGF-β1-FITC (clone TB21) from Antigenix America. Intracytoplasmic expression of TGF-β1 and IL-10 was assessed before and after stimulation of PBMC or TIL for 12 h with PMA (20 μg/ml) and ionomycin (1 μM/ml) (American Type Culture Collection) and for 4 h with Brefeldin A (1 μg/ml) (Sigma-Aldrich). We also determined the expression of surface-bound TGF-β1 (sTGF-β1) in nonsaponized cells. Appropriate isotype controls were used in all experiments. Before use, all mAbs were titrated using normal resting or activated PBMC to establish optimal staining dilutions. Abs and their respective isotype controls were all purchased from Beckman Coulter, if not otherwise indicated.

Flow cytometry was performed using a FACScan flow cytometer (Beckman Coulter) equipped with Expo32 software (Beckman Coulter). The acquisition and analysis gates were restricted to the lymphocyte gate as determined by characteristic forward and side scatter properties of lymphocytes. Forward and side scatters were set in a linear scale (Fig. 1,A). For analysis, at least 1 × 105 lymphocytes were acquired. The analysis gates were restricted to the CD3+CD4+, CD4+CD25+, and CD4+CD25high T cell subsets, as previously described (Fig. 1,A). In addition to the above-mentioned gates, we gated on the CD4+CD25highICOShigh, CD4+CD25highICOSint, and CD4+CD25highICOSlow cells within the CD4+CD25high gate (see Fig. 1 A). Cells expressing Treg markers and/or “death” molecules were acquired and analyzed in the FL1 or FL2 logarithmic scale using the set gates.

FIGURE 1.

ICOS expression on CD4+CD25high Treg in TIL and PBMC obtained from melanoma patients or NC. A, The gating strategy used to identify the CD4+CD25, CD4+CD25high, and the CD4+CD25highICOShigh or CD4+CD25highICOSlow T cell subsets. B, Average ICOS expression (MFI ± SD) on CD4+CD25high T cells in PBMC obtained from NC and TIL or PBMC obtained from melanoma patients (left). Flow dot blots for ICOS expression on CD4+CD25high Treg in PBMC of one representative NC and in TIL of one representative melanoma patient (right). C, Percentages of T cells expressing low, intermediate, or high levels of ICOS (left) and MFI of ICOS expression (right) within the CD4+CD25high T cell populations obtained from PBMC (gray bars) or TIL (hatched bars). The data are mean values ± SD from experiments performed with cells of 10 NC and 10 melanoma patients.

FIGURE 1.

ICOS expression on CD4+CD25high Treg in TIL and PBMC obtained from melanoma patients or NC. A, The gating strategy used to identify the CD4+CD25, CD4+CD25high, and the CD4+CD25highICOShigh or CD4+CD25highICOSlow T cell subsets. B, Average ICOS expression (MFI ± SD) on CD4+CD25high T cells in PBMC obtained from NC and TIL or PBMC obtained from melanoma patients (left). Flow dot blots for ICOS expression on CD4+CD25high Treg in PBMC of one representative NC and in TIL of one representative melanoma patient (right). C, Percentages of T cells expressing low, intermediate, or high levels of ICOS (left) and MFI of ICOS expression (right) within the CD4+CD25high T cell populations obtained from PBMC (gray bars) or TIL (hatched bars). The data are mean values ± SD from experiments performed with cells of 10 NC and 10 melanoma patients.

Close modal

For this analysis, CD4+ T cells with a mean fluorescence intensity (MFI) of CD25 expression ≥120 were classified as CD25high. We determined this cut off based on our previous data (15) demonstrating that only CD4+CD25high T cells with this MFI or higher in PBMC of NC have suppressive activity. CD4+CD25+ T cells with a lower MFI for CD25 expression do not suppress autologous or allogeneic responder T cells and are probably activated T effector cells. Therefore, we set the cut off for the ICOShigh subset within the CD4+CD25high gate at the same cut off we have used to define the CD25high subset, assuming that real CD4+CD25high Treg are “hyperactivated” T cells that might coexpress ICOS.

PBMC obtained from patients’ or NC and patients’ TIL were SCS using a MoFlo high-speed cell sorter (DakoCytomation). Samples were sorted for separation of the following subsets: CD4+CD25 T cells and CD4+CD25highICOShigh, CD4+CD25highICOSlow, and CD4+CD25high ICOS T cells. The same cut-offs determined for flow analysis of the Treg phenotype were used for SCS (Fig. 1).

SCS fresh CD4+CD25high, CD4+CD25highICOShigh, or CD+CD25highICOSlow T cells obtained from patients’ TIL or PBMC, from PBMC of NC or CD4+CD25 T cells recovered from Tr1 cultures, ICOShigh Treg/CD4+CD25 cocultures, or control cultures, were tested for regulatory function in CSFE-based assays, as previously described (15). The suppressor to RC (S:RC) ratios were 1:1 or 1:5, with at least 0.5 × 105 CFSE-labeled autologous CD4+CD25 RC/well. CD4+ RC were stimulated with soluble OKT3 (10 μg/ml) (ATCC) and IL-2 (150 IU/ml). Cocultures were harvested on day 5 (13).

T cell populations were classified as suppressors, if they inhibited proliferation of the CD4+CD25 RC in the coculture assay and if decreasing the number of CD4+CD25+ T cells relative to the number of CD4+CD25 RC in coculture restored proliferation. CD4+CD25+ T cell populations that satisfied both of these criteria were classified as suppressor T cells. These criteria were applied to all tested T cell populations. All CFSE data were analyzed using the ModFit software provided by Verity Software House. The percentages of suppression were calculated based on the proliferation index of RC alone compared with the proliferation index of cultures containing RC and Treg. The program determines the percent of cells within each peak, and the sum of all peaks in the control culture is taken as 100% of proliferation and 0% of suppression.

To assess whether cell-to-cell contact was necessary for ICOS+ Treg to mediate suppression, polycarbonate 24-well transwell inserts (TR; 0.4 μM) (Corning Costar) were used. At least 5 × 104 CFSE-labeled CD4+CD25 RC were stimulated with soluble OKT3 (1 μg/ml) in the presence of 150 IU IL-2/ml in the lower chambers of the plates. Cells were cultured for 5 days and tested for suppression of RC proliferation as described (7).

The origin of CD4+CD25CD132+IL-10+ Tr1 cells was investigated in two series of experiments as follows:

a. Coincubation of ICOS+ Treg (CD4+CD25high) obtained from TIL with autologous PBMC-derived CD4+CD25 T cells.

To study effects of activated ICOS+ Treg in TIL on autologous CD4+CD25 T cells isolated from PBMC, SCS ICOShigh or ICOSlow CD4+CD25+ TIL were coincubated for 5 days with at least 0.5 × 105 autologous CD4+CD25 RC isolated from PBMC at the 1:1 ratio. CD4+CD25 RC were labeled with the fluorescent dye, CFSE, before cocultures in all assays to differentiate them from the ICOS+ Treg subset. To assess whether cell-to-cell contact was necessary for ICOS+ Treg to induce Tr1 in autologous CD4+CD25 RC, polycarbonate 24-well TR (0.4 μM; Corning Costar) were used in half of the assays. To analyze whether ICOS+ Treg can induce and activate Tr1 in nonactivated vs TCR- and IL-2-activated CD4+CD25 RC, half of the cocultures were set up with unstimulated CD4+CD25 RC and the other half with OKT3 (10 μg/ml)- and IL-2 (150 IU/ml)-stimulated CD4+CD25 T cells. On day 5 of coculture, cells were harvested and the Tr1 phenotype and expression of death molecules was determined by flow cytometry, gating on the CD4+CFSE+ and the CD4+CFSE T cell subsets.

Next, the ability of CD4+CD25 RC coincubated with ICOShigh or ICOSlow Treg to function as suppressor cells was determined. CFSE-based suppression assays were set up using RC autologous CD4+CD25 T cells not previously exposed to ICOShigh or ICOSlow Treg. CD4+CD25 T cells that were coincubated with Treg in the absence of TR were used as S cells. However, to be able to separate these putative S cells from the original cocultures, we sorted for the CD4+ CD25CFSE+ subset. After sorting, these cells were plated for 5 days in medium supplemented with a low dose of IL-2 (50 IU/ml) to down-regulate CFSE fluorescence before being added to suppression assays as CFSE-negative S cells. Their suppressor function was measured using fresh CD4+CD25CFSE+ RC in 5-day proliferation assays, as described above. We have previously determined that a period of 5–10 days is needed to completely dilute CFSE label in CD4+ T cells previously labeled with 1.5 μM of CFSE.

To analyze the role of the tumor microenvironment on ICOS+ Treg, sorted CD4+CD25highICOShigh, CD4+CD25highICOSlow, and CD4+CD25 RC were also cultured in the presence of low doses of “Tr1 cytokines” IL-2, IL-10, and IL-15, as previously described by us (16). These cultures were harvested on day 10 and tested for the Tr1 phenotype, cytokine production, and suppressor function.

b. Conversion of CD4+CD25highICOS+ Treg to Tr1 cells.

We have recently reported that culture of PBMC or TIL from cancer patients in the presence of cytokines IL-2, IL-10, and IL-15 favors expansion of in vivo-primed Tr1 lymphocytes (14). To study the impact of these cytokines on CD4+CD25highICOSlow or CD4+CD25highICOShigh Treg separated by SCS from TIL were cultured in the Tr1 in vitro activated (IVA) culture system, previously described by us (14). Briefly, cells were cultured at 5 × 105/well in 96-well plates (Corning Costar). Cytokines IL-2 (10 IU/ml), IL-10 (20 IU/ml), and IL-15 (20 IU/ml), all purchased from PeproTech, were added to these cultures on days 0, 3, and 6. On day 9, culture medium was replaced by fresh medium containing anti-CD3 Ab (1 μg/ml) (ATCC). On day 10, lymphocytes and culture medium were harvested and the Tr1 phenotype and suppressor function were determined. As positive controls, autologous CD4+CD25 T cells obtained by SCS from TIL or PBMC of patients or NC were also cultured in the Tr1 IVA culture system. Negative controls were cultured in media in the presence of IL-2 alone (10 IU/ml).

For the last 24 h of Tr1 cultures, supernatants (SN) were replaced by fresh complete medium containing anti-CD3 mAb (1 μg/ml), but no exogenous cytokines. The levels of IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-17, GM-CSF, IFN-γ, and TNF-α were measured in the SN using immunobead-based multiplex assays (Luminex). Panels of capture Ab-coated beads and labeled detection Abs were purchased from Biosource International. The reagents were pretested and qualified by the manufacturer to ensure the absence of cross-reactivity among Ab-coated beads. The assay sensitivity varied from 5 to 15 pg/ml, depending on the analyte. The 10-plex assay and IL-17 Ab bead kit were approved for combination and performed in the same experimental runs. The assays were performed using the Bio-Plex system (the instrument and software for data analysis purchased from Bio-Rad). The assay was performed according to manufacturer’s instructions.

Tissue samples were embedded in OCT, and 5-μm frozen sections were cut in a cryostat, dried, fixed for 10 min in cold acetone and ethanol (1:1), and again dried in air. The following mAbs were used for staining: mouse anti-human CD4/FITC (diluted 1/100 in PBS), anti-human CD25/PE (1/50) (BD Pharmingen), rat anti-human Foxp3 (1/100) (eBioscience), and rat anti-human IL-10 (1/50) (Santa Cruz Biotechnology). In addition, polyclonal Abs rabbit anti-human TGF-β1 (1/50) and polyclonal goat anti-human ICOS (1/50) (Santa Cruz Biotechnology) were used. As secondary Abs for Foxp3, IL-10, TGF-β1, and Alexa Fluor 488-labeled goat anti-rat and goat anti-rabbit Abs were used (1/500). To detect ICOS, Cy-5-labeled donkey anti-goat (Jackson ImmunoResearch Laboratories) Ab was used at 1/1000 dilution. To eliminate nonspecific binding of secondary Abs, tissue sections were initially incubated with 2% BSA (BSA diluted in PBS) for 1 h and washed five times in 0.5% BSA. Sections were incubated with Abs for 1 h at room temperature in a moist chamber. Next, slides were washed in 0.5% BSA and then incubated with the secondary Abs under the same conditions and in the dark. Primary Abs were omitted in all negative controls. Sections were incubated in a medium with 4′,6-diamidino-2-phenylindole (Vector Laboratories) to trace cell nuclei. Slides were evaluated in the Olympus Provis fluorescence microscope under ×400 magnification (mag). For digital image analysis, the software Adobe Photoshop 6.0 version was used.

Data were summarized by descriptive statistics (mean and SD for continuous variables; frequency and percentage for categorical variables). All statistical analyses were performed using the Student’s t test, and values of p < 0.05 were considered significant.

Flow cytometry analyses of TIL isolated from tumor tissues showed that the CD3+CD4+ T cell subset accounted for 45 ± 18% (mean ± SD) of all T cells, with a range of 30–65%. The gating strategy used to identify this cell subset is shown in Fig. 1,A. Among these CD4+ T cells, an average of 15 ± 2.5% were CD25+ (Table I). This enrichment in CD4+CD25+ T cells was significant (p < 0.002) compared with circulating CD3+CD4+ T cells (6 ± 3%) in melanoma patients or in NC (2 ± 1.5%). We also determined the proportions of CD25high cells present in the CD3+CD4+ subset in TIL (7.5 ± 3.5%; n = 10) and in PBMC of melanoma patients (3.5 ± 1.5%; n = 10) as well as NC (1 ± 0.75%; n = 10), as shown in Table I. The mean percentages of CD4+CD25+ as well as CD4+CD25high T cells were higher in TIL and in PBMC of melanoma patients (p < 0.002 and p < 0.01, respectively) compared with that in PBMC of NC.

Table I.

Distribution of CD4+ cell subsets in PBMC or TIL in patients with melanoma and PBMC of NCa

PBMC NC (%)PBMC Melanoma (%)TIL Melanoma (%)p Valuesb
CD3+CD4+ 50 ± 12 42 ± 20 45 ± 18 NSDc 
CD4+CD25+ 2 ± 1.5 6 ± 3 15 ± 2.5 <0.002 
CD4+CD25high within CD4+CD25+ 1 ± 0.75 3.5 ± 1.5 7.5 ± 3.5 <0.01 
PBMC NC (%)PBMC Melanoma (%)TIL Melanoma (%)p Valuesb
CD3+CD4+ 50 ± 12 42 ± 20 45 ± 18 NSDc 
CD4+CD25+ 2 ± 1.5 6 ± 3 15 ± 2.5 <0.002 
CD4+CD25high within CD4+CD25+ 1 ± 0.75 3.5 ± 1.5 7.5 ± 3.5 <0.01 
a

The data are mean values ± SD obtained with cells of 10 patients with melanoma and 10 NC.

b

The p values are for differences between PBMC and TIL of melanoma patients.

c

NSD = No significant difference.

For this analysis, CD4+ T cells with a MFI of CD25 expression ≥120 were classified as CD25high (Fig. 1 A). An average MFI for CD4+CD25high cells in TIL was 175 ± 25, while that for CD4+CD25high cells in PBMC was 135 ± 15. Thus, CD4+ TIL expressed higher levels of CD25 than did circulating CD4+ T cells.

Next, we determined ICOS expression on the CD4+CD25high and CD4+CD25 subsets in PBMC as well as TIL in melanoma patients or in PBMC of NC. As shown in Fig. 1,B, ICOS was only weakly expressed on CD4+CD25high T cells in PBMC of NC or melanoma patients. In contrast, ICOS was highly expressed on CD4+CD25high T cells isolated from TIL in melanoma patients (Fig. 1,B). ICOS was not detected on the CD4+CD25 T cell subset in PBMC of NC or patients or in TIL (data not shown). Further, we analyzed the % positive cells as well as average MFI of ICOShigh, ICOSint, and ICOSlow T cell subsets within the CD4+CD25high T cell subset in PBMC from NC and patients or from TIL. As illustrated in Fig. 1,C, the CD4+CD25high T cell subset in PBMC from melanoma patients was mainly composed by ICOSlow T cells with few ICOSint or ICOShigh T cells, and the MFI in all three subsets was low. Similar data were observed with cells of NC (data not shown). In contrast, the CD4+CD25high T cell subset in TIL comprised ICOShigh and ICOSint T cells (p < 0.01) with few ICOSlow T cells. Of note, the MFI of the ICOShigh and ICOSint subsets in TIL was greatly enhanced (p ≤ 0.001 and p ≤ 0.01, respectively) compared with the ICOS low subset (Fig. 1 C).

Expression of T cell activation markers, Treg markers, “death” molecules as well as IL-10, sTGF-β1 (membrane-bound), and iTGF-β1 (intracellular) on gated CD4+CD25high T cells was determined in isolated TIL and PBMC of melanoma subjects or NC by multiparameter flow cytometry.

Back-gating on CD4+CD25high T cells, we show in Fig. 2,A that in PBMC of NC, only a proportion of the cells expressed markers usually associated with nTreg: CD62L, GITR, Foxp3, or CTLA-4, and only about half were CD45RO+ memory cells expressing Granzyme B. These results suggest that in NC, most CD4+CD25high nTreg are in the precursor stage and only a subset are activated or fully differentiated to perform suppressor function. In contrast, both PBMC and TIL obtained from melanoma patients had CD4+CD25high T cells with an activated/memory phenotype (CD62LlowCD45RO+CD45RA) (Fig. 2,B). These memory T cells in TIL and PBMC of melanoma patients were largely Foxp3+, CTLA-4+, and iTGF-β1+, with the majority expressing Fas (CD95) (Fig. 2,B). Few CD4+CD25high T cells were positive for CD45RA (6.8 ± 2.1%) or CD127 (14 ± 10%) (data not shown). Of note, TIL and PBMC obtained from melanoma patients showed strong expression of Granzyme A and Granzyme B (Fig. 2,B), whereas Treg in PBMC of NC expressed Granzyme B but not Granzyme A (Fig. 2 A).

FIGURE 2.

Phenotypic characteristics of CD4+CD25high Treg isolated from PBMC of NC or TIL and PBMC of melanoma patients. A, Percentages of cells positive for various markers within the CD4+CD25high T cell subset. Flow cytometry-based analysis was performed using PBMC of NC. B, Percentages of cells positive for various markers within CD4+CD25high T cell subsets in TIL (black bars) or PBMC of melanoma patients. Asterisks indicate significant differences between PBMC and TIL. C, Representative flow dot plots obtained with TIL of one subject for membrane-bound sTGF-β1, iTGF-β, GITR, and IL-10. D, MFI for expression of various markers on the surface of CD4+CD25highICOShigh or CD4+CD25highICOSlow T cells in TIL obtained from melanoma patients. E, Representative dot plots obtained with TIL of one subject for expression of Foxp3 and iTGF-β in ICOShigh or ICOSlow CD4+CD25high T cell subsets. Histograms show mean values ± SD from experiments performed with cells of 10 NC and 10 melanoma patients.

FIGURE 2.

Phenotypic characteristics of CD4+CD25high Treg isolated from PBMC of NC or TIL and PBMC of melanoma patients. A, Percentages of cells positive for various markers within the CD4+CD25high T cell subset. Flow cytometry-based analysis was performed using PBMC of NC. B, Percentages of cells positive for various markers within CD4+CD25high T cell subsets in TIL (black bars) or PBMC of melanoma patients. Asterisks indicate significant differences between PBMC and TIL. C, Representative flow dot plots obtained with TIL of one subject for membrane-bound sTGF-β1, iTGF-β, GITR, and IL-10. D, MFI for expression of various markers on the surface of CD4+CD25highICOShigh or CD4+CD25highICOSlow T cells in TIL obtained from melanoma patients. E, Representative dot plots obtained with TIL of one subject for expression of Foxp3 and iTGF-β in ICOShigh or ICOSlow CD4+CD25high T cell subsets. Histograms show mean values ± SD from experiments performed with cells of 10 NC and 10 melanoma patients.

Close modal

In TIL, CD4+CD25high T cells expressed a different profile of markers than did this same subset of T cells in PBMC of melanoma patients (Fig. 2,B). For example, most CD4+CD25high T cells in TIL were GITR+, FasL+, IL-10+, and sTGF-β1+ (Fig. 2, B and C) relative to CD4+CD25high T cells in PBMC (p < 0.0012), only few of which were positive for GITR and all of which were negative for FasL, IL-10, and sTGF-β1 (Fig. 2,B). A higher percentage (p < 0.0012) of CD4+CD25high T cells expressed the migration and lymph node homing receptors CCR7 and CD62L in PBMC of melanoma patients than in TIL (Fig. 2,A). FasL was absent from the surface of CD4+CD25high PBMC in patients, while the majority of Treg in TIL were FasL+ (Fig. 2,B). Although GITR was expressed on 54% of CD4+CD25high cells in TIL, CD4+CD25high T cells in PBMC of melanoma patients expressed little GITR (Fig. 2,B). Also, in TIL, close to half of CD4+CD25high Treg had up-regulated CD132 (IL2Rγ) (Fig. 2 B).

Next, expression levels of activation markers, Treg molecules, and “death” molecules were compared on CD4+CD25high ICOShigh and CD4+CD25high ICOSlow T cell subsets sorted from TIL of melanoma patients (Fig. 2,D). Higher expression of markers related to activation, suppressor, and “killer” function of nTreg was evident (p < 0.0015) in a subset of CD4+CD25high ICOShigh T cells (Fig. 2, D and E). These results support our assumption that Treg require strong T cell activation signals to become potent suppressor cells.

In aggregate, our data indicate that considerable phenotypic differences exist between the CD4+CD25high populations in TIL and PBMC of melanoma patients. Only Treg in TIL express GITR, IL-10 and TGF-β1 (intracellular and membrane-bound), both granzyme A and B, and up-regulate expression of CD132. Likewise, CD4+CD25high Treg in PBMC and TIL of melanoma patients were phenotypically distinct from those in PBMC of NC.

To evaluate suppressor activity mediated by sorted CD4+CD25highICOSlow or CD4+CD25highICOShigh T cell subsets, cocultures were set up with CFSE-labeled autologous CD4+CD25 T cells defined here as RC. The cocultures were established at the ratios of 1S:1RC or 1S:5RC, using RC proliferating in response to OKT3 in the presence of IL-2. To analyze proliferation of RC, gating was set to include CD4+CFSE+ T cells, as previously described. The percentages of inhibition of RC proliferation measured varied, depending on the S:RC ratios used and the ICOS expression level. As illustrated in Fig. 3, CD4+CD25highICOShigh Treg isolated from TIL of melanoma patients almost completely suppressed proliferation (mean ± SD = 98 ± 7.5%) of autologous RC (p < 0.001) (see left panel). In contrast, CD4+CD25highICOSlow Treg isolated from TIL of melanoma patients only partially suppressed proliferation of RC (68 ± 15%) at the same S:RC ratio of 1:1 (see left panel). Similarly, CD4+CD25highICOShigh Treg isolated from PBMC of melanoma patients had higher suppression activity (82 ± 12%) than CD4+CD25highICOSlow Treg isolated from PBMC of the same patients (38 ± 10%) (data not shown). The CD4+CD25highICOShigh as well as CD4+CD25highICOSlow Treg subsets isolated from PBMC of NC had weak suppressor activity (26 ± 12%) (data not shown). Further, only the CD4+CD25highICOShigh Treg isolated from TIL suppressed RC proliferation at a S:RC ratio of 1:5 (68 ± 12%), whereas TIL- or PBMC-derived CD4+CD25highICOSlow Treg obtained from patients or NC had no suppressor activity at this ratio (data not shown). These results support our hypothesis that only activated Treg expressing ICOS and receiving costimulatory signals up-regulate activation molecules and become potent suppressor cells. Thus, in patients with cancer, nTreg that are in direct contact with the tumor are “hyperactivated” suppressor cells, whereas Treg in PBMC express significantly less ICOS and are activated to a lesser degree. In contrast, the CD4+CD25high Treg subset in NC is mainly composed by activated CD25+ non-Treg, which may or may not serve as Treg precursors, as previously reported.

FIGURE 3.

Suppression of proliferation by CD4+CD25highICOS+ T cells SCS from TIL. Following sorting, autologous CD4+CD25 RC were labeled with CFSE and stimulated with OKT3 in the presence of IL-2, as described in Materials and Methods. CD4+CD25highICOShigh or CD4+CD25highICOSlow T cells in TIL sorted from a representative melanoma patient were added at the start of the culture to achieve the ratio of 1S:1RC. Half of the assay was performed in the presence of a TR separating the Treg subset (S) from the CD4+CD25 RC (right). Flow cytometry performed on day 5 of culture gave the plots representing cell division as measured by the dilution of CFSE and analyzed using the ModFit software. The average percentages of proliferation inhibition of RC in each culture are shown.

FIGURE 3.

Suppression of proliferation by CD4+CD25highICOS+ T cells SCS from TIL. Following sorting, autologous CD4+CD25 RC were labeled with CFSE and stimulated with OKT3 in the presence of IL-2, as described in Materials and Methods. CD4+CD25highICOShigh or CD4+CD25highICOSlow T cells in TIL sorted from a representative melanoma patient were added at the start of the culture to achieve the ratio of 1S:1RC. Half of the assay was performed in the presence of a TR separating the Treg subset (S) from the CD4+CD25 RC (right). Flow cytometry performed on day 5 of culture gave the plots representing cell division as measured by the dilution of CFSE and analyzed using the ModFit software. The average percentages of proliferation inhibition of RC in each culture are shown.

Close modal

A large majority (75–96%) of CD4+CD25high T cells in TIL of melanoma patients were positive for iIL-10 and iTGF-β1 as well as sTGF-β1 by flow cytometry (Fig. 2,A). Importantly, as described above (Fig. 2 C), CD4+CD25highICOShigh T cells in TIL expressed the highest levels of inhibitory cytokines and mediated stronger suppression relative to CD4+CD25highICOSlow Treg, which expressed low levels of iIL-10 and TGF-β1. Therefore, we hypothesized that Treg function of CD4+CD25highICOShigh T cells in TIL was mediated via IL-10 and TGF-β1 secretion and required cell-to-cell contact. To test this hypothesis, CD4+CD25highICOShigh or CD4+CD25highICOSlow Treg isolated from TIL were coincubated at the S:RC ratio of 1:1 with CFSE-labeled autologous CD4+CD25 RC responding to OKT3 and IL-2 in the presence or absence of TR that permit diffusion of soluble factors but prevent cell-to-cell contact.

Coincubation of CD4+CD25highICOShigh T cells isolated from TIL in the presence of TR inhibited suppression of RC proliferation by ∼50% (Fig. 3, upper right panel). In contrast, incubation of CD4+CD25highICOSlow T cells isolated from TIL in the presence of TR almost completely abrogated suppression of RC proliferation (Fig. 3, lower right panel). These results confirm our previously published results (6), indicating that cytokine-negative nTreg depend entirely on direct cell-to-cell contact for suppression of RC proliferation, whereas in those nTreg that secrete inhibitory cytokines, direct cell-to-cell contact is not required for suppression of RC proliferation but it might enhance cytokine-mediated suppression. This is especially evident with ICOShigh Treg; these cells express high levels of membrane-bound TGF-β1, and direct cell-to-cell contact might facilitate suppression of RC proliferation by this cytokine.

We and others have previously shown that Treg require activation via the TCR and IL-2/Stat5b signaling to develop and mediate effective suppression (11, 17, 18, 19, 20). However, the mechanisms responsible for suppression of immune cell proliferation by strongly activated Treg are not clear. We have also shown that both nTreg as well as Tr1 cells infiltrate human solid tumors and that Tr1 cells form cell clusters with nTreg in the tumor microenvironment (7). Based on this observation, we hypothesized that highly activated ICOS+ nTreg can induce Tr1 cells in autologous CD4+CD25 T cells that are attracted from the peripheral circulation into the tumor.

SCS ICOShigh or ICOSlow CD4+CD25high T cells obtained from TIL were coincubated with autologous PBMC-derived CD4+CD25 RC at the S:RC ratio of 1:1. As described above, the assays were performed in the presence or absence of TR and using CD4+CD25 RC activated or not activated with OKT3 and IL-2. In the assays that were performed in the absence of TR, CD4+CD25 T cells were labeled with CFSE to differentiate them from the cocultured Treg. On day 5 of coculture, expression of activation markers, Treg markers, and “death” molecules was determined using flow cytometry and gating on CD4+CD25CFSE+ vs CD4+CD25CFSE T cell subsets. Fig. 4,A shows that before any culture, ex vivo-harvested CD4+CD25 RC were largely negative for Foxp3, CTLA-4, IL-10, or TGF-β and only 20–30% expressed CD132 and ICOS. When ICOShighCD4+CD25high Treg isolated from TIL were coincubated with autologous nonactivated CD4+CD25 RC in the absence of TR, the cells generated in these cocultures expressed the phenotype of Tr1 cells, which is characterized by expression of CD132 and IL-10 in the absence of CD25 (Fig. 4,B). These Tr1 cells induced by ICOShigh Treg expressed a memory T cell phenotype (CD4+CD25CD62LlowCD45RO+CD45RA), low levels of Foxp3 and CTLA-4, and had significantly higher (p < 0.006) levels of “suppressor” cytokines IL-10 and TGF-β1 (intracellular and membrane-bound) than those in ex vivo-tested CD4+CD25 precursors (Fig. 4 A, B, and D). These “induced” Tr1 cells also expressed higher levels of ICOS (p < 0.006) and CD132 (p < 0.001) and high levels of the cytotoxins Granzyme A and Granzyme B.

FIGURE 4.

CD4+CD25highICOShigh Treg isolated from TIL induce Tr1 cells in autologous CD4+CD25 T cells isolated from PBMC. A, The average mean fluorescence (MFI ± SD) of cells positive for various markers within the CD4+CD25 T cell population isolated from PBMC of melanoma patients (n = 5), before coculture with autologous CD4+CD25highICOShigh Treg isolated from TIL, is shown (white bars). B, The average mean fluorescence (MFI ± SD) of cells positive for various markers within the CD4+CD25 T cell population isolated from PBMC of melanoma patients (n = 5), after coculture with autologous CD4+CD25highICOShigh Treg isolated from TIL in the presence of a TR. C, The average mean fluorescence (MFI ± SD) of cells positive for various markers within the CD4+CD25 T cell population isolated from PBMC of melanoma patients (n = 5), after coculture with autologous CD4+CD25highICOShigh Treg isolated from TIL in the absence of a TR. D, Representative flow dot plots obtained from one subject with melanoma for expression of CD132, IL-10, and iTGF-β1 on CD4+CD25 T cells isolated from the corresponding coculture. All cocultures were performed with nonactivated autologous CD4+CD25 T cells.

FIGURE 4.

CD4+CD25highICOShigh Treg isolated from TIL induce Tr1 cells in autologous CD4+CD25 T cells isolated from PBMC. A, The average mean fluorescence (MFI ± SD) of cells positive for various markers within the CD4+CD25 T cell population isolated from PBMC of melanoma patients (n = 5), before coculture with autologous CD4+CD25highICOShigh Treg isolated from TIL, is shown (white bars). B, The average mean fluorescence (MFI ± SD) of cells positive for various markers within the CD4+CD25 T cell population isolated from PBMC of melanoma patients (n = 5), after coculture with autologous CD4+CD25highICOShigh Treg isolated from TIL in the presence of a TR. C, The average mean fluorescence (MFI ± SD) of cells positive for various markers within the CD4+CD25 T cell population isolated from PBMC of melanoma patients (n = 5), after coculture with autologous CD4+CD25highICOShigh Treg isolated from TIL in the absence of a TR. D, Representative flow dot plots obtained from one subject with melanoma for expression of CD132, IL-10, and iTGF-β1 on CD4+CD25 T cells isolated from the corresponding coculture. All cocultures were performed with nonactivated autologous CD4+CD25 T cells.

Close modal

In contrast, when ICOShighCD4+CD25high Treg isolated from TIL were coincubated with nonactivated CD4+CD25 RC in the presence of TR, the shift to the Tr1 profile was not significant, indicating that direct cell-to-cell contact is necessary for the induction of Tr1 cells from autologous CD4+CD25 T cell precursors (Fig. 4, C and D). Of note, when ICOShighCD4+CD25high Treg were coincubated with CD4+CD25 T cells that were previously activated via the TCR and IL-2, no Tr1 phenotype was induced. Instead, proliferation of the CD4+ T cells was induced and cells with a characteristic “Th2” cytokine profile emerged (Table II). When ICOSlow Treg were coincubated with autologous preactivated or nonactivated CD4+CD25 RC in the absence or presence of TR, no phenotypic changes compared with their phenotype before coculture were observed (Table II). In aggregate, these results suggest that ICOS might participate in the induction and differentiation of CD4+CD25 RC into immunosuppressive Tr1 cells or active Th2 cells, depending on the activation state of the responder population.

Table II.

Th1/Th2 profile and proliferation of PBMC-derived CD4+CD25 RC cells coincubated with autologous CD4+CD25high ICOS+ Trega

CD25 (MFI ± SD)Foxp3 (MFI ± SD)IL-10 (MFI ± SD)IL-2 (pg/ml)IL-4 (pg/ml)% Proliferation
Nonactivated RC coincubated with:       
 CD4+CD25highICOShigh 12 ± 8 15 ± 12 3850 ± 250b 52 ± 85b 22 ± 13b 25 ± 11%b 
 CD4+CD25highICOSlow 18 ± 12 23 ± 12 95 ± 550 2338 ± 4428 169 ± 3895 96 ± 5% 
Preactivated RC (OKT3 and IL-2) coincubated with:       
 CD4+CD25highICOShigh 20 ± 15 13 ± 10 2888 ± 1445.6 3897 ± 6110b 950 ± 4760b 92 ± 7.8%b 
 CD4+CD25highICOSlow 10 ± 5.5 19 ± 13 120 ± 380 2252 ± 7224 140 ± 2351 94 ± 5% 
CD25 (MFI ± SD)Foxp3 (MFI ± SD)IL-10 (MFI ± SD)IL-2 (pg/ml)IL-4 (pg/ml)% Proliferation
Nonactivated RC coincubated with:       
 CD4+CD25highICOShigh 12 ± 8 15 ± 12 3850 ± 250b 52 ± 85b 22 ± 13b 25 ± 11%b 
 CD4+CD25highICOSlow 18 ± 12 23 ± 12 95 ± 550 2338 ± 4428 169 ± 3895 96 ± 5% 
Preactivated RC (OKT3 and IL-2) coincubated with:       
 CD4+CD25highICOShigh 20 ± 15 13 ± 10 2888 ± 1445.6 3897 ± 6110b 950 ± 4760b 92 ± 7.8%b 
 CD4+CD25highICOSlow 10 ± 5.5 19 ± 13 120 ± 380 2252 ± 7224 140 ± 2351 94 ± 5% 
a

The data are from experiments testing ICOS+ Treg subsets obtained from TIL of five patients with melanoma, as described in Materials and Methods.

b

Indicates significant differences in the phenotype compared to that before co-culture.

We next analyzed whether the phenotype of Tr1 cells induced by ICOShigh Treg isolated from TIL corresponds with their suppressor function. The Tr1 cells induced by ICOShigh Treg in the presence or absence of TR were coincubated with autologous CFSE-labeled CD4+CD25 RC at the 1S:1RC ratio as described in Materials and Methods. The Tr1 cells that were induced by ICOShigh Treg in the absence of TR and had a “characteristic” Tr1 phenotype partially suppressed proliferation of CD4+CD25 RC (65 ± 12%) (Fig. 5), whereas Tr1 cells that were induced in the presence of TR and expressed a weak Tr1 phenotype had low (30 ± 15%) suppressor activity (Fig. 5). Thus, cell-to-cell contact between inducer and RC was necessary for the generation of Tr1 cells in this coculture system.

FIGURE 5.

Suppressor activity of Tr1 cells induced from CD4+CD25 T cells isolated from PBMC of a melanoma patient by autologous TIL-derived CD4+CD25highICOShigh Treg. On day 5 of ICOShighTreg/CD4+CD25 T cell cocultures, CD4+CD25 T cells were harvested, as described in Materials and Methods. Autologous untreated CD4+CD25 RC were labeled with CFSE and stimulated with OKT3 in the presence of IL-2, as described in Materials and Methods. CD4+CD25 T cells previously cocultured with autologous TIL-derived CD4+CD25highICOShigh Treg in the presence or absence of a transwell inset were added at the start of the culture to give the ratio of 1S:1RC. Flow cytometry, to measure cell division, was performed on day 5. The dilution of CFSE was analyzed using the ModFit software. The average percentages of the inhibition of RC proliferation are shown.

FIGURE 5.

Suppressor activity of Tr1 cells induced from CD4+CD25 T cells isolated from PBMC of a melanoma patient by autologous TIL-derived CD4+CD25highICOShigh Treg. On day 5 of ICOShighTreg/CD4+CD25 T cell cocultures, CD4+CD25 T cells were harvested, as described in Materials and Methods. Autologous untreated CD4+CD25 RC were labeled with CFSE and stimulated with OKT3 in the presence of IL-2, as described in Materials and Methods. CD4+CD25 T cells previously cocultured with autologous TIL-derived CD4+CD25highICOShigh Treg in the presence or absence of a transwell inset were added at the start of the culture to give the ratio of 1S:1RC. Flow cytometry, to measure cell division, was performed on day 5. The dilution of CFSE was analyzed using the ModFit software. The average percentages of the inhibition of RC proliferation are shown.

Close modal

In the above-described experiments, we showed that CD4+CD25highICOShigh T cells obtained from TIL promoted induction and differentiation of Tr1 cells from autologous CD4+CD25 precursors. As these experiments involved coincubation and then separation of the coincubated cell populations, a possibility had to be considered that CD4+CD25highICOShigh Treg convert into Tr1 cells. To test the hypothesis that CD25highICOShigh Treg in TIL might act as precursors of Tr1, sorted CD4+CD25 RC, ICOShighCD25high T cells, and ICOSlowCD25high T cells from TIL of melanoma patients were incubated for 10 days in media supplemented with low doses of IL-2, IL-10, and IL-15 every 3 days under conditions previously reported to favor outgrowth of CD4+CD25IL-10+TGF-β1+ Tr1 cells (16). As shown in Fig. 6, Tr1 cells were induced in CD4+CD25 RC populations after 10 days in culture, confirming the results we recently reported with cells obtained from head and neck squamous cell carcinoma patients. These Tr1 cells had up-regulated marker expression of Tr1-associated surface or intracellular molecules, as shown in Fig. 6,A, upper row. ICOSlowCD25high TIL did not up-regulate expression of the same markers, as shown in Fig. 6,A, middle row, and Fig. 6,B. However, ICOShighCD25high TIL converted into cells with a phenotype similar to Tr1 cells (Fig. 6,B), with down-regulated expression of IL2Rα (CD25) (MFI < 5) and up-regulated expression of IL2Rγ (CD132; MFI >152; Fig. 6,A, lower row). Furthermore, these Tr1 cells remained positive for IL-10 (MFI: 63 ± 12.5) and TGF-β1 expression (MFI: 38 ± 7.5), in contrast to ICOSlowCD25high TIL (MFI: 3 ± 4.5 and 11 ± 5.2, respectively) (Fig. 6,B). Also, T cells, which outgrew from CD25CD4+ TIL in the Tr1 culture, retained a high expression of IL7Rα (CD127), whereas ICOShigh and ICOSlow TIL remained low for IL7Rα expression (Fig. 6,B). Phenotypic analysis of these different groups of Tr1 cells expanded in the presence of the Tr1 cytokines for apoptosis-related molecules revealed a reciprocal relationship between these molecules and ICOS expression (Fig. 6 C): while non-Tr1 cells expanded from the ICOSlowCD25high Treg subset showed a significantly higher expression of FasL and Granzyme B, those expanded from CD4+CD25 and CD4+CD25highICOShigh T cell subsets had a significantly higher expression of Fas and down-regulated Granzyme B and FasL. The data suggest that these cells might be able to differentially regulate T cell activation-induced cell death.

FIGURE 6.

Transition of ICOShigh nTreg to Tr1 cells. A, Flow cytometry dot blots showing relative expression of selected markers on T lymphocytes cultured in the Tr1 IVA supplemented with low doses of cytokines IL-2, IL-10, and IL-15 for 10 days, as described in Materials and Methods, and analyzed after harvest with flow cytometry. CD4+CD25 T cells (upper row), CD4+CD25highICOShigh (middle row), or CD4+CD25highICOSlow (lower row) sorted from TIL of a representative melanoma patient were cultured, as described above. Gates were set on CD3+CD4+ cells. The data are representative for cells generated in IVA, performed with cells of 10 melanoma patients. The figures indicate percentages of positive cells. B, Phenotypic characteristics shown as MFI ± SD of different CD4+ T cell subsets sorted from TIL of melanoma patients (n = 10) and cultured in Tr1 IVA for 10 days, as described above. Asterisks indicate significant differences (∗, p < 0.05); cells stained with mAb were evaluated by multicolor flow cytometry with gates set on CD3+CD4+. C, Phenotypic profile of apoptotic markers expressed on different T cell subsets separated from TIL of melanoma patients, as described above.

FIGURE 6.

Transition of ICOShigh nTreg to Tr1 cells. A, Flow cytometry dot blots showing relative expression of selected markers on T lymphocytes cultured in the Tr1 IVA supplemented with low doses of cytokines IL-2, IL-10, and IL-15 for 10 days, as described in Materials and Methods, and analyzed after harvest with flow cytometry. CD4+CD25 T cells (upper row), CD4+CD25highICOShigh (middle row), or CD4+CD25highICOSlow (lower row) sorted from TIL of a representative melanoma patient were cultured, as described above. Gates were set on CD3+CD4+ cells. The data are representative for cells generated in IVA, performed with cells of 10 melanoma patients. The figures indicate percentages of positive cells. B, Phenotypic characteristics shown as MFI ± SD of different CD4+ T cell subsets sorted from TIL of melanoma patients (n = 10) and cultured in Tr1 IVA for 10 days, as described above. Asterisks indicate significant differences (∗, p < 0.05); cells stained with mAb were evaluated by multicolor flow cytometry with gates set on CD3+CD4+. C, Phenotypic profile of apoptotic markers expressed on different T cell subsets separated from TIL of melanoma patients, as described above.

Close modal

We also evaluated the ability of T cells expanded from different TIL subsets to suppress autologous T cell proliferation. T cells outgrowing from the three different types of Tr1 cultures were titrated into CFSE-labeled autologous CD4+CD25 RC at different ratios and stimulated with OKT3 and IL-2 for 5 days. Fig. 7 A shows that T cells that outgrew in Tr1 cultures established from CD4+CD25 TIL mediated greater suppression (67% ± 7.7) than T cells outgrowing from CD4+ICOSlowCD25high TIL in the presence of Tr1 cytokines (23% ± 9.1). T cells from cultures with CD4+ICOShighCD25high starting population and cytokines showed even greater suppressor activity (92% ± 5.2) and the difference was highly significant at p < 0.001.

FIGURE 7.

Functional characteristics of Tr1 cells derived from ICOShigh nTreg. A, Percentages (mean values ± SD) of suppression of proliferation in CFSE-labeled CD4+CD25 RC mediated by autologous T cell subsets, which were SCS from TIL of melanoma patients (n = 10), and cultured with “Tr1 cytokines” for 10 days, as described in Materials and Methods, at the S:RC ratio of 1:1. B, Histograms generated using the ModFit software flow diagrams show proliferation of CFSE-labeled RC stimulated with anti-CD3 Ab and cocultured with S for five RC are CD4+CD25 freshly separated autologous T cells obtained from TIL of a representative melanoma patient. S are autologous T cell subsets SCS after culture in the presence of “Tr1 cytokines” for 10 days. S cells were added to R cells at the start of the cocultures at a 1:1 ratio. Percent inhibition of proliferation relative to proliferation of RC alone is indicated in every panel. The results are from one representative experiment of 10 performed with cells obtained from different melanoma patients.

FIGURE 7.

Functional characteristics of Tr1 cells derived from ICOShigh nTreg. A, Percentages (mean values ± SD) of suppression of proliferation in CFSE-labeled CD4+CD25 RC mediated by autologous T cell subsets, which were SCS from TIL of melanoma patients (n = 10), and cultured with “Tr1 cytokines” for 10 days, as described in Materials and Methods, at the S:RC ratio of 1:1. B, Histograms generated using the ModFit software flow diagrams show proliferation of CFSE-labeled RC stimulated with anti-CD3 Ab and cocultured with S for five RC are CD4+CD25 freshly separated autologous T cells obtained from TIL of a representative melanoma patient. S are autologous T cell subsets SCS after culture in the presence of “Tr1 cytokines” for 10 days. S cells were added to R cells at the start of the cocultures at a 1:1 ratio. Percent inhibition of proliferation relative to proliferation of RC alone is indicated in every panel. The results are from one representative experiment of 10 performed with cells obtained from different melanoma patients.

Close modal

We analyzed the SN of different types of Tr1 cultures for levels of cytokines using Luminex technology as described in Materials and Methods. The objective was to identify the type of response that is promoted in each starting population cultured in a Tr1-favoring milieu. As opposed to SN from the control cultures, which consisted of CD4+CD25 TIL cultured in the presence of IL-2 (50 IU/ml) for 10 days, SN from T cell cultures containing CD4+CD25 TIL and “Tr1 cytokines” showed very low levels of proinflammatory cytokines IFN-γ, IL-1β, IL-4, TNF-α, and IL-17 but high levels of IL-10 and IL-2. Surprisingly, SN from T cell cultures with ICOShigh or ICOSlow Treg in the presence of “Tr1 cytokines” showed a comparable anti-inflammatory cytokine profile (Table III). However, as shown in Fig. 7, the MFI for expression of these cytokines in Tr1, which originated from ICOShigh Treg, is significantly higher and these cells are stronger suppressors, compared with Tr1 originated from ICOSlow Treg. These findings suggest that Tr1 cells originating from ICOShigh Treg are able to create a stronger immunosuppressive environment.

Table III.

Levels of cytokines in SN of Tr1 cell cultures generated from various starting subsets of TILa

TIL subsetpg/ml
IFNγIL-1βIL-2IL-4TNFαIL-17IL-10
CD4+CD25 control 16684 ± 1554 63 ± 20.8 26831 ± 1445.6 97 ± 21.0 22688 ± 2760 487 ± 85.5 402 ± 32.3 
CD4+CD25 364 ± 72 198 ± 33.1 2341 ± 171.7 52 ± 22.4 2140 ± 335.1 83.7 ± 56.7 3605 ± 450.0 
CD+CD25+ICOSlow 23.3 ± 29.1 23 ± 21.7 2396 ± 780 21 ± 28.2 35.3 ± 32.9 7.3 ± 6.7 5754 ± 405.4 
CD+CD25+ICOShigh 20.3 ± 19.6 38 ± 37 2533 ± 510 32 ± 30 31.7 ± 31.5 34.7 ± 44.6 5781 ± 462.7 
p Value (vs control) <0.0002 <0.02 <0.00001 <0.01 <0.0001 <0.001 <0.0001 
TIL subsetpg/ml
IFNγIL-1βIL-2IL-4TNFαIL-17IL-10
CD4+CD25 control 16684 ± 1554 63 ± 20.8 26831 ± 1445.6 97 ± 21.0 22688 ± 2760 487 ± 85.5 402 ± 32.3 
CD4+CD25 364 ± 72 198 ± 33.1 2341 ± 171.7 52 ± 22.4 2140 ± 335.1 83.7 ± 56.7 3605 ± 450.0 
CD+CD25+ICOSlow 23.3 ± 29.1 23 ± 21.7 2396 ± 780 21 ± 28.2 35.3 ± 32.9 7.3 ± 6.7 5754 ± 405.4 
CD+CD25+ICOShigh 20.3 ± 19.6 38 ± 37 2533 ± 510 32 ± 30 31.7 ± 31.5 34.7 ± 44.6 5781 ± 462.7 
p Value (vs control) <0.0002 <0.02 <0.00001 <0.01 <0.0001 <0.001 <0.0001 
a

Treg subsets were SCS from TIL of 10 patients with melanoma and cultured in the presence of “Tr1 cytokines” to generate Tr1 cells in vitro, as previously described (14 ). SN of Tr1 cells were tested for cytokine levels by Luminex.

Fig. 8,a shows that using immunofluorescence, numerous ICOS+CD4+ T cells can be visualized in a representative melanoma tissue. In Fig. 8,a2, pink CD4+CD25+ICOS+ nTreg are localized in the proximity of CD4+CD25ICOSlow/neg cells (see inset), implying that these cell subsets can interact in situ. We have also examined distribution in the tumor of ICOSL+ cells (data not shown). In addition to ICOSL+ melanoma cells, which were localized at the border between the lymphocytic infiltrates and tumor nests, numerous infiltrating CD4+ T cells expressed ICOSL. These observations suggest that ICOSL is abundantly expressed in the tumor microenvironment. In Fig. 8,b, the white circles denote the areas of distribution of CD25high Foxp3+ nTreg relative to a cluster of CD25dimFoxp3dim cells, which could be either Tr1 cells or nTreg “transiting” to Tr1 cells in the tumor microenvironment. Finally, Fig. 8 c shows that numerous IL-10+ lymphocytes accumulating in the tumor are ICOS+ and include both CD25+ and CD25neg populations. Multicolor immunofluorescence in situ confirms that ICOS+ Treg with the characteristics of nTreg and Tr1 cells accumulate in melanoma tissues and are in direct contact with each other.

FIGURE 8.

Multicolor immunofluorensence for ICOS+ T cells in melanoma. CD4+CD25+ and CD4+CD25+ICOS+ T cells are shown in sections of a representative tumor of 10 examined. a, Tumor sections stained for CD4+CD25+ and ICOS+ cells: (a1) CD4+ cells are green, CD25+ cells are red, and CD4+CD25+ cells are yellow; and (a2) ICOS+ cells are blue and CD4+CD25+ICOS+ cells are pink. Note that CD4+CD25+ICOS+ Treg are located adjacent to CD4+CD25ICOS T cells (inset). b, A section stained for CD25, Foxp3, and ICOS; in b1, CD25+ cells are red, Foxp3 cells are green, and CD25+Foxp3+ cells appear yellow, and cells that are CD25lowFoxp3loware probably Tr1 cells; in b2, ICOS+ cells are blue and nuclei are pseudocolored white; and in b3, an overlay showing that Tr1 cells (CD25lowFoxp3low) express ICOS (blue). c, A section stained for CD25, IL-10, and ICOS; in c1, CD25+ cells are red, IL-10+ cells are green, and CD25+IL-10+ cells are yellow; in c2, ICOS+ cells are blue and nuclei are pseudocolored white; and in c3, an overlay showing that CD25+IL-10+ Treg are pink, i.e., ICOS+ and that nearly all of IL-10+ cells are also ICOS+. Mag. ×400. Inset in 8a, mag. ×600.

FIGURE 8.

Multicolor immunofluorensence for ICOS+ T cells in melanoma. CD4+CD25+ and CD4+CD25+ICOS+ T cells are shown in sections of a representative tumor of 10 examined. a, Tumor sections stained for CD4+CD25+ and ICOS+ cells: (a1) CD4+ cells are green, CD25+ cells are red, and CD4+CD25+ cells are yellow; and (a2) ICOS+ cells are blue and CD4+CD25+ICOS+ cells are pink. Note that CD4+CD25+ICOS+ Treg are located adjacent to CD4+CD25ICOS T cells (inset). b, A section stained for CD25, Foxp3, and ICOS; in b1, CD25+ cells are red, Foxp3 cells are green, and CD25+Foxp3+ cells appear yellow, and cells that are CD25lowFoxp3loware probably Tr1 cells; in b2, ICOS+ cells are blue and nuclei are pseudocolored white; and in b3, an overlay showing that Tr1 cells (CD25lowFoxp3low) express ICOS (blue). c, A section stained for CD25, IL-10, and ICOS; in c1, CD25+ cells are red, IL-10+ cells are green, and CD25+IL-10+ cells are yellow; in c2, ICOS+ cells are blue and nuclei are pseudocolored white; and in c3, an overlay showing that CD25+IL-10+ Treg are pink, i.e., ICOS+ and that nearly all of IL-10+ cells are also ICOS+. Mag. ×400. Inset in 8a, mag. ×600.

Close modal

The generation of effective immune responses to T cell-dependent Ags, including TA, is a dynamic process involving the coordinated action of cytokines and costimulatory molecules expressed on APC and T cells (21). During this process, the type and level of costimulatory signals delivered to T cells will ultimately govern the outcome of Ag presentation in the context of MHC molecules (22). Among major costimulatory receptor-ligand pairs identified to date, the CD28/B7 pathway has been most extensively studied (23, 24). More recently described, the ICOS/ICOSL (ICOS/B7-H2) pathway appears to play a key role in the control of T cell-mediated immune responses and tolerance. Unlike CD28, which is constitutively expressed on the surface of T cells, ICOS requires up-regulation delivered via TCR stimulation (25). Evidence suggests that B7/CD28 interactions are important in clonal expansion and effector functions of naive CD4+ T cells, whereas ICOS/ICOSL interactions control the responses of already activated T cells. Although CD28/B7 interactions are important for IL-2-driven responses, costimulation through ICOS is particularly effective in enhancing IL-10 production, and, in direct comparisons, ICOS is more potent than CD28 in inducing the production of IL-10 (26).

Recently reported data indicate that ICOS may be expressed not only on effector T cells but also on Treg. Storkus and colleagues (27) show that murine CD4+ TIL with an activated/memory phenotype (ICOShighCD62LlowCD45RBlow) suppressed allo-specific T cell proliferation and IFN-γ production from cross-primed anti-CMS4 CD8+ T cells via a mechanism at least partially dependent on IL-10 and TGF-β1. In patients with prostate cancer, Treg in TIL were found to express higher levels of ICOS and GITR compared with Treg in PBMC (13). In addition, recent studies show that the development of IL-10-secreting CD4+CD25+ Treg is mediated by mature dendritic cells and allergens in a process that requires T cell costimulation via the ICOS-ICOSL pathway (28). Importantly, these “ICOS-induced” Treg suppressed the proliferation of allergen-specific RC via IL-10, supporting the hypothesis that Treg- and IL-10-mediated suppression requires signaling via CD28 costimulation (29, 30, 31). De Jong and colleagues (30) also demonstrated that induction of IL-10 in Treg requires ICOS and signaling via CD28 in mice with colitis. Our results extend these findings by demonstrating that ICOS expression on Treg, defined as CD4+CD25high cells, in TIL of patients with melanoma correlates with high expression levels of Foxp3, CTLA-4, IL-10, and TGF-β1, the molecules that have been associated with nTreg suppressor function (7, 20, 32, 33, 34). Treg appear to comprise ICOShigh and ICOSlow subsets, and ICOShigh Treg in TIL express the highest levels of Foxp3, CTLA-4, IL-10, and intracellular as well as membrane-bound TGF-β1. In contrast, Treg in the patient’s PBMC expressed low levels of ICOS and GITR, no IL-10, and only moderate levels of iTGF-β1.

The ICOShigh Treg subset had enhanced suppressor function compared with autologous ICOSlow or ICOS Treg, which might be related to higher expression of inhibitory cytokines by the former. ICOShigh Treg also express very high levels of Fas, FasL, and Granzyme B, indicating that they might be able to act as “killer” cells of CD4+ and/or CD8+ target cells. We have recently reported that human CD4+CD25highFoxp3+ Treg in cancer patients can induce apoptosis in autologous CD8+ T cells via the Fas/FasL pathway and in autologous CD4+ T cells via a Granzyme B-mediated pathway (12). Thus, ICOS expression on Treg in TIL appears to be associated with increased functional potential of these cells.

The biologic importance of ICOS expression on Treg in human TIL has not been evaluated so far. It now appears that the modulation of immune responses by ICOS-expressing nTreg in the tumor depends on the costimulatory signal delivered to RC and on the local cytokine milieu. We showed for the first time that CD4+CD25highICOShigh nTreg can induce Tr1 cells (CD4+CD25CD132+IL-10+) in autologous CD4+CD25 RC. In contrast, when ICOShigh nTreg were coincubated with CD4+CD25 RC preactivated via TCR and IL-2, proliferation of responder with a Th2 cytokine profile ensued and no Tr1 cells were detected in the cocultures. These results confirm that nonactivated CD4+CD25 RC are precursors of Tr1 (16) and that ICOS costimulation may be critical for their differentiation to Tr1 cells.

Cytokines are another critical component. We have reported (16) and further show in this manuscript that Tr1 precursor cells (CD4+CD25) obtained from patients with cancer are activated by IL-15 and IL-10 in the absence of TCR and/or CD28 engagement and, in contrast to CD4+CD25high Treg and CD4+ RC, require only low dose of IL-2 for their differentiation and expansion. ICOS costimulation induces low levels of IL-2 and IL-10 that may be sufficient to accomplish such differentiation and expansion of naive CD4+CD25 cells in the Tr1 pathway. But if the CD4+CD25 precursors are activated via TCR and/or CD28 engagement, ICOS reinforces this interaction by supporting signaling triggered by TCR and CD28 costimulation, as previously suggested (29). This leads to differentiation and expansion of activated CD4+ RC into effector, rather than regulatory, cells. However, in the tumor microenvironment, activated CD4+ RC that encounter ICOS may not proliferate because of the paucity of IL-2 and the excess of inhibitory cytokines. Instead, the microenvironment enriched in ICOShigh Treg, which express high levels of IL-10, favors induction of Tr1 in CD4+CD25 T cells.

IL-10 suppresses proliferation of CD4+ RC via Stat5 up-regulation. However, Taylor and colleagues (29) have recently shown that IL-10 inhibits CD4+ T cell proliferation only in a defined range of TCR triggering, where T cells require costimulation. These findings suggest that CD4+ T cells that differentiate into Tr1 cells upon encountering ICOShighCD4+CD25high Treg in the tumor are suboptimally stimulated. In fact, the absence or paucity of costimulation in the presence of abundant TA at tumor sites is known to lead to tolerance (35, 36, 37). Our data are consistent with this paradigm.

This selective expansion of Tr1 from CD4+CD25 precursors by ICOShigh nTreg in the suboptimal range of TCR engagement and IL-2 might be an important mechanism for the regulation of T lymphocyte homeostasis. When CD4+CD25high Treg have prevented an excessive expansion of T effector cells, and the immune response is declining after IL-2 has been consumed by activated T effector and nTreg, Tr1 cells might be induced as a secondary response that prevents an excessive induction of new immune responses upon re-challenge with the Ag. This possibility is supported by our finding that when CD4+CD25highFoxp3+ICOShigh Treg are cultured in Tr1-favorable conditions, they differentiate into CD25 Tr1 cells that produce IL-10. Of note, Tr1 cells differentiated from CD4+CD25highFoxp3+ICOShigh Treg have stronger suppressor activity than Tr1 induced from CD4+CD25 T cells or those differentiated from CD4+CD25highFoxp3+ICOSlow Treg. Surprisingly, levels of cytokines in the SN of Tr1 cells originated from ICOShigh Treg were not increased in comparison to those in the SN of Tr1 cultures originated from ICOSlow Treg (Table III). However, the MFI for expression of “suppressor” cytokines in Tr1 cells that had originated from ICOShigh Treg was significantly higher compared with Tr1 cells from ICOSlow Treg. This observation suggests that direct cell-to-cell contact, even in the case of Tr1 cells that mediate suppression via cytokines, might be required to maximize the suppressor function of human Treg cells in cancer.

The observation that highly activated human nTreg present in the tumor microenvironment can “transit” to Tr1 or induce Tr1 cells in activated CD4+CD25 RC might explain our recent findings indicating that Tr1 cells accumulate in tumors of cancer patients and mediate strong suppression of anti-tumor immunity (7). In contrast to nTreg, few Tr1 are detected in the peripheral circulation of patients with cancer. We hypothesize that Tr1 precursor cells might be induced by or differentiate from activated nTreg during the contraction phase of an immune response, when the microenvironment provides suboptimal activation signals for CD4+ effector cells. Tr1 cells might be continuously induced in necrotic tumor areas that are poor in IL-2 and, thus, partially contribute to the autocrine-feedback loop of chronic immunosuppression in cancer patients. In tumor sections obtained from melanoma patients (Fig. 8), CD4+CD25+Foxp3+ICOS+ nTreg colocalized with CD4+CD25IL-10+ Tr1 cells, emphasizing their mutual involvement. Our data show that Tr1 cells originating from ICOShigh Treg express high levels of Fas but not FasL or Granzyme B, whereas Tr1 cells originating from ICOSlow Treg do not express Fas but are FasL+ and Granzyme B+. Such a distinct profile suggests that Tr1 cells originating from ICOSlow Treg might control an excessive induction of Tr1 cells derived from ICOShigh precursors by inducing their apoptosis via Fas/FasL or Granzyme B. In contrast, Tr1 cells originating from ICOSlow Treg that are not fully functional might be eliminated by Tr1 cells originating from ICOShigh Treg. The observation that Treg can suppress RC by several distinct mechanisms supports our hypothesis that activation and contraction of human T cell immune responses may be regulated by different subsets of Treg.

In summary, our data suggest that the modulation of the immune responses by ICOS-expressing nTreg present in the tumor depends upon costimulatory signals delivered to the RC and the cytokine milieu. Targeting of costimulatory and coinhibitory molecules on human Treg and effector T cells with mAbs in combination with cytokine-based therapy might be a promising strategy for enhancing or decreasing peripheral tolerance in patients with autoimmune disease or cancer, respectively.

We acknowledge the technical assistance of the University of Pittsburgh Cancer Institute Flow Cytometry Facility, the Center for Biologic Imaging, and the Immune Monitoring Laboratory at the University of Pittsburgh.

The authors have no financial conflict 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.

1

This work was supported in part by National Institutes of Health Grants R01 DE13918, P01 DE12321, and P01CA109688 (to T.L.W.), and by Philip Morris USA Inc. and Philip Morris International (to C.B., S.L., and T.L.W.).

4

Abbreviations used in this paper: ICOSL, ICOS ligand; nTreg, naturally occurring T regulatory; TA, tumor Ag; TIL, tumor-infiltrating lymphocytes; NC, normal control; RC, responder cell; sTGF-β1, surface-bound TGF-β1; int, intermediate; MFI, mean fluorescence intensity; SCS, single-cell sorted (or sorting); TR, transwell insert; SN, supernatant; mag, magnification; GITR, glucocorticoid-induced TNF receptor; i, intracellular; IVA, in vitro activated.

1
Linsley, P. S., J. Bradshaw, M. Urnes, L. Grosmaire, J. A. Ledbetter.
1993
. CD28 engagement by B7/BB-1 induces transient down-regulation of CD28 synthesis and prolonged unresponsiveness to CD28 signaling.
J. Immunol.
150
:
3161
-3169.
2
Mahajan, S., A. Cervera, M. MacLeod, S. Fillatreau, G. Perona-Wright, S. Meek, A. Smith, A. Macdonald, D. Gray.
2007
. The role of ICOS in the development of CD4 T cell help and the reactivation of memory T cells.
Eur. J. Immunol.
37
:
1796
-1808.
3
Sakaguchi, S., K. Fukuma, K. Kuribayashi, T. Masuda.
1985
. Organ-specific autoimmune diseases induced in mice by elimination of T cell subset: I. Evidence for the active participation of T cells in natural self-tolerance; deficit of a T cell subset as a possible cause of autoimmune disease.
J. Exp. Med.
161
:
72
-87.
4
Asano, M., M. Toda, N. Sakaguchi, S. Sakaguchi.
1996
. Autoimmune disease as a consequence of developmental abnormality of a T cell subpopulation.
J. Exp. Med.
184
:
387
-396.
5
Woo, E. Y., C. S. Chu, T. J. Goletz, K. Schlienger, H. Yeh, G. Coukos, S. C. Rubin, L. R. Kaiser, C. H. June.
2001
. Regulatory CD4+CD25+ T cells in tumors from patients with early-stage non-small cell lung cancer and late-stage ovarian cancer.
Cancer Res.
61
:
4766
-4772.
6
Liyanage, U. K., T. T. Moore, H. G. Joo, Y. Tanaka, V. Herrmann, G. Doherty, J. A. Drebin, S. M. Strasberg, T. J. Eberlein, P. S. Goedegebuure, D. C. Linehan.
2002
. Prevalence of regulatory T cells is increased in peripheral blood and tumor microenvironment of patients with pancreas or breast adenocarcinoma.
J. Immunol.
169
:
2756
-2761.
7
Strauss, L., C. Bergmann, M. Szczepanski, W. Gooding, J. T. Johnson, T. L. Whiteside.
2007
. A unique subset of CD4+CD25highFoxp3+ T cells secreting IL-10 and TGF-β1 mediates suppression in the tumor microenvironment.
Clin. Cancer Res.
13
:
4345
-4354.
8
Curiel, T. J., G. Coukos, L. Zou, X. Alvarez, P. Cheng, P. Mottram, M. Evdemon-Hogan, J. R. Conejo-Garcia, L. Zhang, M. Burow, et al
2004
. Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival.
Nat. Med.
10
:
942
-949.
9
Strauss, L., C. Bergmann, J. Johnson, T. L. Whiteside.
2006
. Functional CD4+CD25highFoxp3+ T cells in the circulation of patients with head and neck squamous cell carcinoma (HNSCC) and correlation with advanced disease.
Arch. Otolaryngol. Head Neck Surg.
132
:
883
-884.
10
Shevach, E. M., R. S. McHugh, C. A. Piccirillo, A. M. Thornton.
2001
. Control of T-cell activation by CD4+ CD25+ suppressor T cells.
Immunol. Rev.
182
:
58
-67.
11
Strauss, L., C. Bergmann, T. L. Whiteside.
2007
. Function and phenotype of human CD4+CD25+Foxp3+ T cell clones in patients with head and neck cancer (HNSCC) and normal controls.
Int. J. Cancer
121
:
2473
-2483.
12
Strauss, L., C. Bergmann, J. Johnson, T. L. Whiteside.
2006
. Expression of Fas ligand (FasL) isoforms on peripheral T-cell subsets in patients with head and neck squamous cell carcinoma (HNSCC).
J. Immumother.
29
:
670
(Abstr.).
13
Miller, A. M., K. Lundberg, V. Ozenci, A. H. Banham, M. Hellstrom, L. Egevad, P. Pisa.
2006
. CD4+CD25high T cells are enriched in the tumor and peripheral blood of prostate cancer patients.
J. Immunol.
177
:
7398
-7405.
14
Whiteside, T. L., S. Miescher, H. R. MacDonald, V. Von Fliedner.
1986
. Separation of tumor-infiltrating lymphocytes from tumor cells in human solid tumors: a comparison between velocity sedimentation and discontinuous density gradients.
J. Immunol. Methods
90
:
221
-233.
15
Strauss, L., T. L. Whiteside, A. Knights, C. Bergmann, A. Knuth, A. Zippelius.
2007
. Selective survival of naturally occurring human CD4+CD25+Foxp3+ regulatory T cells cultured with rapamycin.
J. Immunol.
178
:
320
-329.
16
Bergmann, C., L. Strauss, R. Zeidler, S. Lang, T. L. Whiteside.
2007
. Expansion and characteristics of human T regulatory type 1 cells in co-cultures simulating tumor microenvironment.
Cancer Immunol. Immunother.
56
:
1429
-1442.
17
Burchill, M. A., J. Yang, C. Vogtenhuber, B. R. Blazar, M. A. Farrar.
2007
. IL-2 receptor β-dependent STAT5 activation is required for the development of Foxp3+ regulatory T cells.
J. Immunol.
178
:
280
-290.
18
Thornton, A. M., C. A. Piccirillo, E. M. Shevach.
2004
. Activation requirements for the induction of CD4+CD25+ T cell suppressor function.
Eur. J. Immunol.
34
:
366
-376.
19
Cohen, A. C., K. C. Nadeau, W. Tu, V. Hwa, K. Dionis, L. Bezrodnik, A. Teper, M. Gaillar, J. Heinrich, A. M. Krensky, et al
2006
. Cutting edge: decreased accumulation and regulatory function of CD4+CD25high T cells in human STAT5b deficiency.
J. Immunol.
177
:
2770
-2774.
20
Fontenot, J. D., M. A. Gavin, A. Y. Rudensky.
2003
. Foxp3 programs the development and function of CD4+CD25+ regulatory T cells.
Nat. Immunol.
4
:
330
-336.
21
Robbins, P. F., Y. Kawakami.
1996
. Human tumor antigens recognized by T cells.
Curr. Opin. Immunol.
8
:
628
-636.
22
Appleman, L. J., V. A. Boussiotis.
2003
. T cell anergy and costimulation.
Immunol. Rev.
192
:
161
-180.
23
Chambers, C. A..
2001
. The expanding world of co-stimulation: the two-signal model revisited.
Trends Immunol.
22
:
217
-223.
24
Carreno, B. M., M. Collins.
2002
. The B7 family of ligands and its receptors: new pathways for costimulation and inhibition of immune responses.
Annu. Rev. Immunol.
20
:
29
-53.
25
Hutloff, A., A. M. Dittrich, K. C. Beier, B. Eljaschewitsch, R. Kraft, I. Anagnostopoulos, R. A. Kroczek.
1999
. ICOS is an inducible T-cell co-stimulator structurally and functionally related to CD28.
Nature
397
:
263
-266.
26
Witsch, E. J., M. Peiser, A. Hutloff, K. Buchner, B. G. Dorner, H. Jonuleit, H. W. Mages, R. A. Kroczek.
2002
. ICOS and CD28 reversely regulate IL-10 on re-activation of human effector T cells with mature dendritic cells.
Eur. J. Immunol.
32
:
2680
-2686.
27
Berhanu, A., J. Huang, S. C. Watkins, H. Okada, W. J. Storkus.
2007
. Treatment-enhanced CD4+Foxp3+ glucocorticoid-induced TNF receptor family related high regulatory tumor-infiltrating T cells limit the effectiveness of cytokine-based immunotherapy.
J. Immunol.
178
:
3400
-3408.
28
Akbari, O., G. J. Freeman, E. H. Meyer, E. A. Greenfield, T. T. Chang, A. H. Sharpe, G. Berry, R. H. DeKruyff, D. T. Umetsu.
2002
. Antigen-specific regulatory T cells develop via the ICOS-ICOS-ligand pathway and inhibit allergen-induced airway hyperreactivity.
Nat. Med.
8
:
1024
-1032.
29
Riley, J. L., P. J. Blair, J. T. Musser, R. Abe, K. Tezuka, T. Tsuji, C. H. June.
2001
. ICOS costimulation requires IL-2 and can be prevented by CTLA-4 engagement.
J. Immunol.
166
:
4943
-4948.
30
de Jong, Y. P., S. T. Rietdijk, W. A. Faubion, A. C. Abadia-Molina, K. Clarke, E. Mizoguchi, J. Tian, T. Delaney, S. Manning, J. C. Gutierrez-Ramos, et al
2004
. Blocking inducible co-stimulator in the absence of CD28 impairs Th1 and CD25+ regulatory T cells in murine colitis.
Int. Immunol.
16
:
205
-213.
31
Tai, X., M. Cowan, L. Feigenbaum, A. Singer.
2005
. CD28 costimulation of developing thymocytes induces Foxp3 expression and regulatory T cell differentiation independently of interleukin 2.
Nat. Immunol.
6
:
152
-162.
32
Walker, M. R., D. J. Kasprowicz, V. H. Gersuk, A. Benard, M. Van Landeghen, J. H. Buckner, S. F. Ziegler.
2003
. Induction of FoxP3 and acquisition of T regulatory activity by stimulated human CD4+CD25 T cells.
J. Clin. Invest.
112
:
1437
-1443.
33
Miyara, M., S. Sakaguchi.
2007
. Natural regulatory T cells: mechanisms of suppression.
Trends Mol. Med.
13
:
108
-116.
34
Fahlen, L., S. Read, L. Gorelik, S. D. Hurst, R. L. Coffman, R. A. Flavell, F. Powrie.
2005
. T cells that cannot respond to TGF-β escape control by CD4+CD25+ regulatory T cells.
J. Exp. Med.
201
:
737
-746.
35
Sotomayor, E. M., I. Borrello, F. M. Rattis, A. G. Cuenca, J. Abrams, K. Staveley-O’Carroll, H. I. Levitsky.
2001
. Cross-presentation of tumor antigens by bone marrow-derived antigen-presenting cells is the dominant mechanism in the induction of T-cell tolerance during B-cell lymphoma progression.
Blood
98
:
1070
-1077.
36
Enk, A. H., H. Jonuleit, J. Saloga, J. Knop.
1997
. Dendritic cells as mediators of tumor-induced tolerance in metastatic melanoma.
Int. J. Cancer
73
:
309
-316.
37
Beecham, E. J., Q. Ma, R. Ripley, R. P. Junghans.
2000
. Coupling CD28 co-stimulation to immunoglobulin T-cell receptor molecules: the dynamics of T-cell proliferation and death.
J. Immunother.
23
:
631
-642.