IL-10 and PD-1 cooperate to limit the activity of tumor-specific CD8⁺ T cells

Abstract

Immune checkpoint inhibitors show great promise as therapy for advanced melanoma, heightening the need to determine the most effective use of these agents. Here, we report that programmed death-1^high (PD-1^high) tumor antigen (TA)-specific CD8⁺ T cells present at periphery and at tumor sites in patients with advanced melanoma upregulate IL-10 receptor (IL-10R) expression. Multiple subsets of peripheral blood mononucleocytes from melanoma patients produce IL-10, which acts directly on IL-10R⁺ TA-specific CD8⁺ T cells to limit their proliferation and survival. PD-1 blockade augments expression of IL-10R by TA-specific CD8⁺ T cells, thereby increasing their sensitivity to the immunosuppressive effects of endogenous IL-10. Conversely, IL-10 blockade strengthened the effects of PD-1 blockade in expanding TA-specific CD8⁺ T cells and reinforcing their function. Collectively, our findings offer a rationale to block both IL-10 and PD-1 to strengthen the counteraction of T cell immunosuppression and enhance the activity of TA-specific CD8⁺ T cell in advanced melanoma patients.

Introduction

T cells recognize tumor antigens (TA) expressed by melanoma cells and can induce tumor regression in animals and in humans (1). However, in the presence of high antigen load in chronic viral infections and in cancer, antigen-specific CD8⁺ T cells become dysfunctional/exhausted and lose their capacities to proliferate, produce cytokines and lyse tumor cells (2, 3). These dysfunctional antigen-specific CD8⁺ T cells upregulate a number of inhibitory receptors including PD-1 and PD-1 blockade augments their expansion and functions in vitro (4-8). The capability of PD-1 blockade to provide persistent clinical benefit to approximately 30-40% of patients with advanced melanoma has now been demonstrated in multiple clinical trials (9, 10). To further improve the clinical efficacy of PD-1 blockade, it appears critical to identify additional strategies to counteract the major negative immunoregulatory pathways impairing TA-specific CD8⁺ T cells in the tumor microenvironment (TME).

IL-10 is a potent anti-inflammatory molecule produced by innate and adaptive immune cells including T cells, NK cells, antigen-presenting cells as well as tumor cells including melanoma (11-15). The immunosuppressive role of endogenous IL-10 in impeding antigen-presenting cells is supported by the demonstration that neutralizing IL-10 with anti-IL-10R antibodies is required for the stimulation of potent Th1 OVA-specific and TA-specific T cell responses in mice treated with toll-like receptor ligands (16, 17). The role of IL10 role in cancer immunology remains controversial. In experimental tumor models, IL-10 appears to either promote or facilitate tumor rejections (18-26). The effects of IL-10 and IL-10 blockade on human TA-specific CD8⁺ T cells have not been thoroughly evaluated yet. In chronic viral infections, IL-10 and PD-1 pathways act synergistically through distinct pathways to suppress T cell functions, and dual IL-10 and PD-1 blockade appears more effective in restoring antiviral CD8⁺ and CD4⁺ T cell responses and viral clearance than either single blockade alone (27, 28). Whether IL-10 added to PD-1 blockade further enhances TA-specific CD8⁺ T cell functions in melanoma patients remains unknown.

Here, we report for the first time that PD-1^high CD8⁺ T cells directed against the cancer-germline antigen NY-ESO-1 and PD-1^high CD8⁺ tumor-infiltrating lymphocytes (TILs) isolated from patients with advanced melanoma, upregulate IL-10R. Although PD-1 blockade in the presence of cognate antigen increases the expansion and functions of NY-ESO-1–specific CD8⁺ T cells, it also augments IL-10R expression by TA-specific CD8⁺ T cells. We show that IL-10 blockade adds to PD-1 blockade to increase the expansion and functions of NY-ESO-1–specific CD8⁺ T cells, supporting the role of dual IL-10 and PD-1 blockade to enhance TA-specific CTL responses to melanoma.

Materials and Methods

Subjects

Blood samples and tumor specimen were obtained under the University of Pittsburgh Cancer Institute Institutional Review Board (IRB)-approved protocols 00-079 and 05-140 from twelve HLA-A2⁺ patients with NY-ESO-1⁺ stage IV melanoma and spontaneous NY-ESO-1–specific CD8⁺ T-cells (supplementary Table 1). The PBMCs used in this study were obtained from melanoma patients with no prior immunotherapy. The same patients were used across all assays.

Phenotypic analysis

CD8⁺ T lymphocytes were purified from PBMCs of patients using MACS Column Technology (Miltenyi Biotec, San Diego, CA). Alternatively, PBMCs were incubated for 6 d in culture medium containing 50 IU/ml rhIL-2 (PeproTech, Rocky Hill, NJ) with peptide NY-ESO-1 157–165 or medium alone in the presence of 10 μg/ml anti-IL-10R (clone 3F9, Biolegend, San Diego, CA) or anti-PD-L1 (clone MIH1, eBioscience, San Diego, CA) or isotype control antibodies and/or 20 ng/ml rhIL-10 (PeproTech). Cells were incubated either with HLA-A2/NY-ESO-1 157-165, HLA-A2/CMV 495-503, HLA-A2/EBV-BMLF-1 280-288, HLA-A2/Flu-M 58-66, or HLA-A2/MART-1 26-35 tetramers (TC metrix Ltd, Epalinges, Switzerland) prior to staining with PD-1-PerCPCy5.5, IL-10R-PE (Biolegend), and CD8- PE-Cy7, CD14-ECD, CD19-ECD, CD56-biotin (Beckman Coulter, Brea CA), and streptavidin-ECD (Invitrogen, Grand Island, NY) conjugated antibodies or reagent. Alternatively, after tetramer labeling, cells were stained with PD-1-PECy7 (Biolegend), CD8-V500, CD69-FITC or CD57-FITC, CD38-PerCp-Cy5.5 (BD Biosciences, San Jose, CA), HLA-DR-ECD or CD25-ECD (Beckman Coulter). Alternatively, PBMCs were stained with CD11c-Alexa700 (eBioscience), CD19-APCCy7, CD56-FITC (BD Biosciences), CD8-PECy7, CD4-PerCPCy5.5 (Biolegend), CD14-ECD, IL-10R-PE. A violet amine reactive dye (Invitrogen) was used to assess the viability of the cells. p-STAT3-Alexa 488 (BD Biosciences) was used to identify the phosphorylated form of STAT3 (Ser727). 2.5×10⁶ events were collected on a FACSAria machine (BD Biosciences) and analyzed with FlowJo software (Tree Star, Ashland, OR).

IL-10 detection

The concentrations of IL-10 in supernatant or sera were determined using BD OptEIA Human IL-10 ELISA Set (BD Biosciences). To test IL-10 production, CD8⁺ T cells were purified from PBMCs (MACS Column Technology), and labeled with tet-APC, CD8-PECy7, CD4-PE and violet. 6×10⁴ FACS-sorted cells were distributed into 96 wells with 200 μl medium containing 50 IU/ml rhIL-2, T2 cells (2:1 ratio) pulsed with peptide NY-ESO-1 157–165 or control peptide HIVpol 476-484 (10 μg/ml). Supernatant was collected for ELISA assay after 48-hour incubation. Alternatively, CD4⁺ and CD8⁺ T cells were separated by MACS Column Technology, the rest of the cells were labeled with CD14-ECD, CD11c-Alexa700, CD3-PerCPCy5.5, CD56-PE and CD19-FITC and sorted. Total RNA was extracted from each cell subset, and IL-10 mRNA was detected using RT-QPCR as previously described (43).

CFSE proliferation assay

Proliferation assay was performed as described previously (8). CD8⁺ TILs were separated by MACS Column Technology and cultured in medium containing 50 IU/ml rhIL-2 for 2 d followed by CFSE labeling and stimulation with anti-CD3 antibodies and autologous non-CD3 cells obtained from the same tumor for 6 d. Cells were stained and analyzed by flow cytometry.

Intracellular cytokine staining assay and apoptosis studies

IVS assays were performed as described previously (8). Briefly, PBMCs were stimulated for 6 d in the presence of 10 μg/ml anti-IL-10R and/or anti-PD-1 (clone EH12.2H7, Biolegend) blocking mAb or isotype control antibodies prior to intracellular cytokine staining. Alternatively, cells were harvested and then surface stained with APC-labeled A2/NY-ESO-1 tetramers, and subsequently with CD8-PECy7 and Annexin V-FITC (ApoScreen Annexin V-FITC Apoptosis Kit, Beckman Coulter).

Statistics

Statistical hypotheses were tested with the Wilcoxon signed rank test (for paired results from the same patient) and unpaired t test (for IL-10 concentration) with GraphPad Prism statistical analysis program. Tests were 2-sided and considered significant for P < 0.05.

Results

IL-10R is upregulated by PD-1^high NY-ESO-1–specific CD8⁺ T cells

Using HLA-A2 (A2) tetramers (tet), we first investigated the expression of IL-10R and PD-1 on NY-ESO-1–specific, MART-1–specific, virus-specific [cytomegalovirus (CMV), Epstein-Barr virus (EBV) and influenza virus (Flu)] and total CD8⁺ T cells that are detectable ex vivo in PBMCs of nine HLA-A* 0201⁺ (HLA-A2⁺) stage IV melanoma patients. The percentages of IL-10R⁺ cells among NY-ESO-1–specific CD8⁺ T cells (mean 8.8% ± SD 2.8%) were significantly higher than those of MART-1–specific (2.4% ± 1.5%), Flu-specific (1.6% ± 1.3%), EBV-specific (3.4% ± 2.7%) and total tet^- (1.7% ± 1.0%) CD8⁺ T cells (Supplementary Fig. S1 A and B, left), albeit not significantly different from percentages of IL-10R⁺ CMV-specific CD8⁺ T cells (5.1% ± 4.2%). Similar observations were made in terms of mean fluorescence intensity (MFI; Supplementary Fig. S1 B, right). In agreement with previous findings (11, 29), IL-10R was also constitutively expressed by multiple cell subsets of PBMCs with higher expression detected on myeloid dendritic cells (mDC) and B cells (P < 0.01, Supplementary Fig. S1 C).

As previously shown, NY-ESO-1–specific CD8⁺ T cells expressed higher PD-1 levels (frequencies and MFI) than MART-1–specific and virus-specific CD8⁺ T cells present in PBMCs of patients with advanced melanoma (6). Notably, PD-1^high CD8⁺ T cells were detected in NY-ESO-1–specific CD8⁺ T cells but not in MART-1–specific, virus-specific and total tet^- CD8⁺ T cells (Fig. 1A and Supplementary Fig. S1 D). The percentages of IL-10R⁺ cells within PD-1^high NY-ESO-1–specific CD8⁺ T cells (mean 23.6% ± SD 8.1%) were significantly higher than within PD-1^int (4.2% ± 1.7%) and PD-1^low (1.8% ± 1.9%) NY-ESO-1–specific CD8⁺ T cells. Similar observations were made in terms of MFI (Fig. 1B, right).

Figure 1.

IL-10R is upregulated by PD-1^high NY-ESO-1–specific CD8⁺ T cells. A and B, Dot plots from one representative patient (A) and summary data for all nine patients with advanced melanoma (B) showing ex vivo IL-10R expression by PD-1^high and/or, PD-1^int and PD-1^low subsets of A2/NY-ESO-1 157–165, A2/MART-1 26–35, A2/CMV 495-503, A2/Flu-M 58–66, A2/EBV BMLF1 280–288 tet⁺ and total tet^- CD8⁺ T cells. * P < 0.05 and ** P < 0.01. Horizontal bars depict the mean percentage or MFI of IL-10R expression. Data shown are representative of two independent experiments performed in duplicate.

In agreement with previous studies (13, 30, 31), we detected higher levels of IL-10 in sera of advanced melanoma patients than in healthy donors (Supplementary Fig. S2 A). While multiple cellular subsets produced IL-10, monocytes and DCs expressed the highest IL-10 levels in both melanoma patients and healthy donors (Supplementary Fig. S2 B). Interestingly, in one patient with high NY-ESO-1–specific CD8⁺ T cell frequency, ex vivo sorted PD-1⁺ NY-ESO-1 tet⁺ CD8⁺ T cells produced IL-10 in the presence of cognate peptide-pulsed antigen presenting cells (APCs) (Supplementary Fig. S2 C), supporting that dysfunctional PD-1^high antigen-specific CD8⁺ T cells can produce IL-10 upon TCR activation (32).

Collectively, our results show that circulating PD-1^high NY-ESO-1–specific CD8⁺ T cells upregulate IL-10R expression and that IL-10 is produced at low levels by multiple cells in PBMCs of patients with advanced melanoma.

NY-ESO-1–specific CD8⁺ T cells upregulate IL-10R along with PD-1 upon TCR activation

To determine the activation status of IL-10R⁺ PD-1^high NY-ESO-1–specific CD8⁺ T cells, we evaluated their expression of HLA-DR, CD38 and CD57. IL-10R⁺ PD-1^high cells expressed higher levels of HLA-DR, CD38 and CD57 than IL-10R^- PD-1^high, IL-10R^- PD-1^int and IL-10R^- PD-1^low NY-ESO-1–specific CD8⁺ T cells, suggesting that they represent a highly activated NY-ESO-1–specific CD8⁺ T cell subset (Fig. 2A).

Figure 2.

NY-ESO-1–specific CD8⁺ T cells upregulate IL-10R and PD-1 upon TCR activation. A, Pooled data from melanoma patients (n = 8) showing the percentage (%) (left) and MFI (right) of ex vivo expression of HLA-DR, CD38 and CD57 on IL-10R⁺ and IL-10R^- subsets in PD-1^high and on IL-10R^- subsets in PD-1^int and PD-1^low NY-ESO-1–specific CD8⁺ T cells. B, IL-10R expression on NY-ESO-1 tet⁺ CD8⁺ T cells assessed ex vivo and after indicated hours following in vitro stimulation with cognate peptide (NY-ESO-1 peptide) or irrelevant peptide (HIV peptide), (n=6). C, The percentage of IL-10R⁺PD-1^high in NY-ESO-1 tet⁺ CD8⁺ T cells (left) and in total CD8⁺ T cells (right) tested in B. Data shown are representative of at least two independent experiments. * P < 0.05 and ** P < 0.01. Horizontal bars depict the mean percentage or MFI of cells that express the corresponding molecule.

To investigate whether IL-10R upregulation occurs upon TCR activation, we next assessed the sequential expression of IL-10R and PD-1 by spontaneous NY-ESO-1–specific CD8⁺ T cells present in PBMCs of patients with advanced melanoma over a 6-day in vitro stimulation (IVS) in the presence of cognate or irrelevant peptide. We observed a significant increase in IL-10R expression (% and MFI) by a fraction of NY-ESO-1–specific CD8⁺ T cells following cognate antigen stimulation from day 4 to day 6 of IVS (Fig. 2B). As previously reported, NY-ESO-1–specific CD8⁺ T cells also upregulated PD-1 expression upon stimulation with cognate peptide (8), with a majority of the T cells displaying a PD-1^high phenotype starting at day 4 (Supplementary Fig. S3 A and B). The percentages of IL-10R⁺ PD-1^high among NY-ESO-1–specific CD8⁺ T cells as well as the frequencies of IL-10R⁺ PD-1^high NY-ESO-1–specific CD8⁺ T cells among total CD8⁺ T cells increased upon stimulation with cognate peptide (Fig. 2C).

Collectively, our findings show that IL-10R⁺PD-1^high NY-ESO-1–specific CD8⁺ T cells represent a highly activated T cell subset that expands upon prolonged TCR activation.

NY-ESO-1–specific CD8⁺ T cells upregulate IL-10R expression upon PD-1 blockade

To investigate the effects of PD-1 blockade on IL-10R expression by total NY-ESO-1–specific CD8⁺ T cells, PBMCs of melanoma patients were stimulated with NY-ESO-1 peptide for 6 d in the presence of blocking monoclonal antibodies (mAbs) against PD-L1, IL-10R or isotype control antibodies, before tetramer and IL-10R or PD-1 labeling. We observed an increase in IL-10R expression by total NY-ESO-1 tet⁺, but not tet^- CD8⁺ T cells after incubation in the presence of cognate peptide and anti-PD-L1 antibodies, as compared to cognate peptide and IgG control antibodies (mean % IL-10R⁺: 35.3% ± SD 12.7% versus 25.0% ± 9.8% and mean MFI IL-10R: 537± SD 220 versus 394 ± 156, respectively) (Fig. 3A and B). In contrast, IL-10R blockade did not increase PD-1 expression by NY-ESO-1-specific CD8⁺ T cells (Supplementary Fig. S4 A and B).

Figure 3.

NY-ESO-1–specific CD8⁺ T cells upregulate IL-10R expression upon PD-1 pathway blockade. A and B, Dot plots from one representative patient (A) and summary data for eight patients with melanoma showing the percentage (%) (B, left) and MFI (B, right) of IL-10R expression by total A2/NY-ESO-1 157–165 tet⁺ and total tet^- CD8⁺ T cells after culture. PBMCs of melanoma patients were stimulated with NY-ESO-1 peptide in the presence of blocking mAbs against PD-L1 or isotype control antibodies, before tetramer and IL-10R labeling. * P < 0.05 and ** P < 0.01. Horizontal bars depict the mean percentage or MFI of IL-10R expression. Data shown are representative of two independent experiments performed in duplicate.

Collectively, our findings show that NY-ESO-1–specific CD8⁺ T cells upregulate IL-10R upon PD-1 blockade and prolonged antigen stimulation, whereas IL-10R blockade has no effect on PD-1 expression.

IL-10 impedes the expansion of NY-ESO-1–specific CD8⁺ T cells upon antigen stimulation

Since we have shown that IL-10R expression is upregulated by PD-1^high NY-ESO-1–specific CD8⁺ T cells and increases following TCR activation, we next investigated the effect of IL-10 on the expansion and survival of NY-ESO-1-specifc CD8⁺ T cells upon stimulation with cognate antigen. CFSE-labeled PBMCs from melanoma patients were stimulated for 6 d with NY-ESO-1 157-165 peptide in the presence or absence of rhIL-10 and/or anti-IL-10R mAbs. The frequencies of proliferating CFSE^lo NY-ESO-1–specific CD8⁺ T cells significantly decreased after IVS in the presence of IL-10 as compared to stimulation with peptide only (0.5-fold change) and were restored in the presence of blocking anti-IL-10R mAbs (Fig. 4A–C). The percentages of Annexin-V⁺ NY-ESO-1–specific CD8⁺ T cells increased after a 6-day IVS with cognate peptide in the presence of rhIL-10 (1.3-fold change), but not rhIL-10 in combination with anti-IL10R mAbs, as compared to IVS with peptide only (Fig. 4D). Notably, in the presence of rhIL-10, the expression of phosphorylated signal transducer and activator of transcription-3 (p-STAT3) in NY-ESO-1–specific CD8⁺ T cells was significantly higher than in EBV-specific CD8⁺ T cells, which express lower levels of IL-10R, as compared to untreated cells (3- and 1.6-fold increase, respectively). The effect of IL-10 on p-STAT3 expression was abolished in the presence of blocking anti-IL-10R mAbs, suggesting that IL-10 acts directly on T cells through IL-10R activation (Fig. 4E).

Figure 4.

IL-10 inhibits NY-ESO-1–specific CD8⁺ T cell proliferation after prolonged antigen stimulation. A and B, Dot plots from a representative patient (A) and summary for seven patients with melanoma (B) showing the variation in the frequencies of CFSE^lo NY-ESO-1–specific CD8⁺ T cells for 10⁶ CD8⁺ T cells. CFSE-labeled PBMCs from patients were incubated with NY-ESO-1 157–165 peptide and rIL-10 or blocking mAb against IL-10R (aIL-10R) plus rhIL-10 or an isotype control antibody (IgG) before the evaluation of A2/NY-ESO-1 157–165 tet⁺ CD8⁺ T cell proliferation by flow cytometry. C, Fold change of the frequencies of CFSE^lo NY-ESO-1–specific CD8⁺ T cells after IVS with cognate peptide and rhIL-10 or aIL-10R plus rhIL-10. The ratios of the percentages of CFSE^lo NY-ESO-1–specific CD8⁺ T cells in the presence of rhIL-10 or in the presence of aIL-10R plus rhIL-10 and IgG isotype control are shown. D, Pooled data showing the percentage of Annexin-V⁺ determined by flow cytometry for A2/NY-ESO-1 157–165 tet⁺ CD8⁺ T cells (n = 7) or A2/EBV BMLF1 280–288 tet⁺ CD8⁺ T cells (n = 4), which were incubated for 6 d with NY-ESO-1 157–165 peptide or BMLF1 280–288 peptide and rhIL-10 or aIL-10R plus rhIL-10 or IgG. E, p-STAT3 expression measured by flow cytometry in A2/NY-ESO-1 157–165 tet⁺ CD8⁺ T cells or A2/EBV BMLF1 280–288 tet⁺ CD8⁺ T cells from melanoma patients (n = 5) treated or not with aIL-10R and then stimulated or not with IL-10.

Collectively, our findings show that IL-10 impedes the expansion of NY-ESO-1–specific CD8⁺ T cells by decreasing their proliferative capacity and promoting their apoptosis. We also observed that IL-10 acts directly on IL-10R⁺ NY-ESO-1–specific CD8⁺ T cells.

IL-10R blockade adds to PD-1 blockade to increase the expansion and functions of NY-ESO-1–specific CD8⁺ T cells

We then investigated whether IL-10R blockade alone or in combination with PD-1 blockade increased NY-ESO-1–specific CD8⁺ T cell expansion and functions in response to cognate antigen. We observed a modest increase in the frequencies of proliferating (CFSE^lo) and total NY-ESO-1 tet⁺ CD8⁺ T cells in the presence of cognate peptide and anti-IL-10R mAbs as compared to cognate peptide and IgG control antibodies (1.5- and 1.3-fold changes, respectively; Fig. 5A and B and Supplementary Fig. S5 A and B). In line with previous findings (6, 7), PD-1 blockade significantly increased the frequencies of CFSE^lo and total NY-ESO-1–specific CD8⁺ T cells (1.7- and 1.4-fold changes, respectively). There was no significant difference between single IL-10R blockade and single PD-1 blockade on NY-ESO-1–specific CD8⁺ T cells proliferation. Interestingly, dual IL-10R and PD-1 blockade further increased the frequencies of proliferating and total NY-ESO-1–specific CD8⁺ T cells as compared to incubation with IgG control antibodies, anti-IL-10R mAbs alone or anti-PD-1 mAbs alone, resulting in the highest fold changes in the frequencies of CFSE^lo and total NY-ESO-1 tet⁺ CD8⁺ T cells (2.5- and 1.9-fold changes, respectively; Fig. 5A and B and Supplementary Fig. S5 A and B).

Figure 5.

IL-10R blockade adds to PD-1 blockade to increase the expansion and functions of NY-ESO-1–specific CD8⁺ T cells. A, Representative flow cytometry analysis from one melanoma patient showing the percentages of CFSE^lo NY-ESO-1–specific CD8⁺ T cells among total CD8⁺ T cells in CFSE-based proliferation assay (n = 9). B, Fold change of the frequencies of CFSE^lo (left) and total (right) NY-ESO-1–specific CD8⁺ T cells after IVS with cognate peptide and aPD-1 and/or aIL-10R. The ratio of the percentages of CFSE^lo and total NY-ESO-1–specific CD8⁺ T cells in the presence of indicated antibody treatment and isotype control antibody is shown. C, Representative flow cytometry analysis from one melanoma patient showing the percentages of IFN-γ- and TNF-producing NY0-ESO-1–specific CD8⁺ T cells among total CD8⁺ T cells (upper) and the percentages of IFN-γ⁺TNF⁺ NY-ESO-1–specific CD8⁺ T cells among NY-ESO-1–specific CD8⁺ T cells (lower). PBMCs were incubated with NY-ESO-1 157–165 peptide or with HIVpol 476–484 peptide and aPD-1 and/or aIL-10R or an isotype control antibody (IgG), prior to evaluating intracellular cytokine production of NY-ESO-1–specific CD8⁺ T cells upon stimulation with cognate peptide, (n=8). D, Fold change of the frequencies of IFN-γ ⁺, TNF⁺ and IFN-γ⁺TNF⁺ NY-ESO-1–specific CD8⁺ T cells after IVS with cognate peptide and aPD-1 and/or aIL-10R. The ratio of the frequency of cytokine-producing NY-ESO-1–specific T cells from melanoma patients in the presence of indicated antibody treatment and isotype control antibody is shown. Data shown are representative of two independent experiments performed in duplicate.

The frequencies of NY-ESO-1–specific CD8⁺ T cells that produced IFN-γ, but not TNF, increased after a 6-day IVS in the presence of cognate peptide and anti-IL-10R mAbs as compared to incubation with IgG control antibodies (1.3-fold change; Fig. 5C and D upper panels and Supplementary Fig. S5 C and D). PD-1 blockade alone increased the frequencies of IFN-γ- and TNF-producing NY-ESO-1–specific CD8⁺ T cells (1.6- and 1.5-fold changes, respectively). Dual IL-10 and PD-1 blockade further increased the frequencies of IFN-γ- and TNF-producing NY-ESO-1–specific CD8⁺ T cells (2.1- and 1.9-fold changes, respectively; Fig. 5C and D upper panels and Supplementary Fig. S5 C and D). In addition, we observed an increase in the frequencies of IFN-γ⁺ TNF⁺ NY-ESO-1–specific CD8⁺ T cells (1.9 fold change, Fig. 5C and D lower panels), suggesting that dual IL-10R and PD-1 blockade expands polyfunctional TA-specific CD8⁺ T cells.

Collectively, our findings show that IL-10 blockade adds to PD-1 blockade to enhance the expansion and functions of NY-ESO-1–specific CD8⁺ T cells.

CD8⁺ TILs present in metastatic melanoma co-upregulate IL-10R and PD-1

To determine whether our findings on circulating TA-specific CD8⁺ T cells were relevant to CD8⁺ TILs present in the TME, we assessed IL-10R and PD-1 expression by CD8⁺ TILs obtained from metastatic lesions of nine melanoma patients. The percentages of IL-10R⁺ cells were significantly higher in CD8⁺ TILs (mean 30.0% ± SD 19.6%) than in total NY-ESO-1 tet⁺ (8.8% ± 2.8%) and tet^- (1.7% ± 1.0%) CD8⁺ T cells present in PBMCs of melanoma patients (Fig. 6A and B left). Similar observations were made in terms of MFI (Fig. 6B right). CD8⁺ TILs also upregulated PD-1 expression and the percentages of IL-10R⁺ cells within PD-1^high CD8⁺ TILs (mean 39.8% ± SD 24.4%) were significantly higher than within PD-1^int (29.2% ± 27.6%) and PD-1^low (20.7% ± 23.4%) CD8⁺ TILs. Similar observations were made in terms of MFI (Fig. 6C and D).

Figure 6.

IL-10R is highly upregulated by PD-1⁺ CD8⁺ TILs. A and B, Dot plots from one representative patient (A) and summary data (B) showing ex vivo IL-10R expression by NY-ESO-1 tet^- CD8⁺ T cells from PBMCs of healthy donors (n=9) and by CD8⁺ TILs from melanoma patients (n=9). C and D, Dot plots from one representative patient (C) and summary data for all nine patients with advanced melanoma (D) showing ex vivo IL-10R expression by PD-1^high, PD-1^int and PD-1^low subsets of CD8⁺ TILs. E, Flow cytometry analysis from one melanoma patient showing the percentages of CFSE^lo CD8⁺ TILs among total CD8⁺ TILs. CFSE-labeled CD8⁺ TILs were incubated with anti-CD3-pulsed non-T cell fraction of one melanoma tumor single cell suspension in the presence of anti-IL-10R and/or anti-PD-1 or IgG control antibodies before the evaluation of the CD8⁺ T cells proliferation by flow cytometry. * P < 0.05 and ** P < 0.01. Horizontal bars depict the mean percentage or MFI of IL-10R expression by NY-ESO-1 tet^- CD8⁺ T cells or CD8⁺ TILs. Data shown are representative of two independent experiments performed in duplicate.

To investigate whether IL-10 blockade alone or in combination with PD-1 blockade increases the expansion of CD8⁺ TILs, CD8⁺ T cells isolated from one metastatic tumor single cell suspension were incubated in vitro for 6 d in the presence of anti-CD3 antibodies and autologous non-CD3 cells obtained from the same tumor. We observed an increase in the proliferation of CD8⁺ TILs in the presence of anti-IL-10R or anti-PD-1 antibodies as compared to IgG control antibodies. Dual IL-10R and PD-1 blockade further increased the frequencies of proliferating CD8⁺ TILs (Fig. 6E).

Collectively, our findings show that CD8⁺ TILs present in metastatic melanoma co-upregulate PD-1 and IL-10R and that dual IL-10 and PD-1 blockade promotes the expansion of CD8⁺ TILs.

Discussion

In this manuscript, we report that PD-1^high TA-specific CD8⁺ T cells in PBMCs and TILs of patients with advanced melanoma upregulate IL-10R expression. Among PD-1⁺ antigen-specific CD8⁺ T cells, it is now well established that PD-1^high CD8⁺ T cells represent a dysfunctional T cell subset occurring upon persistent antigen stimulation in chronic infections and in cancer (4, 5, 8, 33, 34). TA-specific dysfunctional CD8⁺ T cells present at periphery and at tumor sites upregulate a number of inhibitory receptors, including PD-1, BTLA and Tim-3 that bind to their respective ligands expressed by APCs and tumor cells in the TME (7, 8, 35). Here, we show that TA-specific CD8⁺ T cells increase IL-10R expression upon TCR activation and PD-1^highIL-10R⁺ cells upregulate HLA-DR and CD38, supporting that IL-10R is a T cell activation marker expressed by chronically activated TA-specific CD8⁺ T cells in patients with advanced melanoma.

IL-10 is an immunoregulatory cytokine, which is produced by multiple innate and adaptive immune cells as well as melanoma cells (11). IL-10 inhibits antigen-specific T cell responses by impeding antigen-presentation and costimulation, inducing human CD4⁺ T cell anergy (36, 37). A number of observations have also suggested that IL-10 may exert immunostimulatory effects on CD8⁺ T cells depending on their state of activation (38). Elevated IL-10 levels have been reported in patients with HCV, HBV and HIV infections and correlate with T cell dysfunction and uncontrolled viral replication (17, 39-41).

The role of IL-10 in cancer immunology remains controversial. IL-10 transgenic mice, which produce moderate and transient levels of IL-10 upon TCR activation, are unable to control tumor growth (18, 19), supporting the immunosuppressive and protumoral effects of IL-10. Paradoxically, the administration of high-dose IL-10 and pegylated IL-10 appears to promote tumor regression in animals, and enhances the expansion and functions of CD8⁺ TILs that express elevated levels of IL-10R, resulting in tumor regression (20, 25, 26). In addition, chemically-induced transplanted tumors appear to grow better in IL-10^-/- mice (24). In combination with cancer vaccines, and depending on the schedule of injection, high-dose IL-10 either promotes T cell-mediated tumor rejection or tumor progression, illustrating the two opposite effects of IL-10 on T cells in vivo depending on their cell activation state (23). In sharp contrast with the experimental models supporting the antitumoral activity of high-dose IL-10, we measured low level circulating IL-10 in melanoma patients' sera. While multiple immune cells in PBMCs of patients with advanced melanoma produce low-level IL-10, we found that the highest IL-10 levels are produced by circulating CD11c⁺ and CD14⁺ cells in agreement with a previous study (40). In one patient with a high frequency of NY-ESO-1–specific CD8⁺ T cells that upregulate PD-1, we observed that TA-specific CD8⁺ T cells produce IL-10. These findings are in line with previous reports of IL-10 production by exhausted/dysfunctional LCMV-specific CD8⁺ T cells that upregulate inhibitory receptors in chronic viral infections (32, 41).

In the present study, we show that IL-10 decreases TA-specific CD8⁺ T cell proliferation upon prolonged stimulation with cognate antigen and increases TA-specific CD8⁺ T cell apoptosis. Furthermore, upon IL-10 exposure, p-STAT3 expression in NY-ESO-1–specific CD8⁺ T cells was significantly higher than in EBV-specific CD8⁺ T cells that do not upregulate PD-1 and IL-10R, suggesting that IL-10 exerts its immunosuppressive effects by acting directly on IL-10R⁺ TA-specific CD8⁺ T cells. Collectively, our data support that endogenous IL-10 produced at low levels by multiple innate and adaptive immune cells impedes the proliferation and survival of chronically activated TA-specific CD8⁺ T cells in patients with advanced melanoma. Notably, IL-10 production by circulating CD14⁺ cells of patients with advanced melanoma appears to correlate with poor clinical outcome (40).

The immunosuppressive role of endogenous IL-10 on T cells has served as a rationale for neutralizing IL-10 to stimulate potent antigen–specific T cell responses in experimental models. Neutralization of IL-10 with anti-IL10R mAbs in combination with LPS renders soluble antigen capable of stimulating potent Th1 type T cell responses in vivo (16). IL-10 blockade improves virus-specific T cell function in mice chronically infected by LCMV virus (42) and dual IL-10 and PD-1 blockade appears more effective in restoring virus-specific T cell function and controlling persistent viral infections (27). Monocytes isolated from HIV patients upregulate PD-1 expression, produce IL-10 upon PD-1 ligation in vitro to impair CD4⁺ T cell activation and dual IL-10 and PD-1 blockade restores T cell responses to HIV in vitro. (28). In cancer immunology, IL-10 blockade in combination with intratumoral injection of CPG reverses tumor-infiltrating dendritic cell dysfunction to prime potent antitumor T cell responses leading to tumor regression in mice (17). One novel finding in this manuscript is that IL-10 blockade adds to PD-1 blockade to increase the proliferation and functions of TA-specific CD8⁺ T cells isolated from PBMCs and TILs. Most interestingly, we observed that TA-specific CD8⁺ T cells further upregulate IL-10R upon PD-1 blockade. It is therefore tempting to speculate that upon PD-1 blockade, activated TA-specific CD8⁺ T cells may be rendered more sensitive to the immunosuppressive effects of endogenous IL-10. Therefore, IL-10 blockade may represent a potent therapeutic strategy to counteract the direct negative regulatory effects of IL-10 on T cells in the tumor microenvironment, which could further increase antitumor T cell responses in combination with PD-1 blockade.

In summary, our findings demonstrate the upregulation of IL-10R by PD-1^high TA-specific CD8⁺ T cells at periphery and at tumor sites in patients with advanced melanoma. They also show that IL-10 blockade adds to PD-1 blockade to further enhance the expansion and functions of TA-specific CD8⁺ T cells. Such findings may have significant implications in terms of potent immunotherapy of melanoma to further improve the clinical efficacy of PD-1 blockade in the clinic.