CD8+ T cells can be polarized into several different subsets as defined by the cytokines they produce and the transcription factors that govern their differentiation. Here, we identified the polarizing conditions to induce an IL22-producing CD8+ Tc22 subset, which is dependent on IL6 and the aryl hydrocarbon receptor transcription factor. Further characterization showed that this subset was highly cytolytic and expressed a distinct cytokine profile and transcriptome relative to other subsets. In addition, polarized Tc22 were able to control tumor growth as well as, if not better than, the traditional IFNγ-producing Tc1 subset. Tc22s were also found to infiltrate the tumors of human patients with ovarian cancer, comprising up to approximately 30% of expanded CD8+ tumor-infiltrating lymphocytes (TIL). Importantly, IL22 production in these CD8+ TILs correlated with improved recurrence-free survival. Given the antitumor properties of Tc22 cells, it may be prudent to polarize T cells to the Tc22 lineage when using chimeric antigen receptor (CAR)-T or T-cell receptor (TCR) transduction–based immunotherapies.

The typical CD8+ T cell is regarded as a highly cytotoxic cell type that produces large amounts of IFNγ (1). Studies over the past decade challenge this notion by identifying distinct lineages of CD8+ T cells, each producing a unique profile of cytokines and transcription factors, as well as displaying varying cytolytic capacities. Studies have identified IFNγ-producing Tc1s, IL4+ Tc2s (2–5), IL9+ Tc9s (6, 7), and IL17+ Tc17s (8–12). Each Tc subset differs in their killing capacity, with Tc1s being highly cytotoxic and Tc17s being poorly cytotoxic (10). The polarization of Tc subsets is mediated by the same cytokines as their CD4+ Th counterparts, involving different combinations of IL4, IL6, IL12, and TGFβ (13). Indeed, this also leads to the utilization of many of the same signaling pathways and transcription factors as their Th equivalents. For example, Tc17 polarization is dependent on the transcription factors RORγT, IRF4, and STAT3 (10, 14, 15). A variety of Tc subsets are generated in vivo during immune responses to infections and autoimmune diseases (6, 9, 15) and many different Tc subsets infiltrate mouse and human tumors (12, 16). However, the precise physiologic role of these Tc subsets remains unclear.

IL22 is a member of the IL10 family of cytokines that acts on cells expressing IL22R1, namely epithelial cells, keratinocytes, hepatocytes, and pancreatic β cells. IL22 helps to maintain the epithelial barrier by promoting tissue repair and wound healing, and to induce antimicrobial peptides and proinflammatory cytokines (17). In some instances, IL22 inhibits tumor growth (18, 19), but IL22 is generally considered to be protumorigenic by promoting inflammation as well as tumor cell proliferation and survival (20–22). IL22 is mainly produced by cells of lymphoid origin, including innate lymphoid cells and T cells (23). A novel subset of human CD8+ IL22+ Tc22s was identified after examining the cytokine profile of CD8+ T cells infiltrating psoriatic and atopic dermatitis lesions (24–26). In addition to infiltrating inflamed skin lesions, Tc22s arise in response to HIV (27) and IL22-producing CD8+ T cells infiltrate squamous cell and hepatocellular carcinomas (28, 29). Although the polarizing conditions for Tc22s are not defined, IL21 can induce IL22 production in CD8+ T cells (30). Regardless, it is unclear whether these cells are a distinct Tc22 lineage, thus we sought to further characterize the various Tc subsets, particularly Tc22s.

Mice and cell lines

C57BL/6, AhR floxed with CD4 cre (referred to as AhR −/−) and Tbet −/− mice were purchased from the Jackson Laboratory and Taconic. Generation of P14 mice, which express a transgenic T-cell receptor (TCR) specific for the gp33 peptide of the lymphocytic choriomeningitis virus (LCMV) and H-2Db, was described previously (31). EL4 cells were obtained from Dr. Rolf M. Zinkernagel (University of Zurich, Zurich, Switzerland). B16F10-gp33 cells were obtained from Dr. Rolf M. Zinkernagel (University of Zurich). These cells contained a neomycin resistance gene and were cultured with G418 selection reagent (800 μg/mL, InVivogen) for 7 days prior to freezing. Cells were used for tumor inoculation after 1 to 2 passages post-thaw. For both cell lines, no cell line authentication occurred in the past year and no Mycoplasma testing was performed. All mice were maintained at the Ontario Cancer Institute animal facility according to institutional guidelines and with approval of the Ontario Cancer Institute Animal Ethics Committee.

Tumor experiments

Eight- to 12-week-old B6 mice were inoculated subcutaneously with 4 × 105 B16F10-gp33 cells resuspended in 100 μL HBSS. Ten to 11 days later, mice bearing tumors of approximately 5 mm diameter were randomly allocated to different treatment groups, some of which received 1 × 106 polarized CD8+ P14 T cells in 200 μL of HBSS injected intravenously (i.v.) via the tail vein. Tumor size was continually assessed using calipers until mice reached experimental endpoint (diameter ≥1.5 cm or severe ulceration/necrosis). No blinding occurred.

Tc subset polarization

CD8+ T cells were magnetically purified with a Negative Selection, Magnetic CD8+ T Cell Isolation Kit (Miltenyi Biotec, catalog no. 130-104-075) according to the manufacturer's instructions. Spleens and lymph nodes of P14 and wild-type (WT) mice were mashed through a 70-μm strainer, incubated with selection beads, and passed through a magnetic column to isolate untouched CD8+ T cells. These CD8+ T cells were subsequently cocultured either with mature bone marrow–derived dendritic cells (BMDC) pulsed with gp33 peptide (1 μmol/L) from LCMV (KAVYNFATM; ref. 32) or cultured with mature BMDCs + α-CD3 (145-2C11, 1 μg/mL) for three days in IMDM supplemented with 10% FCS, l-glutamine (2 mmol/L), β-mercaptoethanol (50 μmol/L), penicillin and streptomycin (100 U/mL). BMDCs were generated by culturing BM cells isolated from the legs of WT mice for 7 to 9 days in the presence of 40 ng/mL GM-CSF, with a refresh of the media on day 3 and 6 (32). On day 7, 8, or 9, BMDCs were matured overnight with LPS (100 ng/mL, Invivogen). To generate Tc subsets, polarizing cocktails were added at the start of the coculture as follows: Tc0, no additional cytokines; Tc1, IL12 (5 ng/mL); Tc2, IL4 (20 ng/mL) + anti-IFNγ (XMG1.2, 10 μg/mL); Tc9, IL4 (10 ng/mL) + TGFβ1 (10 ng/mL) + anti-IFNγ (10 μg/mL); and Tc17, IL6 (20 ng/mL) + TGFβ1 (3 ng/mL) + IL23 (10 ng/mL) + anti-IFNγ (10 μg/mL). Tc22s were polarized with IL6 (20 ng/mL) + TNFα (40 ng/mL) + 6-Formylindolo(3,2-b)carbazole (FICZ at 2 ng/mL) + anti-IFNγ (10 μg/mL) + anti-TGFβ (1D11.16.8, 10 μg/mL). On day 3 poststimulation, cells were stained for flow cytometry or used for functional assays. Purity of cells used for downstream assays were >98% CD8+ as determined by flow cytometry staining. All cytokines were purchased from BioLegend. Anti-IFNγ was purchased from BioLegend, while anti-TGFβ was purchased from eBioscience and R&D Systems. FICZ and CH-223191 (used at 500 ng/mL) were purchased from Enzo Life Sciences and EMD Millipore, respectively.

Flow cytometry and antibodies

For intracellular cytokine staining, cells were restimulated for 5 to 6 hours with Cell Stimulation Cocktail (eBioscience, catalog no. 00-4970-03) in the presence of Brefeldin A (eBioscience, catalog no. 00-4506-51) both used according to the manufacturers' recommended concentrations. Cells were then stained for surface markers in FACS buffer (PBS + 2% FCS) for 30 minutes and fixed using Cytofix/Cytoperm (BD Pharmingen, catalog no. BDB554714) or Fixation/Permeabilization Buffer Set (eBioscience, catalog no. 88-8824-00) according to manufacturers' recommended protocols. Intracellular cytokine staining was then performed for 30 minutes in permeabilization buffer. Cells were washed in FACS buffer or permeabilization buffer following every staining/fixation step according to the manufacturer's recommended protocol. All antibodies were used according to manufacturers' recommended concentrations. In some instances, cells were incubated with Fc block (BD Biosciences, catalog no. 553141) for 10 minutes prior to surface staining according to the recommended protocol, or a fixable viability dye (eBioscience, catalog no. 65-0866-14) at 1:1,000 for 30 minutes after surface staining prior to fixation. Phosphoflow was performed using BD Phosflow Perm Buffer III (BD Biosciences, catalog no. BDB558050) on T cells that had been stimulated for 30 minutes in polarizing conditions according to the manufacturer's recommended protocol. Flow cytometry data was acquired on a FACSCanto II (BD Biosciences) or LSR Fortessa (BD Biosciences) and analyzed using FlowJo software (Tree Star). A sample gating strategy for human and mouse T cells can be found in Supplementary Fig. S1. Antibodies for flow cytometry were purchased from eBioscience, BioLegend, R&D Systems, and BD Pharmingen. Antibody clones used for flow cytometry were from: eBioscience [CD8 (53–6.7), IFNγ (XMG1.2), IL22 (IL22JOP), IL17 (ebio17B7), TNFα (MP6-XT22), IL4 (BVD6-24G2), Tbet (4B10), AhR (4MEJJ), CD62L (MEL-14), CD44 (IM7), KLRG1 (2F1), CCR7 (4B12), CD25 (PC61), CD127 (A7R34), IL18R (P3TUNYA), 4-1BB (17B5), OX40 (OX-86), CD30 (mCD30.1), GITR (DTA-1), ICOS (7E.17G9), CD86 (GL-1), SLAM (mShad150), CD101 (Moushi 101), Granzyme B (GB12), Perforin (eBioOMAK-D), FasL (MFL3), TRAIL (N2B2)]; BioLegend [IL22 (Poly5164), IL9 (RM9A4), CCR4 (2G12), CCR6 (29-2L17), CD27 (LG.3A10)]; BD Biosciences [IL2 (JES6-5H4), TNFR2 (TR75-89), pSTAT3 (4/P-STAT3)]; and R&D Systems [CCR10 (248918)]. Antibody clones for human T-cell staining used were CD3 (OKT3), CD4 (RPA-T4), CD8 (RPA-T8), IL22 (22URTI), IL17 (ebio64DEC17), and IFNγ (4S.B3), all purchased from eBioscience. Antibodies used for mitochondrial protein staining were the total OXPHOS antibody cocktail (Abcam, catalog no. ab110413) used according to the recommended protocol.

RNA sequencing

RNA was extracted from day 3 polarized Tc subsets (>99% CD8+ purity) using RNeasy Mini Kit (Qiagen, catalog no. 74104) according to manufacturer's instructions. RNA samples were quantified by Qubit BR RNA kit (Thermo Fisher Scientific, catalog no. Q10210) and quality assessed by Agilent Bioanalyzer RNA Nano 6000 Kit (Agilent, catalog no. 5067-1511) to ensure all samples had RIN values greater than 9. RNA libraries were prepared following the Illumina TruSeq Stranded Total RNA Reference workflow (catalog no. 1000000040499) using the TruSeq Stranded Total RNA kit (Illumina, catalog no. 20020596). Two-hundred nanograms of each RNA sample was depleted of ribosomal RNA using Ribo-zero Gold rRNA beads (Illumina, catalog no. 20020599), followed by fragmentation with kit components according to the manufacturer's instructions. First-strand cDNA product was generated from the depleted and fragmented RNA fragments using SuperScript II Reverse Transcriptase (Thermo Fisher Scientific, catalog no. 18064014) with the included random hexamer primers. Second-strand cDNA synthesis was performed using the RNase H and DNA Polymerase I kit components as instructed followed by Ampure XP–mediated purification as instructed (Beckman Coulter Genomics part no. A63881). Double stranded cDNA ends were 3′ adenylated and ligated with Illumina TruSeq dual index adapters (Illumina, catalog no. 20019792) with included components as instructed. Ligation product was purified with Ampure XP beads, followed by library PCR amplification, and further Ampure XP purification with included kit components as instructed. Final cDNA libraries were size validated using the Agilent Technologies 2100 Bioanalyzer with a DNA 1000 chip according to manufacturer's instructions. Final cDNA library concentrations were validated by qPCR using a Kapa KK4873 library quantification kit (Kapa Biosystems, catalog no. 07960336001) on a Bio-Rad CFX96 according to the manufacturer's instructions. All libraries were normalized to 10 nmol/L, pooled, denatured with 0.2 N NaOH, and diluted to a final concentration of 1.4 pmol/L. 1.3 mL of 1.4 pmol/L pooled libraries were loaded in to an Illumina NextSeq cartridge for cluster generation and sequenced on an Illumina NextSeq 500 instrument using the paired-end 75-bp protocol to achieve approximately 40 million reads per sample according to the manufacturer's instructions. FASTQ files were mapped using STAR v2.5.2 aligner (33) with transcript coordinates from GENCODE release M19 (34) using mouse genome GRCm38 (mm10). Reads were summarized per transcript using RSEM v1.3.0 (35) using the Gencode transcript coordinates. The Bioconductor (36) biomaRt (37) package was used to map the original Ensembl transcript IDs to NCBI Gene IDs and Unigene symbols. Following exploratory data visualization, transcripts with total read counts less than 33 across 18 samples were removed before further analysis, leaving 18,556 transcripts for further analysis. Principal component analysis (PCA) was performed on RNA-sequencing (RNA-seq) count values normalized using the variance-stabilizing transformation (38). PCA used transcripts with above-median SDs across samples. DESeq2 (39) differential expression was performed between all pairs of Tc subsets by specifying the Tc subset as the variable of interest and extracting DESeq2 model predictions for comparisons between all pairs of Tc subsets. P values were adjusted for multiple testing using the FDR correction of Benjamini and Hochberg (40). DESeq2 log2-fold change predictions were moderated using the APEGLM algorithm (41) to shrink large but highly variable fold-change values predominantly occurring for transcripts with low read counts. Prior to gene set enrichment analysis (GSEA), mouse-to-human gene homolog mapping was performed using biomaRt, and multimapping putative homologs resolved to produce a unique list of human genes, resulting in a ranked log2-fold change list of 12,754 genes. GSEA v3.0 was used with MSigDB v6.2 gene sets (42) on the Tc22 versus Tc17 DESeq2 differential expression analysis results ranked by log2-fold change to produce gene sets significantly (FDR-adjusted P < 0.1) upregulated in either Tc22 or Tc17. The Cytoscape v3.7.1 (43) EnrichmentMap v3.2.0 (44) plugin was used to visualize and cluster (using GLay network clustering algorithm; ref. 45) GSEA results into network clusters of partially overlapping gene sets. Cluster annotations were generated manually following examination of gene sets in each cluster. Clusters of 3 or fewer connected gene sets were excluded for clarity. All other RNA-seq visualizations were created using the R ggplot2 (46) package.

Real-time PCR

RNA was extracted from cells pellets using RNeasy Plus kit (Qiagen, catalog no. 74136) following the manufacturer's instructions and measured by Nanodrop2000. A total of 1 μg of RNA was reverse transcribed using qScript cDNA Super Mix (Quanta, catalog no. 95048), and gene expression was quantified by real-time PCR using 10 ng for each reaction with the PerfeCTa SYBR Green FastMix (Quanta, catalog no. 95073). The reactions were run in triplicate on the Applied Biosystems 7900HT using the recommended fast two-step cycle protocol. Gene expression for all experiments were normalized to the housekeeping gene GAPDH and expressed as fold change relative to Tc0 calculated using comparative Ct method (ΔΔCt). Primer sequences can be found in Supplementary Table S1.

Cytotoxicity assay

The cytotoxicity assay was performed as described previously (6). Briefly, EL4 cells were pulsed with gp33 peptide from LCMV or a control adenovirus (AV) peptide (SGPSNTPPEI) for 2 hours. Gp33-pulsed cells were labeled with 10 μmol/L CFSE and AV-pulsed EL4 cells were labeled with 1 μmol/L CFSE and mixed together at a 1:1 ratio. The mixture of EL4 cells were incubated with polarized Tc subsets expressing the P14 transgenic TCR for approximately 5 hours and killing was assessed by measuring the ratio of high CFSE–expressing cells to low CFSE–expressing cells by flow cytometry.

Cytokine quantification and ATP

Polarized Tc subsets were restimulated for 24 hours with anti-CD3 (1 μg/mL). Supernatants were collected after 24 hours and cytokines were quantified using LEGENDplex (BioLegend, catalog no. 740741) or by ELISA (eBioscience, catalog no. 5017319) according to the manufacturer's recommended protocol. ATP was determined using an ATP quantification kit (Sigma, catalog no. MAK135-1KT) and the FlexStation3 plate reader (Molecular Devices) according to the recommended protocols.

Human tissue and blood specimens

Fresh tissue from patients with ovarian cancer undergoing standard-of-care surgical procedures (n = 27) were obtained from the UHN Biospecimen Program. Peripheral blood mononuclear cells (PBMC) were obtained from healthy donors (n = 3). All human tissue and blood were obtained through protocols approved by the institutional review board. Written informed consent was obtained from all donors.

Human Tc22 polarization

Fresh or cryopreserved PBMCs from healthy donors were magnetically sorted (Miltenyi Biotec, catalog no. 130-097-095) for naïve T cells (>96% pure) according to the manufacturer's recommended protocol. T cells were seeded into a 96-well plate previously coated with 5 μg/mL anti-CD3 (eBioscience, clone OKT3). To induce Tc22 polarization, the following antibodies and cytokines were added to culture: 1 μg/mL anti-CD28 (eBioscience, clone CD28.2), 5 μg/mL anti-IFNγ (BioLegend, clone B27), 5 μg/mL anti-IL4 (BioLegend, clone 8D4-8), 5 μg/mL anti-TGFβ (eBioscience, clone 1D11.16.8), 20 ng/mL IL6, 10 ng/mL IL21, 10 ng/mL IL23, 40 ng/mL TNFα, and 2 ng/mL FICZ (Enzo Life Sciences). Five days later, cells were stimulated with PMA/ionomycin (eBioscience) + Brefeldin A (eBioscience) for 5–6 hours. Cells were then stained and analyzed for intracellular cytokines by flow cytometry as described above.

Ovarian tumor-infiltrating lymphocyte staining

Ovarian cancer tumor-infiltrating lymphocytes (TIL) were expanded from tumor fragments or digests in IL2 as described previously (47). Briefly, ovarian tumors were cut into approximately 1 mm3 fragments. Tissue fragments were either cultured directly in 24 well plates or enzymatically digested in IMDM (Lonza) containing 1 mg/mL collagenase (Sigma) and 10 μg/mL pulmozyme (Roche). Cells or fragments were expanded in complete medium comprised of IMDM (Lonza), 10% human serum (Gemini), 25 mmol/L HEPES (Lonza), 100 μg/mL penicillin/streptomycin (Lonza), 10 μg/mL gentamycin (Lonza), 55 μmol/L 2-mercaptoethanol (Invitrogen), 2 mmol/L l-glutamine (Lonza), and 1,000 CU/mL IL2 (Proleukin, Novartis). Expanded ovarian TILs were cryopreserved in freezing medium containing 10% Cryoserv DMSO (Mylan Institutional) and 90% human serum (Gemini). Cryopreserved ovarian TILs were thawed and rested for several days in complete IMDM (Hyclone) supplemented as described above. TILs were then restimulated with PMA/ionomycin (eBioscience) + Brefeldin A (eBioscience) for 5–6 hours. Cells were subsequently stained and analyzed for intracellular cytokines by flow cytometry as described previously. For the recurrence-free survival (RFS) curves, only patients with stage IIIC high-grade serous cancer were included as outlined in Supplementary Table S2.

Statistical analysis

Statistical analysis was performed as described in each figure legend. Statistical tests used include one-way ANOVA, a two-way repeated measures ANOVA, and a log-rank test. In some instances, multiple t tests were used with a Holm–Sidak correction to control for multiple comparisons. Results were considered statistically significant when P < 0.05. Statistics were calculated using Graphpad Prism Version 8.

IL6 is essential for Tc22 polarization

To establish our tissue culture conditions used to polarize different CD8+ Tc subsets, we stimulated CD8+ T cells in the presence of the previously defined polarizing conditions (13) to induce Tc1s, Tc2s, Tc9s, and Tc17s, or without any additional cytokines to generate Tc0s (Fig. 1A). Using the polarizing conditions for CD4+ Th22s as our starting point (48, 49), we were able to induce a population of IL22+ Tc22s that minimally expressed IL17 (Fig. 1A). Tc22 polarization was dependent on IL6, and further enhanced when combined with TNFα and the aryl hydrocarbon receptor (AhR) agonist 6-Formylindolo (3,2-b) carbazole (FICZ), in conjunction with neutralizing antibodies for IFNγ and TGFβ (Fig. 1B; Supplementary Fig. S2). Together, these findings indicate that IL6 is the driving cytokine for Tc22 polarization.

Figure 1.

IL6 drives the polarization of CD8+ IL22-producing Tc22s. A, Cytokine production was assessed in CD8+ T cells that were cocultured with LPS-matured BMDCs for 3 days with the following polarizing conditions: Tc0, no cytokines; Tc1, IL12; Tc2, IL4 + anti-IFNγ; Tc9, IL4 + TGFβ + anti-IFNγ; Tc17, IL6 + IL23 + TGFβ + anti-IFNγ; Tc22, IL6 + TNFα + FICZ + anti-TGFβ + anti-IFNγ. Results shown are representative FACS plots gated on CD8+ T cells. B, CD8+ T cells were stimulated with the indicated cytokines and neutralizing antibodies. When IL6 or TNFα were omitted from the culture conditions, a corresponding cytokine neutralizing antibody was added instead. Graphs represent mean percentage of IL22+ IL17 or of IL22 IL17+ T cells ± SEM of technical replicates. Data shown are representative of at least two to three independent experiments.

Figure 1.

IL6 drives the polarization of CD8+ IL22-producing Tc22s. A, Cytokine production was assessed in CD8+ T cells that were cocultured with LPS-matured BMDCs for 3 days with the following polarizing conditions: Tc0, no cytokines; Tc1, IL12; Tc2, IL4 + anti-IFNγ; Tc9, IL4 + TGFβ + anti-IFNγ; Tc17, IL6 + IL23 + TGFβ + anti-IFNγ; Tc22, IL6 + TNFα + FICZ + anti-TGFβ + anti-IFNγ. Results shown are representative FACS plots gated on CD8+ T cells. B, CD8+ T cells were stimulated with the indicated cytokines and neutralizing antibodies. When IL6 or TNFα were omitted from the culture conditions, a corresponding cytokine neutralizing antibody was added instead. Graphs represent mean percentage of IL22+ IL17 or of IL22 IL17+ T cells ± SEM of technical replicates. Data shown are representative of at least two to three independent experiments.

Close modal

Tc22 polarization is inhibited by T-bet and facilitated by AhR

Transcription factors such as T-bet, RORγt (RORC), and AhR play a critical role in driving CD4+ Th polarization (50). Each defined Th subset has one or more “master regulator” transcription factors that are essential for their differentiation. When looking at the CD8+ Tc subsets, we found that they differentially express many of these transcription factors in a manner similar to their CD4+ counterparts (Fig. 2A). Tc1s had the highest T-bet transcripts, Tc2s had the highest GATA3 transcripts, and Tc17s had the highest expression of RORγt.

Figure 2.

Tc22 polarization is inhibited by T-bet and facilitated by AhR. A, Expression of various transcription factors in day 3 polarized Tc subsets was quantified by RT-PCR and expressed as fold change relative to Tc0 ± SEM (n = 4). B, AhR and T-bet expression in Tc0, Tc1, and Tc22s on day 3 post-activation as determined by flow cytometry. C and D, WT and T-bet−/− CD8+ T cells were stimulated under polarizing conditions for 3 days. Cytokine production was assessed in Tc22-polarized cells (C) and other Tc subsets (D) and expressed as mean percentage of positive cells of technical replicates ± SEM. E, WT and AhR −/− CD8+ T cells were stimulated in Tc22-polarizing conditions for 3 days, and cytokine production was assessed. F, CD8+ T cells were stimulated in the presence of Tc22-polarizing conditions without FICZ. Either an AhR antagonist (CH-223191), an AhR agonist (FICZ), or vehicle control was added at the start of the culture, and cytokine production was assessed 3 days later. G, Percentage of IL22+ CD8+ T cells after 3-day stimulation in the indicated polarizing conditions in conjunction with CH-223191, vehicle control, or FICZ. Values were expressed as mean percentage of IL22+ from technical replicates ± SEM. All FACS plots shown were gated on CD8+ T cells. Results were pooled from four independent experiments (A) or are representative of at least two to three independent experiments (B–G; *, P < 0.05; **, P < 0.01 as compared with Tc0s as determined by one-way ANOVA with Tukey test).

Figure 2.

Tc22 polarization is inhibited by T-bet and facilitated by AhR. A, Expression of various transcription factors in day 3 polarized Tc subsets was quantified by RT-PCR and expressed as fold change relative to Tc0 ± SEM (n = 4). B, AhR and T-bet expression in Tc0, Tc1, and Tc22s on day 3 post-activation as determined by flow cytometry. C and D, WT and T-bet−/− CD8+ T cells were stimulated under polarizing conditions for 3 days. Cytokine production was assessed in Tc22-polarized cells (C) and other Tc subsets (D) and expressed as mean percentage of positive cells of technical replicates ± SEM. E, WT and AhR −/− CD8+ T cells were stimulated in Tc22-polarizing conditions for 3 days, and cytokine production was assessed. F, CD8+ T cells were stimulated in the presence of Tc22-polarizing conditions without FICZ. Either an AhR antagonist (CH-223191), an AhR agonist (FICZ), or vehicle control was added at the start of the culture, and cytokine production was assessed 3 days later. G, Percentage of IL22+ CD8+ T cells after 3-day stimulation in the indicated polarizing conditions in conjunction with CH-223191, vehicle control, or FICZ. Values were expressed as mean percentage of IL22+ from technical replicates ± SEM. All FACS plots shown were gated on CD8+ T cells. Results were pooled from four independent experiments (A) or are representative of at least two to three independent experiments (B–G; *, P < 0.05; **, P < 0.01 as compared with Tc0s as determined by one-way ANOVA with Tukey test).

Close modal

The transcriptional control of Th22 polarization is thought to be dependent on T-bet and AhR, although Th22s demonstrate minimal expression of AhR transcripts (51). Our data showed that Tc22s expressed comparable amounts of AhR to Tc1s and Tc0s at the transcript and protein level, but less T-bet than either of these subsets (Fig. 2A and B). To determine whether T-bet was important for the development of the Tc22 lineage, we stimulated T-bet −/− CD8+ T cells in the presence of polarizing cytokines and found that T-bet −/− Tc0, Tc1, Tc17, and Tc22s all had enhanced IL22 expression, indicating that T-bet was a negative regulator of IL22 production in CD8+ T cells (Fig. 2C and D; Supplementary Fig. S3). To evaluate the importance of AhR in Tc22 differentiation, we induced Tc subsets using AhR−/− CD8+ T cells and found IL22 production was markedly reduced (Fig. 2E). AhR also regulated IL22 production in other subsets, as AhR−/− Tc0s, Tc1s, and Tc17s produced less IL22 (Supplementary Fig. S4). Tc22 polarization was enhanced in the presence of the AhR agonist FICZ or inhibited by the AhR antagonist CH-223191 (Fig. 2F). These effects were not limited to Tc22s, since treatment of Tc1s and Tc17s with FICZ and CH-223191 also promoted or inhibited IL22 production, respectively (Fig. 2G). AhR alone was not sufficient to induce IL22 production and Tc22 polarization; virtually no increase in IL22 production was observed in Tc0s treated with FICZ alone (Fig. 2G). It is likely that other transcription factors downstream of IL6 signaling were required, such as STAT3, which we identified to be highly expressed in Tc22s compared with Tc0/Tc1s (Supplementary Fig. S5). Collectively, these data demonstrate that IL6 in conjunction with AhR drove IL22 production and Tc22 differentiation.

Tc subsets display unique transcriptomes

To assess the degree of similarity between Tc22s and other Tc subsets, we evaluated the transcriptome of CD8+ Tc subsets using RNA-seq (Fig. 3). Using unsupervised PCA, we found that each sample was grouped by Tc lineage (Fig. 3A). To quantify differences between Tc lineages, we performed differential expression analysis between all pairs of Tc subsets (Fig. 3B). Any pair of Tc subsets differed by at least 2,000 significant (FDR-adjusted P value < 0.05) differences, suggesting that each subset was a distinct lineage. Tc17s in particular seemed to be the most distinct as they had the greatest number of differentially expressed genes relative to other subsets (Fig. 3B).

Figure 3.

Tc subsets are transcriptionally distinct. RNA was extracted from polarized CD8+ Tc subsets from 3 different mice and sequenced. A, PCA of Tc subset expression profiles. B, Boxplots summarizing the number of significantly (DESeq2 FDR < 0.05) different genes when comparing Tc subsets. C, Heatmap of Tc lineage expression of genes at least 4-fold upregulated in a given Tc lineage compared with all others. Values are per gene Z-normalized log2 (RNA-seq read count+1), averaged per Tc lineage. D, Volcano plot summarizing Tc22 versus Tc17 differential expression results and highlighting several key gene differences between subsets. E, Network clustering of gene sets significantly enriched in GSEA pathway analysis of Tc22 versus Tc17 comparison. All gene set clusters with at least 4 enriched gene sets are shown. ECM, extracellular matrix; ER, endoplasmic reticulum; NK, natural killer.

Figure 3.

Tc subsets are transcriptionally distinct. RNA was extracted from polarized CD8+ Tc subsets from 3 different mice and sequenced. A, PCA of Tc subset expression profiles. B, Boxplots summarizing the number of significantly (DESeq2 FDR < 0.05) different genes when comparing Tc subsets. C, Heatmap of Tc lineage expression of genes at least 4-fold upregulated in a given Tc lineage compared with all others. Values are per gene Z-normalized log2 (RNA-seq read count+1), averaged per Tc lineage. D, Volcano plot summarizing Tc22 versus Tc17 differential expression results and highlighting several key gene differences between subsets. E, Network clustering of gene sets significantly enriched in GSEA pathway analysis of Tc22 versus Tc17 comparison. All gene set clusters with at least 4 enriched gene sets are shown. ECM, extracellular matrix; ER, endoplasmic reticulum; NK, natural killer.

Close modal

To define a transcriptional signature of each Tc subset, we identified which genes were specific to each Tc lineage (Fig. 3C; Supplementary Table S3). Using the cutoff of genes being at least 4-fold upregulated from any other subset, we identified lineage-specific genes for each Tc subset, some of which were for surface markers, transcription factors, and cytokines. For example, IL17a, Cd101, and Cd86 were found to be specific for the Tc17 lineage, while high Ifng was indicative of Tc1s. Next, given that both Tc17s and Tc22s could produce IL22 (Fig. 1A), we compared the transcriptomes of Tc17s and Tc22 to further define each lineage and found substantial differences in expressed genes (Fig. 3D). Through GSEA, we identified enriched gene sets in Tc22s versus Tc17s, which network clustering grouped into several gene set modules (Fig. 3E). For instance, some of the gene sets enriched in Tc17 cells were associated with protein localization and translation, while gene sets involved in proliferation and effector functions were found to be enriched in Tc22 cells (Fig. 3E). Finally, many of the genes we identified as either being highly expressed by a single subset or as being differentially expressed in Tc17 versus Tc22 cells were validated throughout the manuscript (Figs. 4 and 5; Supplementary Fig. S6). Together, these findings suggested that each subset was transcriptionally distinct, further supporting the notion that each subset was a distinct lineage.

Figure 4.

Tc subsets differentially express surface markers. CD8+ T cells were stimulated for 3 days in Tc-polarizing conditions, and surface markers were analyzed by flow cytometry: CD62L and CD44 expression on Tc0s (A); KLRG1, CCR7, CD25, and CD127 (B); TNFR2 and IL18R (C); CCR4, CCR6, and CCR10 (D); TNF-superfamily members 4–1BB, OX40, CD27, CD30, and GITR (E); and costimulatory molecules ICOS, CD86, SLAM, and CD101 (F). All plots shown contain fluorescence minus one (FMO) controls and were gated on CD8+ T cells. Results are representative of at least two to three independent experiments.

Figure 4.

Tc subsets differentially express surface markers. CD8+ T cells were stimulated for 3 days in Tc-polarizing conditions, and surface markers were analyzed by flow cytometry: CD62L and CD44 expression on Tc0s (A); KLRG1, CCR7, CD25, and CD127 (B); TNFR2 and IL18R (C); CCR4, CCR6, and CCR10 (D); TNF-superfamily members 4–1BB, OX40, CD27, CD30, and GITR (E); and costimulatory molecules ICOS, CD86, SLAM, and CD101 (F). All plots shown contain fluorescence minus one (FMO) controls and were gated on CD8+ T cells. Results are representative of at least two to three independent experiments.

Close modal
Figure 5.

Tc subsets have distinct effector functions. A, CD8+ T cells were stimulated for 3 days in Tc22-polarizing conditions, and ICS was performed after a 6-hour restimulation. B, Cytokines in supernatants in Tc subsets after a 24-hour restimulation with anti-CD3. Results are mean cytokine concentration pooled from three independent experiments and normalized to cell number ± SEM. C, IL22 expression in Tc0s, Tc1s, Tc2s, Tc9s, Tc17s, and Tc22s by ICS after a 6-hour restimulation. D, Percentage of lysis by polarized P14 Tc subsets after a 5-hour incubation with gp33 and AV control peptide–pulsed EL4 cells. Results were mean percentage of lysis pooled from three independent experiments. E, Polarized Tc subsets were stained for Granzyme B, Perforin, FasL, and TRAIL. All FACS plots were gated on CD8+ T cells. Results are representative of at least two to three independent experiments. Statistical analysis performed was one-way ANOVA with Tukey test (B) and two-way ANOVA with Tukey test (D; *, P < 0.05; **, P < 0.01 relative to Tc0 cells). E:T, effector-to-target; FMO, fluorescence minus one.

Figure 5.

Tc subsets have distinct effector functions. A, CD8+ T cells were stimulated for 3 days in Tc22-polarizing conditions, and ICS was performed after a 6-hour restimulation. B, Cytokines in supernatants in Tc subsets after a 24-hour restimulation with anti-CD3. Results are mean cytokine concentration pooled from three independent experiments and normalized to cell number ± SEM. C, IL22 expression in Tc0s, Tc1s, Tc2s, Tc9s, Tc17s, and Tc22s by ICS after a 6-hour restimulation. D, Percentage of lysis by polarized P14 Tc subsets after a 5-hour incubation with gp33 and AV control peptide–pulsed EL4 cells. Results were mean percentage of lysis pooled from three independent experiments. E, Polarized Tc subsets were stained for Granzyme B, Perforin, FasL, and TRAIL. All FACS plots were gated on CD8+ T cells. Results are representative of at least two to three independent experiments. Statistical analysis performed was one-way ANOVA with Tukey test (B) and two-way ANOVA with Tukey test (D; *, P < 0.05; **, P < 0.01 relative to Tc0 cells). E:T, effector-to-target; FMO, fluorescence minus one.

Close modal

Tc subsets differentially express surface markers

Next, we evaluated the activation status of each of the polarized CD8+ Tc subsets by examining the expression of several different markers by flow cytometry. All Tc subsets had a similar effector phenotype, as they were CD44hiCD62LCCR7KLRG1 in addition to expressing high levels of CD25 and low levels of CD127 (Fig. 4A and B).

An important aspect of defining and characterizing different CD8+ T-cell lineages is finding unique surface markers expressed by each subset to facilitate their identification. In addition to examining markers commonly used to identify CD4+ Th subsets, we also explored RNA-seq data to look for novel identifying markers. Together, this led us to investigate several types of surface molecules including cytokine receptors (Fig. 4C), chemokine receptors (Fig. 4D), and costimulatory molecules (Fig. 4E and F). In the case of cytokine receptors, the best uniquely identifying receptor we found was IL18R as the Tc0, Tc2, Tc9, Tc17, and Tc22 subsets were IL18Rlo while Tc1s were IL18Rhi (Fig. 4C). For Tc22s, we did not identify a unique cytokine receptor; however, Tc22s expressed lower TNFR2 compared with all of the other subsets (Fig. 4C). When looking at the expression of different chemokine receptors, we found that CCR6 and CCR10 were poorly expressed with minimal differences across all subsets (Fig. 4D). The only exception was CCR4, which was highly expressed on Tc9s compared with other Tc subsets. In general, unlike the CD4+ Th counterparts, differential expression of the chemokine receptors was unable to discriminate between the various CD8+ Tc subsets.

Next, we investigated the expression of costimulatory markers, as several are differentially expressed on CD4+ Th subsets (52). Many differences in expression were observed amongst the various subsets; however, the most striking being the differences in the expression of 4-1BB and OX40 (Fig. 4E). These markers were both highly expressed on all of the Tc subsets except for Tc17s, which surprisingly did not express either molecule (Fig. 4E). Tc17 cells also expressed lower levels of CD30 and GITR compared with other Tc subsets. In contrast, Tc17s highly expressed CD86 and CD101, both of which were identified by RNA-seq as being specific to Tc17s (Fig. 4F). In the case of Tc1s and Tc22s, both of these subsets highly expressed ICOS, but could be distinguished from each other based on high IL18R expression by Tc1s and not Tc22s. In general, within each polarizing condition, all of the cells had similar expression of the various markers that were evaluated, as clear biphasic distributions were not observed in most cases. For example, all of the cells polarized in Tc17 conditions displayed a lack of 4-1BB expression (Fig. 4E). This confirmed that the entire population was polarized to a distinct lineage despite the fact that only a proportion of the cells expressed the lineage-defining cytokine(s) (Fig. 1A). Together, these results showed that costimulatory markers were differentially expressed across subsets, with Tc17s standing out from the rest of the Tc subsets with respect to the majority of markers examined.

Tc subsets have distinct cytokine profiles and cytolytic activity

We further characterized the cytokine expression profile of Tc22s and found the majority of IL22 producers to be negative for IL17 and IFNγ, however some Tc22s also coproduced IFNγ and, to a lesser extent, IL17 (Fig. 5A). The IL22 producers also predominantly expressed TNFα and IL2. When comparing the cytokine expression profile of Tc22s against other Tc subsets, we observed a distinct profile produced by each Tc subset (Fig. 5B; Supplementary Fig. S7). Tc22s were among the highest producers of IL2 and TNFα, but produced less IFNγ than Tc0s and Tc1s. Tc22s did not express the Th2/Tc2-associated cytokines IL4, IL5, and IL13 nor the Th9/Tc9 lineage-defining cytokine IL9. Tc22s minimally expressed the Tc17 cytokines IL17A and IL17F. Tc9s and Tc17s both had equally high amounts of IL17 in the culture supernatant after the overnight restimulation (Fig. 5B). This raises the possibility that Tc9s were converting to Tc17s during the overnight restimulation. Indeed, this phenomenon occurs in their CD4+ counterparts, as IL17 is produced by Th9 cells after restimulation with anti-CD3 (53), potentially due to the fact that IL9 can act on CD4+ T cells to induce IL17 production (54, 55).

Although originally defined as a Th1-associated cytokine, subsequent studies have identified Th17s and Th22s as being prominent sources of IL22 (56). Indeed, we demonstrated a similar finding in CD8+ Tc subsets, as IL22 was produced predominantly by Tc22s and Tc17s, and to a lesser extent by Tc1s (Fig. 5C). When using DCs to polarize the cells, IL22 was sometimes detected in Tc0s, Tc2s and Tc9s to varying degrees, thereby suggesting undefined factors produced by DCs promoted some IL22 expression in CD8+ T cells (Fig. 5C). Nevertheless, although other Tc subsets produced IL22 in some situations, Tc22s were among the highest producers.

An important part of CD8+ T-cell effector function is their cytotoxic abilities. Our data showed that Tc1s, Tc2s, and Tc22s were the most cytolytic, while Tc17s were poor killers (Fig. 5D). Cytolytic activity correlated with granzyme B production as Tc0, Tc1, Tc2, and Tc22 had moderate-to-high levels of granzyme B, while Tc9 and Tc17s were granzyme B low (Fig. 5E). All Tc subsets poorly expressed perforin and FasL, and none of the Tc subsets expressed TRAIL (Fig. 5E). Therefore, granzyme B expression correlated with the cytotoxic activity of the polarized populations.

To further support the data that CD8+ Tc22s and Tc17s were distinct subsets, we investigated whether the IL22+ and IL17+ population in the Tc17 conditions were similar to their respective IL22+ and IL17+ counterparts in Tc22 conditions (Supplementary Fig. S8). When looking at the IL22+ in the Tc17 conditions, we found that these cells highly expressed the Tc17 marker CD101 and poorly expressed 4-1BB and granzyme B, all of which were similar to both the IL17+ fraction as well as total Tc17 cells (Supplementary Fig. S8). In contrast, the IL22+ population arising within the Tc22 conditions did not express CD101, but instead highly expressed 4-1BB and had comparable granzyme B expression to total Tc22 cells. Together, these findings indicated that the IL22+ cells arising in Tc17 conditions were similar to the IL17+ Tc17s and distinct from the IL22+ population within the Tc22 conditions.

Tc22-polarized cells promote tumor regression

Having identified a differential cytolytic ability of the various Tc lineages, we evaluated the effector function of these cells in an in vivo experimental model. As the adoptive transfer of tumor-specific CD8+ T cells has promising clinical results (57, 58), we investigated whether polarizing tumor-specific T cells to Tc22s and other Tc subsets prior to transfer may enhance antitumor functions. Several studies involving the transfer of tumor-specific T cells into mice bearing subcutaneous tumors show that Tc1s are superior to Tc2s (59, 60). Given this, we set out to evaluate how Tc1s compare with other Tc subsets, namely Tc0s, Tc9s, and Tc17s, in the context of our tumor model in which we transferred 1 × 106 polarized CD8+ P14 T cells into mice bearing day 10–11 established B16-gp33 melanoma tumors (∼5–6 mm in diameter). The greatest degree of protection was observed in mice that received Tc1s, as tumor progression was delayed about 10 days longer than mice that received Tc0s (Fig. 6A). Unlike previous reports (7, 11), we found that mice receiving Tc9s or Tc17s had virtually no protection against tumor growth. This is likely due to differences in the model used, as other studies have administered T cells in conjunction with supportive treatments including IL2 injections, tumor–antigen vaccination, and lymphodepletion, thereby providing more favorable conditions for response (7, 11). Another possibility was that some of the previous studies demonstrating a role for Tc17s may actually be examining Tc22s instead, as there is a potential for Tc22 to be misidentified as Tc17 given that cells activated in Tc22 conditions expressed IL17 and RORγT to some degree (Figs. 2A and 5B). Future studies may consider quantifying IL22 production as well as IL17 when polarizing Tc17s or evaluating Tc17 lineage commitment with markers identified here, such as 4-1BB and CD101+ to rule this out.

Figure 6.

Tc22s demonstrate potent antitumor function. A–C, Eight- to 12-week-old mice were inoculated subcutaneously with 4 × 105 B16-gp33 tumor cells in the right flank, and 10 to 11 days later received 1 × 106 polarized P14 Tc subsets as indicated. A, Mean tumor area in mice that received indicated P14 Tc subsets (n = 5 per group). B, Mean tumor area in mice that received either Tc1 or Tc22 (n = 4 per group). C, Survival curve from B representing combined survival across all experiments (no T cells n = 19, Tc1 n = 18, Tc22 n = 19). D, Western blot analysis of mitochondrial proteins in lysates of Tc subsets. E, ATP in Tc subsets from a representative experiment. RLU, relative light units. Human naïve T cells isolated from PBMCs were activated for 5 days in nonpolarizing conditions (F) or Tc22 conditions (G), and cytokine production was analyzed by flow cytometry after gating on viable CD3+ CD8+ T cells. Results shown are representative of at least three independent experiments with different donors. H, Cytokine production in viable tumor-infiltrating CD3+ CD8+ T cells expanded from a patient with ovarian cancer. I, Percentage of IFNγ+, IL17+, or IL22+ CD8+ TILs expanded from patients with ovarian cancer of mixed histologies and stages (n = 27). RFS of patients with stage IIIC high-grade serous ovarian cancer stratified on the basis of IL22 expression (J) or IFNγ expression (K) in CD8+ T cells expanded from ovarian cancer TILs. Cytokine-high group (n = 10) represents the top 50% of patients, while the cytokine-low group (n = 10) represents the bottom 50%. Results shown are representative of two to five independent experiments (A, B, and D–G) or were pooled from multiple independent experiments or patients (C and I–K). Error bars, SEM. Statistical analysis was performed using multiple t tests with Holm–Sidak correction comparing each Tc subset to no T-cell control (A), repeated measures ANOVA with Sidak test relative to Tc1 cells (B), log-rank test relative to Tc1 cell (C), and log-rank test (J and K; *, P < 0.05; **, P < 0.01; n.s., not significant).

Figure 6.

Tc22s demonstrate potent antitumor function. A–C, Eight- to 12-week-old mice were inoculated subcutaneously with 4 × 105 B16-gp33 tumor cells in the right flank, and 10 to 11 days later received 1 × 106 polarized P14 Tc subsets as indicated. A, Mean tumor area in mice that received indicated P14 Tc subsets (n = 5 per group). B, Mean tumor area in mice that received either Tc1 or Tc22 (n = 4 per group). C, Survival curve from B representing combined survival across all experiments (no T cells n = 19, Tc1 n = 18, Tc22 n = 19). D, Western blot analysis of mitochondrial proteins in lysates of Tc subsets. E, ATP in Tc subsets from a representative experiment. RLU, relative light units. Human naïve T cells isolated from PBMCs were activated for 5 days in nonpolarizing conditions (F) or Tc22 conditions (G), and cytokine production was analyzed by flow cytometry after gating on viable CD3+ CD8+ T cells. Results shown are representative of at least three independent experiments with different donors. H, Cytokine production in viable tumor-infiltrating CD3+ CD8+ T cells expanded from a patient with ovarian cancer. I, Percentage of IFNγ+, IL17+, or IL22+ CD8+ TILs expanded from patients with ovarian cancer of mixed histologies and stages (n = 27). RFS of patients with stage IIIC high-grade serous ovarian cancer stratified on the basis of IL22 expression (J) or IFNγ expression (K) in CD8+ T cells expanded from ovarian cancer TILs. Cytokine-high group (n = 10) represents the top 50% of patients, while the cytokine-low group (n = 10) represents the bottom 50%. Results shown are representative of two to five independent experiments (A, B, and D–G) or were pooled from multiple independent experiments or patients (C and I–K). Error bars, SEM. Statistical analysis was performed using multiple t tests with Holm–Sidak correction comparing each Tc subset to no T-cell control (A), repeated measures ANOVA with Sidak test relative to Tc1 cells (B), log-rank test relative to Tc1 cell (C), and log-rank test (J and K; *, P < 0.05; **, P < 0.01; n.s., not significant).

Close modal

In agreement with previously published studies (59–63), our model showed IL12-induced Tc1 cells to have superior antitumor functions compared with previously described Tc subsets (Fig. 6A). Therefore, we performed multiple independent experiments to evaluate the antitumor functions of Tc22s using Tc1s as our reference Tc subset (Fig. 6B and C). We found that Tc22s performed at least as well as, if not better than, Tc1s. The growth of tumors in mice that received either Tc1s or Tc22s was arrested and tumors shrank in the majority of mice (Fig. 6B and C). On average, mice that received Tc22s had prolonged survival compared with those that received Tc1s (Fig. 6C). Surprisingly, in 2 of 19 mice that received Tc22s, their tumors did not return. These mice demonstrated a complete response and remained tumor free for more than 200 days. From these experiments, both Tc1s and Tc22s have superior antitumor properties, while some subsets, including Tc17s, have minimal activity in our model.

To examine the potential mechanisms that would account for increased antitumor activity in the Tc22 subset, we looked at exhaustion markers and found Tc1 and Tc22 cells similarly expressed PD-1 (Supplementary Fig. S9). Next, we looked at the bioenergetics of these subsets as T cells with increased fitness and bioenergetics perform well in the tumor microenvironment (64). To investigate whether this may be the case for the observed antitumor effects of our subsets, we analyzed expression of mitochondrial proteins involved in the electron transport chain and found them to be increased in Tc22 cells, particularly complex II (Fig. 6D). In line with this, we also observed enhanced ATP in Tc22 cells (Fig. 6E). Together, these data suggested that Tc22 have increased bioenergetics, which may have accounted for increased antitumor function.

To better understand the role of Tc22s in the context of human disease, we investigated whether Tc22s could be induced in human CD8+ T cells, and if so, could we detect them among TILs. IL22-producing CD8+ T cells were observed in activated PBMCs isolated from healthy donors (Fig. 6F). Upon activation in the presence of Tc22-polarizing cytokines, human CD8+ T cells skew primarily toward an IL22+ IL17 IFNγ Tc22 phenotype, with the rest of the IL22+ T cells coproducing IFNγ+ (Fig. 6G). IL22-producing CD8+ T cells were also detected in TILs expanded from ovarian cancer tissue of several patients, where they comprised up to approximately 35% of CD8+ T cells (Fig. 6H and I). Despite being cultured in nonpolarizing conditions, these expanded CD8+ TILs retained their Tc22 phenotype, further supporting the notion that Tc22s were a distinct subset in humans. We found that increased frequencies of IL22-producing CD8+ T cells significantly correlated with improved RFS in patients with stage IIIC high-grade serous ovarian cancer, suggesting a clinical relevance for these cells (Fig. 6J). Surprisingly, we did not observe a similar finding with IFNγ as no correlation with RFS was observed, suggesting that IL22 production by CD8+ T cells was a better biomarker of response than IFNγ (Fig. 6K). Taken together, these findings indicated that human CD8+ T cells could polarize towards the Tc22 subset and could be detected in ovarian cancer TILs with a potentially beneficial role in controlling tumor growth.

CD8+ T cells can be classified into one of several subsets as determined by the cytokines they produce. Here, we further expanded the known types of Tc subsets by inducing and characterizing CD8+ IL22+ Tc22s, whose polarization required IL6 and AhR. These Tc22s were distinct from other Tc subsets as they express high levels of IL22 and IL2 but less IFNγ, IL4, IL9, and IL17 relative to other Tc subsets. We further demonstrated that each Tc subset, including Tc22s, are unique in their effector functions and transcriptome profiles. Each subset also demonstrated vastly different degrees of tumor protection in the B16 melanoma model, ranging from ineffective (Tc17, Tc9) to most effective (Tc1, Tc22). These data established Tc22 as a distinct CD8+ T-cell lineage.

The different CD8+ Tc subsets mirror those of their CD4+ counterparts in that they produced a similar profile of cytokines, were polarized by the same types of cytokines, and used the same signaling pathways and transcription factors to drive Tc polarization. Tc22 followed a similar pattern, as induction of Tc22s was dependent on the combined effects of IL6 and the transcription factor AhR. However, IL6 on its own was not sufficient to induce Tc22s as polarization was impaired in AhR−/− CD8+ T cells treated with IL6. Similarly, AhR activation alone was insufficient to induce Tc22s as limited IL22 production was observed in Tc0s treated with the AhR agonist FICZ. These findings were similar to what has been shown in CD4+ T cells, as AhR alone was not sufficient to induce the transcription of IL22 even though the IL22 promoter contains AhR response elements (65). However, STAT3 facilitates the recruitment of AhR to the IL22 promoter leading to the induction of IL22 (65). As STAT3 is downstream of the IL6 receptor (66), it is likely that the induction of Tc22s is dependent on the cooperative effects of AhR and the IL6/STAT3 axis. As such, we observed the highest degree of Tc22 polarization when both IL6 and FICZ were present in the polarizing conditions. This may explain why Tc22s were detected in human inflammatory skin lesions in high numbers with >20% of infiltrating CD8+ T cells being Tc22s in some patients (24), as exposure of tryptophan to UV rays from sunlight can induce the AhR ligand FICZ (67).

In the context of skin inflammation, previous reports have identified a positive correlation between the presence of Tc22s and the severity of disease (24). In this scenario, Tc22s likely contribute to pathology as a result of their high cytotoxicity in conjunction with increased expression of cytokines including IL2 and TNFα. However, in the context of an antitumor response, these properties attributed to Tc22 cells may in fact be beneficial when it comes to controlling tumor growth. To this end, we propose that Tc22-mediated immunity is not necessarily attributed to the production of IL22 itself, but instead IL22 serves as a marker for a subset of CD8+ T cells that is highly cytotoxic and produces antitumor cytokines such as TNFα and IL2, the latter of which is not produced by Tc1s. Together, these cytokines augment the antitumor function of T cells, with IL2 in particular being used clinically in many adoptive T-cell therapy protocols (68). Therefore, polarizing cells toward highly cytotoxic Tc22 cells that co-produce beneficial cytokines such as IL2 and TNFα may be a novel approach to improve adoptive immunotherapy.

The role of T-cell bioenergetics in the antitumor response is becoming increasingly appreciated. Tumor-infiltrating T cells are in a bioenergetics crisis as evident by dysfunctional metabolism (64). Although the mechanisms behind this are not entirely clear, some evidence indicates that inhibitory factors within the tumor microenvironment contribute to this dysfunction (69). Indeed, genetic (64) and pharmacologic (70) approaches aimed at enhancing T-cell bioenergetics show promise in enhancing their antitumor activity potentially by rendering them resistant to tumor-induced metabolic suppression. Given that Tc22 cells also demonstrate increased bioenergetics relative to other Tc subsets, we believe this may account for their increased antitumor activity.

In conclusion, we have identified the polarizing conditions to induce the relatively undefined IL22-producing CD8+ Tc22 subset, which we demonstrated to be transcriptionally and functionally distinct from other Tc subsets. Tc22s demonstrated exceptional killing potential and antitumor functions, as they promoted the regression of established B16 melanoma tumors upon T-cell transfer. IL22-producing T cells were detected in TILs expanded from patients with ovarian cancer, and their presence correlated with an improved RFS. Therefore, novel immunotherapies could be aimed at employing Tc22s for adoptive T-cell transfer or by inducing the polarization of Tc22s within the tumor microenvironment to promote an effective antitumor immune response.

No potential conflicts of interest were disclosed.

Conception and design: M. St. Paul, L.T. Nguyen, P.S. Ohashi

Development of methodology: M. St. Paul

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. St. Paul, S.D. Saibil, S.C. Lien, S. Han, D.T. Mulder, C.R. Garcia-Batres, A.R. Elford, C. Robert-Tissot, S.R. Katz, P.A. Shaw, B.A. Clarke, M.Q. Bernardini, T.J. Pugh

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. St. Paul, S.D. Saibil, S. Han, A. Sayad, D.T. Mulder, C.R. Garcia-Batres, M. Zon, B. Haibe-Kains, T.J. Pugh

Writing, review, and/or revision of the manuscript: M. St. Paul, S.D. Saibil, S. Han, A. Sayad, C. Robert-Tissot, S.R. Katz, P.A. Shaw, B.A. Clarke, B. Haibe-Kains, T.J. Pugh, P.S. Ohashi

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. St. Paul, A.R. Elford, T.J. Pugh

Study supervision: M. St. Paul, P.S. Ohashi

Other (performed experiments): K. Israni-Winger

The authors would like to thank S.Q. Crome, H. MacGregor, B.X. Wang, O. Chan, K. Gill, M. Pniak, J. Nie, and P. Yen for their technical assistance. This work was supported by a Canadian Institute for Health Research Foundation Award to P.S. Ohashi and a Natural Science and Engineering Research Council scholarship to M. St. Paul. P.S. Ohashi holds a Canada Research Chair in Autoimmunity and Tumor Immunity.

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.
Zhang
N
,
Bevan
MJ
. 
CD8+ T cells: foot soldiers of the immune system
.
Immunity
2011
;
35
:
161
8
.
2.
Salgame
P
,
Abrams
JS
,
Clayberger
C
,
Goldstein
H
,
Convit
J
,
Modlin
RL
, et al
Differing lymphokine profiles of functional subsets of human CD4 and CD8 T cell clones
.
Science
1991
;
254
:
279
82
.
3.
Seder
RA
,
Boulay
JL
,
Finkelman
F
,
Barbier
S
,
Ben-Sasson
SZ
,
Le Gros
G
, et al
CD8+ T cells can be primed in vitro to produce IL-4
.
J Immunol
1992
;
148
:
1652
6
.
4.
Sad
S
,
Marcotte
R
,
Mosmann
TR
. 
Cytokine-induced differentiation of precursor mouse CD8+ T cells into cytotoxic CD8+ T cells secreting Th1 or Th2 cytokines
.
Immunity
1995
;
2
:
271
9
.
5.
Croft
M
,
Carter
L
,
Swain
SL
,
Dutton
RW
. 
Generation of polarized antigen-specific CD8 effector populations: reciprocal action of interleukin (IL)-4 and IL-12 in promoting type 2 versus type 1 cytokine profiles
.
J Exp Med
1994
;
180
:
1715
28
.
6.
Visekruna
A
,
Ritter
J
,
Scholz
T
,
Campos
L
,
Guralnik
A
,
Poncette
L
, et al
Tc9 cells, a new subset of CD8(+) T cells, support Th2-mediated airway inflammation
.
Eur J Immunol
2013
;
43
:
606
18
.
7.
Lu
Y
,
Hong
B
,
Li
H
,
Zheng
Y
,
Zhang
M
,
Wang
S
, et al
Tumor-specific IL-9-producing CD8+ Tc9 cells are superior effector than type-I cytotoxic Tc1 cells for adoptive immunotherapy of cancers
.
Proc Natl Acad Sci U S A
2014
;
111
:
2265
70
.
8.
Intlekofer
AM
,
Banerjee
A
,
Takemoto
N
,
Gordon
SM
,
Dejong
CS
,
Shin
H
, et al
Anomalous type 17 response to viral infection by CD8+ T cells lacking T-bet and eomesodermin
.
Science
2008
;
321
:
408
11
.
9.
Hamada
H
,
Garcia-Hernandez
MDLL
,
Reome
JB
,
Misra
SK
,
Strutt
TM
,
McKinstry
KK
, et al
Tc17, a unique subset of CD8 T cells that can protect against lethal influenza challenge
.
J Immunol
2009
;
182
:
3469
81
.
10.
Huber
M
,
Heink
S
,
Grothe
H
,
Guralnik
A
,
Reinhard
K
,
Elflein
K
, et al
A Th17-like developmental process leads to CD8(+) Tc17 cells with reduced cytotoxic activity
.
Eur J Immunol
2009
;
39
:
1716
25
.
11.
Hinrichs
CS
,
Kaiser
A
,
Paulos
CM
,
Cassard
L
,
Sanchez-Perez
L
,
Heemskerk
B
, et al
Type 17 CD8+ T cells display enhanced antitumor immunity
.
Blood
2009
;
114
:
596
9
.
12.
Kryczek
I
,
Wei
S
,
Zou
L
,
Altuwaijri
S
,
Szeliga
W
,
Kolls
J
, et al
Cutting edge: Th17 and regulatory T cell dynamics and the regulation by IL-2 in the tumor microenvironment
.
J Immunol
2007
;
178
:
6730
3
.
13.
Mittrücker
H-W
,
Visekruna
A
,
Huber
M
. 
Heterogeneity in the differentiation and function of CD8+ T cells
.
Arch Immunol Ther Exp
2014
;
62
:
449
58
.
14.
Yen
H-R
,
Harris
TJ
,
Wada
S
,
Grosso
JF
,
Getnet
D
,
Goldberg M
V
, et al
Tc17 CD8 T cells: functional plasticity and subset diversity
.
J Immunol
2009
;
183
:
7161
8
.
15.
Huber
M
,
Heink
S
,
Pagenstecher
A
,
Reinhard
K
,
Ritter
J
,
Visekruna
A
, et al
IL-17A secretion by CD8+ T cells supports Th17-mediated autoimmune encephalomyelitis
.
J Clin Invest
2013
;
123
:
247
60
.
16.
Sheu
BC
,
Lin
RH
,
Lien
HC
,
Ho
HN
,
Hsu
SM
,
Huang
SC
. 
Predominant Th2/Tc2 polarity of tumor-infiltrating lymphocytes in human cervical cancer
.
J Immunol
2001
;
167
:
2972
8
.
17.
Sabat
R
,
Ouyang
W
,
Wolk
K
. 
Therapeutic opportunities of the IL-22-IL-22R1 system
.
Nat Rev Drug Discov
2013
;
13
:
21
38
.
18.
Weber
GF
,
Gaertner
FC
,
Erl
W
,
Janssen
K-P
,
Blechert
B
,
Holzmann
B
, et al
IL-22-mediated tumor growth reduction correlates with inhibition of ERK1/2 and AKT phosphorylation and induction of cell cycle arrest in the G2-M phase
.
J Immunol
2006
;
177
:
8266
72
.
19.
Zhang
F
,
Shang
D
,
Zhang
Y
,
Tian
Y
. 
Interleukin-22 suppresses the growth of A498 renal cell carcinoma cells via regulation of STAT1 pathway
.
PLoS One
2011
;
6
:
e20382
.
20.
Lim
C
,
Savan
R
. 
The role of the IL-22/IL-22R1 axis in cancer
.
Cytokine Growth Factor Rev
2014
;
25
:
257
71
.
21.
Huber
S
,
Gagliani
N
,
Zenewicz
LA
,
Huber
FJ
,
Bosurgi
L
,
Hu
B
, et al
IL-22BP is regulated by the inflammasome and modulates tumorigenesis in the intestine
.
Nature
2012
;
491
:
259
63
.
22.
Kryczek
I
,
Lin
Y
,
Nagarsheth
N
,
Peng
D
,
Zhao
L
,
Zhao
E
, et al
IL-22(+)CD4(+) T cells promote colorectal cancer stemness via STAT3 transcription factor activation and induction of the methyltransferase DOT1L
.
Immunity
2014
;
40
:
772
84
.
23.
Zenewicz
LA
,
Flavell
RA
. 
Recent advances in IL-22 biology
.
Int Immunol
2011
;
23
:
159
63
.
24.
Nograles
KE
,
Zaba
LC
,
Shemer
A
,
Fuentes-Duculan
J
,
Cardinale
I
,
Kikuchi
T
, et al
IL-22-producing “T22” T cells account for upregulated IL-22 in atopic dermatitis despite reduced IL-17-producing TH17 T cells
.
J Allergy Clin Immunol
2009
;
123
:
1244
52
.
25.
Res
PCM
,
Piskin
G
,
de Boer
OJ
,
van der Loos
CM
,
Teeling
P
,
Bos
JD
, et al
Overrepresentation of IL-17A and IL-22 producing CD8 T cells in lesional skin suggests their involvement in the pathogenesis of psoriasis
.
PLoS One
2010
;
5
:
e14108
.
26.
Teraki
Y
,
Sakurai
A
,
Izaki
S
. 
IL-13/IL-22–coproducing T cells, a novel subset, are increased in atopic dermatitis
.
J Allergy Clin Immunol
2013
;
132
:
971
4
.
27.
Oliveira
LMS
,
Lima
JF
,
Cervantes
CA
,
Casseb
JS
,
Mendonça
M
,
Duarte
AJ
, et al
Increased frequency of circulating Tc22/Th22 cells and polyfunctional CD38− T cells in HIV-exposed uninfected subjects
.
Sci Rep
2015
;
5
:
13883
.
28.
Jiang
R
,
Tan
Z
,
Deng
L
,
Chen
Y
,
Xia
Y
,
Gao
Y
, et al
Interleukin-22 promotes human hepatocellular carcinoma by activation of STAT3
.
Hepatology
2011
;
54
:
900
9
.
29.
Zhang
S
,
Fujita
H
,
Mitsui
H
,
Yanofsky
VR
,
Fuentes-Duculan
J
,
Pettersen
JS
, et al
Increased Tc22 and Treg/CD8 ratio contribute to aggressive growth of transplant associated squamous cell carcinoma
.
PLoS One
2013
;
8
:
e62154
.
30.
Liu
Y
,
Yang
B
,
Ma
J
,
Wang
H
,
Huang
F
,
Zhang
J
, et al
Interleukin-21 induces the differentiation of human Tc22 cells via phosphorylation of signal transducers and activators of transcription
.
Immunology
2011
;
132
:
540
8
.
31.
Pircher
H
,
Bürki
K
,
Lang
R
,
Hengartner
H
,
Zinkernagel
RM
. 
Tolerance induction in double specific T-cell receptor transgenic mice varies with antigen
.
Nature
1989
;
342
:
559
61
.
32.
Dissanayake
D
,
Hall
H
,
Berg-Brown
N
,
Elford
AR
,
Hamilton
SR
,
Murakami
K
, et al
Nuclear factor-κB1 controls the functional maturation of dendritic cells and prevents the activation of autoreactive T cells
.
Nat Med
2011
;
17
:
1663
7
.
33.
Dobin
A
,
Davis
CA
,
Schlesinger
F
,
Drenkow
J
,
Zaleski
C
,
Jha
S
, et al
STAR: ultrafast universal RNA-seq aligner
.
Bioinformatics
2013
;
29
:
15
21
.
34.
Frankish
A
,
Diekhans
M
,
Ferreira
AM
,
Johnson
R
,
Jungreis
I
,
Loveland
J
, et al
GENCODE reference annotation for the human and mouse genomes
.
Nucleic Acids Res
2019
;
47
:
D766
73
.
3
35.
Li
B
,
Dewey
CN
. 
RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome
.
BMC Bioinformatics
2011
;
12
:
323
.
36.
Huber
W
,
Carey
VJ
,
Gentleman
R
,
Anders
S
,
Carlson
M
,
Carvalho
BS
, et al
Orchestrating high-throughput genomic analysis with Bioconductor
.
Nat Methods
2015
;
12
:
115
21
.
37.
Durinck
S
,
Spellman
PT
,
Birney
E
,
Huber
W
. 
Mapping identifiers for the integration of genomic datasets with the R/Bioconductor package biomaRt
.
Nat Protoc
2009
;
4
:
1184
91
.
38.
Huber
W
,
von Heydebreck
A
,
Sültmann
H
,
Poustka
A
,
Vingron
M
. 
Variance stabilization applied to microarray data calibration and to the quantification of differential expression
.
Bioinformatics
2002
;
18
(
Suppl 1
)
:
S96
104
.
39.
Love
MI
,
Huber
W
,
Anders
S
. 
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
2014
;
15
:
550
.
40.
Benjamini
Y
,
Hochberg
Y
. 
Controlling the false discovery rate: a practical and powerful approach to multiple testing
.
J R Stat Soc Ser B
1995
;
57
:
289
300
.
41.
Zhu
A
,
Ibrahim
JG
,
Love
MI
. 
Heavy-tailed prior distributions for sequence count data: removing the noise and preserving large differences
.
Bioinformatics
2018
;
35
:
2084
92
.
42.
Subramanian
A
,
Tamayo
P
,
Mootha
VK
,
Mukherjee
S
,
Ebert
BL
,
Gillette
MA
, et al
Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles
.
Proc Natl Acad Sci U S A
2005
;
102
:
15545
50
.
43.
Shannon
P
,
Markiel
A
,
Ozier
O
,
Baliga
NS
,
Wang
JT
,
Ramage
D
, et al
Cytoscape: a software environment for integrated models of biomolecular interaction networks
.
Genome Res
2003
;
13
:
2498
504
.
44.
Merico
D
,
Isserlin
R
,
Stueker
O
,
Emili
A
,
Bader
GD
. 
Enrichment Map: a network-based method for gene-set enrichment visualization and interpretation
.
PLoS One
2010
;
5
:
e13984
.
45.
Su
G
,
Kuchinsky
A
,
Morris
JH
,
States
DJ
,
Meng
F
. 
GLay: community structure analysis of biological networks
.
Bioinformatics
2010
;
26
:
3135
7
.
46.
Wickham
H
.
Ggplot2: elegant graphics for data analysis
.
New York
:
Springer-Verlag
; 
2016
.
Available from
: https://ggplot2.tidyverse.org/.
47.
Crome
SQ
,
Nguyen
LT
,
Lopez-Verges
S
,
Yang
SYC
,
Martin
B
,
Yam
JY
, et al
A distinct innate lymphoid cell population regulates tumor-associated T cells
.
Nat Med
2017
;
23
:
368
75
.
48.
Duhen
T
,
Geiger
R
,
Jarrossay
D
,
Lanzavecchia
A
,
Sallusto
F
. 
Production of interleukin 22 but not interleukin 17 by a subset of human skin-homing memory T cells
.
Nat Immunol
2009
;
10
:
857
63
.
49.
Eyerich
S
,
Eyerich
K
,
Pennino
D
,
Carbone
T
,
Nasorri
F
,
Pallotta
S
, et al
Th22 cells represent a distinct human T cell subset involved in epidermal immunity and remodeling
.
J Clin Invest
2009
;
119
:
3573
85
.
50.
Zhu
J
,
Yamane
H
,
Paul
WE
. 
Differentiation of effector CD4 T cell populations
.
Annu Rev Immunol
2010
;
28
:
445
89
.
51.
Basu
R
,
O'Quinn
DB
,
Silberger
DJ
,
Schoeb
TR
,
Fouser
L
,
Ouyang
W
, et al
Th22 cells are an important source of IL-22 for host protection against enteropathogenic bacteria
.
Immunity
2012
;
37
:
1061
75
.
52.
Nakae
S
,
Iwakura
Y
,
Suto
H
,
Galli
SJ
. 
Phenotypic differences between Th1 and Th17 cells and negative regulation of Th1 cell differentiation by IL-17
.
J Leukoc Biol
2007
;
81
:
1258
68
.
53.
Jäger
A
,
Dardalhon
V
,
Sobel
RA
,
Bettelli
E
,
Kuchroo
VK
. 
Th1, Th17, and Th9 effector cells induce experimental autoimmune encephalomyelitis with different pathological phenotypes
.
J Immunol
2009
;
183
:
7169
77
.
54.
Elyaman
W
,
Bradshaw
EM
,
Uyttenhove
C
,
Dardalhon
V
,
Awasthi
A
,
Imitola
J
, et al
IL-9 induces differentiation of TH17 cells and enhances function of FoxP3+ natural regulatory T cells
.
Proc Natl Acad Sci U S A
2009
;
106
:
12885
90
.
55.
Nowak
EC
,
Noelle
RJ
. 
Interleukin-9 as a T helper type 17 cytokine
.
Immunology
2010
;
131
:
169
73
.
56.
Rutz
S
,
Eidenschenk
C
,
Ouyang
W
. 
IL-22, not simply a Th17 cytokine
.
Immunol Rev
2013
;
252
:
116
32
.
57.
Dafni
U
,
Michielin
O
,
Lluesma
SM
,
Tsourti
Z
,
Polydoropoulou
V
,
Karlis
D
, et al
Efficacy of adoptive therapy with tumor-infiltrating lymphocytes and recombinant interleukin-2 in advanced cutaneous melanoma: a systematic review and meta-analysis
.
Ann Oncol Off J Eur Soc Med Oncol
2019
;
30
:
1902
13
.
58.
D'Angelo
SP
,
Melchiori
L
,
Merchant
MS
,
Bernstein
D
,
Glod
J
,
Kaplan
R
, et al
Antitumor activity associated with prolonged persistence of adoptively transferred NY-ESO-1 c259 T cells in synovial sarcoma
.
Cancer Discov
2018
;
8
:
944
57
.
59.
Kemp
RA
,
Ronchese
F
. 
Tumor-specific Tc1, but not Tc2, cells deliver protective antitumor immunity
.
J Immunol
2001
;
167
:
6497
502
.
60.
Helmich
BK
,
Dutton
RW
. 
The role of adoptively transferred CD8 T cells and host cells in the control of the growth of the EG7 thymoma: factors that determine the relative effectiveness and homing properties of Tc1 and Tc2 effectors
.
J Immunol
2001
;
166
:
6500
8
.
61.
Nishimura
F
,
Dusak
JE
,
Eguchi
J
,
Zhu
X
,
Gambotto
A
,
Storkus
WJ
, et al
Adoptive transfer of type 1 CTL mediates effective anti-central nervous system tumor response: critical roles of IFN-inducible protein-10
.
Cancer Res
2006
;
66
:
4478
87
.
62.
Garcia-Hernandez
MDLL
,
Hamada
H
,
Reome
JB
,
Misra
SK
,
Tighe
MP
,
Dutton
RW
. 
Adoptive transfer of tumor-specific Tc17 effector T cells controls the growth of B16 melanoma in mice
.
J Immunol
2010
;
184
:
4215
27
.
63.
Yu
Y
,
Cho
H-I
,
Wang
D
,
Kaosaard
K
,
Anasetti
C
,
Celis
E
, et al
Adoptive transfer of Tc1 or Tc17 cells elicits antitumor immunity against established melanoma through distinct mechanisms
.
J Immunol
2013
;
190
:
1873
81
.
64.
Scharping
NE
,
Menk
AV
,
Moreci
RS
,
Whetstone
RD
,
Dadey
RE
,
Watkins
SC
, et al
The tumor microenvironment represses t cell mitochondrial biogenesis to drive intratumoral T cell metabolic insufficiency and dysfunction
.
Immunity
2016
;
45
:
374
88
.
65.
Yeste
A
,
Mascanfroni
ID
,
Nadeau
M
,
Burns
EJ
,
Tukpah
AM
,
Santiago
A
, et al
IL-21 induces IL-22 production in CD4+ T cells
.
Nat Commun
2014
;
5
:
3753
.
66.
Niemand
C
,
Nimmesgern
A
,
Haan
S
,
Fischer
P
,
Schaper
F
,
Rossaint
R
, et al
Activation of STAT3 by IL-6 and IL-10 in primary human macrophages is differentially modulated by suppressor of cytokine signaling 3
.
J Immunol
2003
;
170
:
3263
72
.
67.
Fritsche
E
,
Schäfer
C
,
Calles
C
,
Bernsmann
T
,
Bernshausen
T
,
Wurm
M
, et al
Lightening up the UV response by identification of the arylhydrocarbon receptor as a cytoplasmatic target for ultraviolet B radiation
.
Proc Natl Acad Sci U S A
2007
;
104
:
8851
6
.
68.
Rosenberg
SA
. 
IL-2: the first effective immunotherapy for human cancer
.
J Immunol
2014
;
192
:
5451
8
.
69.
Patsoukis
N
,
Bardhan
K
,
Chatterjee
P
,
Sari
D
,
Liu
B
,
Bell
LN
, et al
PD-1 alters T-cell metabolic reprogramming by inhibiting glycolysis and promoting lipolysis and fatty acid oxidation
.
Nat Commun
2015
;
6
:
6692
.
70.
Saibil
SD
,
St. Paul
M
,
Laister
RC
,
Garcia-Batres
CR
,
Israni-Winger
K
,
Elford
AR
, et al
Activation of peroxisome proliferator-activated receptors α and δ synergizes with inflammatory signals to enhance adoptive cell therapy
.
Cancer Res
2019
;
79
:
445
51
.