key: cord-0985637-z66yixzw
authors: Zhao, Yapu; Zhao, Huiyuan; Sun, Yuehong; Hao, Jianlei; Qi, Xiaofei; Zhou, Xinglong; Wu, Zhenzhou; Wang, Puyue; Kaech, Susan M.; Weaver, Casey T.; Flavell, Richard A.; Zhao, Liqing; Yao, Zhi; Yin, Zhinan
title: IL‐4 induces a suppressive IL‐10‐producing CD8(+) T cell population via a Cdkn2a‐dependent mechanism
date: 2013-06-14
journal: Journal of Leukocyte Biology
DOI: 10.1189/jlb.0213064
sha: 8d2a523d5c625a947a59c53634a9657bd203abe9
doc_id: 985637
cord_uid: z66yixzw

CD8(+) T cells play an important role in immune regulation and effective immune responses against tumor cells, viral infection, and intracellular pathogens. In this report, using tiger or 10BiT mice, we defined a population of IL‐10‐producing CD8(+) T cells that were induced by IL‐4. These IL‐10(+)CD8(+) T cells possessed a strong inhibitory effect on the CD4(+) T cell proliferation in an IL‐10‐dependent and cell contact‐dependent fashion. In comparison with IL‐10(−)CD8(+) T cells, IL‐10(+)CD8(+) T cells expressed an array of Th2‐like cytokines (IL‐4, IL‐5), perforin, and granzymes, as well as the cell cycle regulatory protein Cdkn2a. Interestingly, knockdown of cdkn2a using siRNA reduced IL‐4‐induced IL‐10 production significantly. Furthermore, CD8(+) T cells from Cdkn2a(−/−) mice produced a significantly lower amount of IL‐10, and the effect was limited to CD8(+) T cells but not observed in CD4(+) T cells and APCs. Finally, IL‐10(+)CD8(+) T cells played a protective role in the TNBS‐induced murine colitis model, indicating a critical role of this population of CD8(+) T cells in regulatory immune responses. Taken together, we have defined a population of IL‐10‐producing CD8(+) Tregs induced by IL‐4 and mediated by Cdkn2a.

Lymphocyte heterogeneity is required for optimal immune responses against pathogens and self-homeostasis. Several functionally distinct subsets of CD8 ϩ T cells have been defined. Traditional Tc1 and Tc2 CD8 ϩ T cells play overlapping and different roles in tumor immunity [1] , viral infection [2] , and some allergic diseases [3] . Tc17 cells can be generated in vitro and in vivo and exert critical functions in tumor rejection and viral clearance [4, 5] . CD8 ϩ Tregs play an important role in maintaining immune self-homeostasis and resistance to autoimmune diseases. CD8 ϩ CD122 ϩ Tregs prevent and cure naïve CD4 ϩ T cell-induced IBD [6] . Moreover, transferring CD8 ϩ CD28 Ϫ T cells into CD8deficient mice can suppress development of experimental autoimmune encephalomyelitis [7] . CD8 ϩ IL-10-producing T cells were found in several murine disease models, including coronavirus-induced encephalitis [8] , acute influenza virus infection [9] , and Salmonella and Leishmania infection [10] . Also, they were found in human HIV-1 infection [11] and chronic hepatitis C virus infection [12] .

IL-10 is a multifunctional cytokine produced by a variety of cell types, including Th2 cells, DCs, activated macrophages, B cells, and mast cells [13] . Recent reports have demonstrated that IL-10 is indispensable for Treg function [14] and suppresses proinflammatory T cell immunity [15] . Several groups, including us, have established that IL-10 is critical for maintaining the suppressive function of Tregs in arthritis [16] and colitis [17] . IL-10 suppresses TNF-␣ production by macrophages [18] , Th1 cell cytokine production, and T cell proliferation [19] . IL-10 restrains Th17 cell-mediated pathology [20] and CD45RB lo cell-mediated colitis [21] . Although Tregs [13] and type 1 Treg [22] , macrophages [13] , and various immune cells were found to be impor-tant sources of IL-10, the function of IL-10 ϩ CD8 ϩ T cells in inflammation remains to be investigated.

A number of transcription factors involved in regulating IL-10 expression in CD4 ϩ T cells and CD8 ϩ T cells have been defined. GATA-3 [23] and Blimp1 [24] mainly regulate IL-10 expression through a specific signal pathway in CD4 ϩ or CD8 ϩ T cells. Several transcription factors have been found to regulate IL-10 expression through different mechanisms in macrophages or monocytes. c-Maf [25] , stat1 [26] , and stat3 [27] have been reported to be involved in IL-10 transcription regulation in the LPS signal transduction pathway, whereas Sp1 [28] or Sp3 [28] directly binds to some specific motifs of the IL-10 promoter to alter IL-10 mRNA levels. However, the comprehensive transcription networks that are responsible for IL-10 production and their regulation by the cytokine environments are still unknown.

The cdkn2a locus in mouse encodes two distinct tumor-suppressor proteins: p19 ARF and p16 INK4a [29] . p19 ARF mainly regulates p53 in response to aberrant growth or oncogenic stresses, such as c-Myc activation [29] , whereas p16 INK4a , which has been mutated or deleted in several tumor tissues [30, 31] , plays an important role in regulating the cell cycle. Cdkn2a Ϫ/Ϫ mice [32] are susceptible to tumor generation and growth. However, whether p19 ARF and p16 INK4a have any roles in T cell differentiation is unclear.

The TNBS-induced murine colitis is an experimental model to be used to study the pathogenesis and therapy of human IBD. In this model, a combination of TNBS and colonic proteins induces excessive production of IFN-␥ by Th1 cells, which then activates monocytes/macrophages to produce a series of cytokines and chemokines, such as TNF-␣ and IL-6 [33] . Down-regulation of IL-10 is also observed in this model, indicating an immunomodulatary role of cytokines in the pathogenesis of such an animal model [34] .

In this report, we demonstrated that CD8 ϩ T cells could be induced to produce IL-10 in the presence of IL-4 upon activation. These CD8 ϩ Tregs suppressed CD4 ϩ T cell proliferation in vitro through IL-10-and cell contact-dependent mechanisms. Furthermore, we determined that the cell cycle regulatory protein Cdkn2a controlled IL-4-induced IL-10 production in CD8 ϩ T cells. Finally, these CD8 ϩ T cells played a protective role in TNBS-induced murine colitis in vivo.

10BiT mice, in which Thy1.1 expression is used to indicate IL-10 gene expression [35] , were generated by Dr. Casey Weaver from University of Alabama and kindly provided by Dr. Susan Kaech from Yale School of Medicine. Tiger mice, in which GFP-positive cells are used to indicate cells expressing IL-10 [10] and CD4dnIL-10R␣ mice [21] , were kindly provided by Dr. Richard Flavell from Yale School of Medicine. OT-1 mice (OVA 257-264 /H-2K b -restricted) were crossed with 10BiT mice to obtain 10BiT.OT-1 dualtransgenic mice. Cdkn2a Ϫ/Ϫ mice [32] were purchased from the NCI Mouse Repository (National Cancer Institute, Frederick, MD, USA) and were crossed with 10BiT mice to obtain Cdkn2a Ϫ/Ϫ .10BiT mice. C57BL/6J (B6 WT) mice were purchased from the Academy of Military Medical Science (Beijing, China). All animals were maintained under specific pathogen-free conditions at Nankai University. All experiments were performed in accordance with guidelines for animal care, created by Nankai University Experimental Animal Ethics Committee.

rmIL-2 and rmIL-4 and OVA 257-264 peptide (SIINFEKL) were purchased from R&D Systems (Minneapolis, MN, USA). rmGM-CSF and rmM-CSF were purchased from PeproTech (London, UK). Purified IFN-␥ mAb (clone XMG1.2) and purified CD210 mAb (clone 1B1.3a) were from eBioscience (San Diego, CA, USA). GolgiStop, anti-mouse CD3 mAb (clone 145-2C11) and anti-mouse CD28 mAb (clone PV1) were purchased from BD Biosciences (San Jose, CA, USA). FITC-conjugated anti-mouse CD11c (clone N418), PE-conjugated anti- 

Naïve CD8 ϩ T cells (CD8 ϩ CD62L hi CD44 lo ) were sorted from tiger or 10BiT mice and cultured with soluble anti-CD3 (2 g/ml) or Con A (2.5 g/ml) together with APCs in the presence of IL-4 (10 ng/ml) and IL-2 (2 ng/ml) for 48 h. After expansion for 2 days, cells were restimulated with PMA (50 ng/ml) and ionomycin (1 g/ml) for 6 h, and GFP ϩ or Thy1.1 ϩ T cells were detected by flow cytometry. APCs were from bone marrow-derived DCs [10] or fixed splenocytes treated with mitomycin C, as described previously [36] .

In Th2 conditions, naïve CD4 ϩ T cells (2.5ϫ10 6 /ml), sorted from reporter mice, were cultured with immobilized anti-CD3 (10 g/ml), soluble anti-CD28 (1 g/ml), IL-2 (2 ng/ml), IL-4 (10 ng/ml), and anti-IFN-␥ (10 g/ ml) for 48 h. After expansion for another 48 h, cells were restimulated for 6 h by PMA and ionomycin.

In Treg conditions, naïve CD4 ϩ T cells (2.5ϫ10 6 /ml) were cultured with immobilized anti-CD3 (10 g/ml), soluble anti-CD28 (1 g/ml), IL-2 (2 ng/ml), and TGF-␤ (5 ng/ml) for 72 h [37] .

Naïve CD8 ϩ T cells from 10BiT mice were activated with Con A together with APCs in the presence of IL-2 and IL-4 for 48 h. After expansion for 2 days, cells were restimulated with PMA and ionomycin as described above. Cells were sorted into CD8 ϩ Thy1.1 ϩ /Thy1.1 Ϫ T cells (2.5ϫ10 5 /well), which were then cocultured separately with CFSE-labeled, naïve CD4 ϩ T cells (2.5ϫ10 5 /well) from B6 WT or CD4dnIL-10R␣ mice (the ratio was 1:1) in 96-well flat-bottom plates and mixed with APCs, IL-2, and soluble anti-CD3 (2 g/ml) in the culture media. After 72 h, cells were harvested; the CFSE profiles of the CD4 ϩ T cells were analyzed by flow cytometry.

A 96-well Transwell plate with 0.4 m pore size was also used (Corning, Corning, NY, USA) for the coculture as described above. CFSE-labeled B6 WT naïve CD4 ϩ T cells (2.5ϫ10 5 /well), mixed with APCs, IL-2, and soluble anti-CD3 (2 g/ml) in the culture media, were placed in the lower chamber alone or together with sorted CD8 ϩ Thy1.1 ϩ T cells or with CD8 ϩ Thy1.1 ϩ T cells (2.5ϫ10 5 /well) in the upper chamber. Cells were cul-tured for 72 h, and CD4 ϩ T cells in the lower chambers were collected and stained with PerCP-Cy5.5-conjugated anti-mouse CD4, and CFSE dilution was analyzed by flow cytometry.

Mouse IL-10 ELISA kits were purchased from BioLegend, and ELISA was performed according to the manufacturer's protocol.

Total RNA was extracted from cells using TRIzol Reagent (Invitrogen) and reverse-transcribed by QuantScript RT Kit (Tiangen, Beijing, China). mRNA expression was quantified by real-time PCR. SYBR PreMix HotMaster Taq (Tiangen) was used in real-time PCR. The primer sequences used were listed in Supplemental Table 1 .

For global gene expression analysis, sorted CD8 ϩ GFP ϩ /GFP Ϫ T cells were lysed by TRIzol Reagent (Invitrogen). Microarray analysis was performed as described in the CapitalBio manual. Microarray data are available in the Gene Expression Omnibus database (Accession Number GSE37739; www. ncbi.nlm.nih.gov/geo/query/acc.cgi?accϭGSE37739).

FAM-conjugated siRNAs were synthesized by GenePharma (Shanghai, China). The specific siRNA sequences used were listed in Supplemental Table 2 , and Lipofectamine 2000 reagent (Invitrogen) was used for transfection. siRNA or N.C. was transfected into mouse splenocytes at a final concentration of 20 nM. Transfection was performed together with cells culturing in the presence of Con A, IL-2, and IL-4 for 48 h, and then cells

Naïve CD8 ϩ T cells from tiger mice were sorted, mixed with DCs (5:1 ratio), and stimulated with Con A (2.5 g/ml) ϩ IL-2 (2 ng/ml) or activated with soluble anti-CD3 (2 g/ml) ϩ IL-2 in the absence ("Ϫ") or presence of IL-4 (ϩIL-4; 10 ng/ml) for 48 h and then expanded for 2 days. Cells were then restimulated for 6 h with PMA and ionomycin and stained for CD8. One of three independent experiments is shown. (B) Naïve CD8 ϩ T cells from tiger mice were stimulated with Con A ϩ IL-2 ("Ϫ") in the presence or absence of IL-4 or anti-IFN-␥ (10 g/ml) for 48 h and then expanded for 2 days. Cells were then restimulated for 6 h with PMA and ionomycin and stained for CD8. One of three independent experiments is shown. (C) Splenocytes from 10BiT.OT-1 mice were stimulated with Con A or OVA 257-264 (1 g/ml) and IL-2 in the presence or absence of IL-4 for 48 h and then expanded for 2 days. Cells were then restimulated as described above and stained for CD8 and Thy1.1. One of three reproducible experiments is shown. (D) GFP ϩ CD8 ϩ T cells and GFP Ϫ CD8 ϩ T cells from A were sorted for IL-10 mRNA quantification by real-time PCR. Data are shown as the mean ϩ sem of three replicate measurements from three independent experiments. ***P Ͻ 0.001. (E) GFP ϩ CD8 ϩ T cells and GFP Ϫ CD8 ϩ T cells were sorted and cultured without IL-2. Supernatants were collected at different time-points for IL-10 ELISA. Data are shown as mean ϩ sem of triplicate of three experiments performed. ***P Ͻ 0.001. (F) Naïve CD4 ϩ and CD8 ϩ T cells were sorted from tiger and 10BiT reporter mice, cultured as described above, and analyzed by flow cytometry. One representative experiment of three is shown. FL1/2-H, Fluorescence 1/2-height; APC, allophycocyanin. transfected with FAM-conjugated siRNA (FAM ϩ cells) were sorted. Genesilencing efficiency was detected by real-time PCR.

Colitis was induced by administration of TNBS in B6 WT mice on Day 0, as described previously [38] . Mice were i.v.-injected with GFP ϩ CD8 ϩ (1ϫ10 6 cells/mouse) or GFP Ϫ CD8 ϩ (1ϫ10 6 cells/mouse) T cells or PBS vehicle, 24 h before and 24 h after TNBS administration. Mice treated with 50% ethanol were used as a negative control. Mice were weighed daily and killed on Day 5. Colon tissues were harvested, fixed for H&E staining, and RNA-extracted for IFN-␥ mRNA level analysis. Histopathological scoring rules of colon tissue sections were as described [39] .

A BD FACSCalibur (BD Biosciences) instrument was used for flow cytometry analysis; the software was CellQuest Pro (BD Biosciences) or FlowJo (Tree Star, Ashland, OR, USA). BD FACSAria (BD Biosciences) was used for cell sorting; the software was FACSDiva (BD Biosciences).

Statistical significance was evaluated by a two-tailed unpaired Student's t-test using InStat version 3.06 software for Windows (GraphPad, San Diego, CA, USA).

To define the role of IL-4 in CD8 ϩ T cell differentiation, naïve (CD44 lo CD62L hi ) CD8 ϩ T cells from tiger mice, in which GFPpositive cells are used to indicate cells expressing IL-10, were cultured with Con A or soluble anti-CD3 in the presence of IL-4. We demonstrated that IL-4 induced a significantly higher number of GFP ϩ CD8 ϩ T cells under Con A and anti-CD3 conditions (Fig. 1A) . In our preliminary studies, this effect was Fig. 1 , and GFP Ϫ CD8 ϩ T cells (gray, solid-line histograms) or GFP ϩ CD8 ϩ T cells (black, solid-line histograms) were gated, and surface molecules, as indicated, were analyzed. Isotype-matched control antibody staining for GFP Ϫ CD8 ϩ T cells (gray-filled histograms) or GFP ϩ CD8 ϩ T cells (black, dotted-line histograms) were also shown. One typical staining of three is shown. (B) Gene expression analysis comparing GFP ϩ CD8 ϩ with GFP Ϫ CD8 ϩ T cells from the above cultures. (C) GFP ϩ CD8 ϩ and GFP Ϫ CD8 ϩ T cells were sorted from the above cultures, and cDNA was prepared for analyzing the expression levels of indicated genes. Data represent mean ϩ sem of three combined experiments: *P Ͻ 0.05; **P Ͻ 0.01; ***P Ͻ 0.001. eomes, Eomesodermin; FUS, fused in sarcoma; T-bet, T-box expressed in T cells; Hif-1a, hypoxia-inducible factor 1␣; Nr4a1, nuclear receptor subfamily 4 group A member 1.

IL-4 dose-dependent (data not shown). The induction of IL-10 by IL-4 was independent of IFN-␥ (Fig. 1B) . Interestingly, using CD8 ϩ T cells from 10BiT.OT-1 transgenic mice, in which Thy1.1 expression is used to indicate IL-10 gene expression, we determined that IL-4 also induced antigen-responding IL-10 production by CD8 ϩ T cells (Fig. 1C) . To confirm the correlation between GFP and IL-10 expression level, GFP ϩ and GFP Ϫ CD8 ϩ T cells were sorted from the culture, and the ex- pression level of IL-10 was analyzed by real-time PCR. Indeed, GFP ϩ CD8 ϩ T cells expressed a significantly higher level of IL-10 ( Fig. 1D) , which was confirmed further by detection of IL-10 protein levels using ELISA (Fig. 1E) . Interestingly, under the same culture conditions, CD4 ϩ T cells expressed a significantly lower level of IL-10 (Fig. 1F) , and these two reporter mice (tiger and 10BiT) showed similar results. Therefore, both of these reporter mice were used in our following studies. Our results present new evidence that CD8 ϩ T cells could be induced to produce IL-10 by IL-4 upon activation.

To characterize these IL-4-induced, IL-10-producing CD8 ϩ T cells, a series of surface markers was analyzed by taking advantage of GFP ϩ /GFP Ϫ CD8 ϩ T cells ( Fig. 2A) . GFP ϩ CD8 ϩ T cells showed an activated phenotype, with higher expression of CD44, CD25, and CD69. These GFP ϩ CD8 ϩ T cells also expressed higher levels of ICOS and PD-1. Moreover, these cells expressed higher levels of IL-7R and IL-15R, indicating a stronger ability for survival of these CD8 ϩ T cells. Whereas these cells did not express Foxp3, no differences in the expression of GITR, CTLA4, and TGF-␤ were observed between GFP ϩ CD8 ϩ T and GFP Ϫ CD8 ϩ T cells.

To compare the broad gene profiles between these two types of CD8 ϩ T cells under the identical conditions, GFP ϩ and GFP Ϫ CD8 ϩ T cells were sorted, and cDNA microarray was performed. A series of functional genes and transcription factors was found to be differentially expressed between these two populations of CD8 ϩ T cells (Fig. 2B) . To confirm the expression profiles, cDNA samples prepared from these two cell populations were used for real-time PCR analysis. IL-10-producing CD8 ϩ T cells were confirmed to express higher levels of IL-4, IL-5, CCL3, CCL4, and perforin (Fig. 2C ).

Based on the fact that IL-10 is a suppressive cytokine, we sought to determine the suppressive function of IL-10-producing CD8 ϩ T cells. Naïve CD8 ϩ T cells from 10BiT mice were activated with Con A together with APCs in the presence of IL-2 and IL-4 to induce Thy1.1 ϩ expression. Cultured CD8 ϩ T cells were sorted into Thy1.1 ϩ and Thy1.1 Ϫ cells, followed by the coculture experiment. The proliferation of naïve CD4 ϩ T cells in the presence of Thy1.1 ϩ CD8 ϩ T cells were overlaid with those cultured with Thy1.1 Ϫ CD8 ϩ T cells. Indeed, Thy1.1 ϩ CD8 ϩ T cells dramatically suppressed proliferation of CD4 ϩ T cells (Fig. 3A) . Interestingly, addition of an anti-IL-10R antibody blocked the suppression (Fig.  3B) . The suppression effect was also impaired when using CD4 ϩ T cells isolated from CD4dnIL-10R␣ mice [21] , in which CD4 ϩ T cells express a high level of IL-10R but without the cytoplasmic domain for signaling through these receptors as the responders (Fig. 3C) . Our results thus demonstrate a critical role of IL-10 in mediating the suppressive effect of these IL-10-producing CD8 ϩ T cells.

To further determine whether the suppressive effect of IL-10-producing CD8 ϩ T cells was cell contact-dependent, transwell plates were used to separate the effector cells and responders. The separation indeed reduced the suppression effect, indicating a contactdependent mechanism (Fig. 3D) . Overall, IL-10-producing CD8 ϩ T cells suppressed CD4 ϩ T cell proliferation through IL-10-and cell contact-dependent mechanisms. 

To define the key transcription factors that might mediate the induction of IL-10 by IL-4, several siRNA sequences were screened to knock down those transcription factors that were upregulated in GFP ϩ CD8 ϩ T cells based on the cDNA microarray results (Fig. 2B) . The efficiency of these siRNAs reached Ͼ70%, as confirmed by real-time PCR (Fig. 4A) . Strikingly, only knockdown of cdkn2a and blimp1 significantly reduced IL-4-induced IL-10 expression (Fig. 4B) . To examine whether Cdkn2a and blimp1 had any interaction, the expression of blimp1 or Cdkn2a in the opposite siRNA-treated cells was analyzed. Interestingly, no direct effect on the expression level of IL-10 induced by one factor was caused by the opposite factor (Fig. 4C ).

To further study the role of Cdkn2a in IL-4-induced IL-10 expression, naïve CD8 ϩ T cells were sorted from B6 WT and Cdkn2a Ϫ/Ϫ mice, and we demonstrated by real-time PCR that CD8 ϩ T cells expressed significantly less IL-10 in the absence of Cdkn2a (Fig. 5A) . This conclusion was confirmed further by detection of IL-10 protein levels using ELISA (Fig. 5B) . Interestingly, IL-10 expression by iTregs and Th2-differentiated CD4 ϩ T cells was not affected (Fig. 5C and D) . Moreover, IL-10 expression from splenic macrophages and CD11c ϩ DCs triggered by LPS was also not affected (Fig. 5E) , and the same result was also obtained from the bone marrow-derived DCs and macrophages (data not shown). Taken together, our results thus define a unique role for Cdkn2a in CD8 ϩ T cell differentiation toward IL-10 production induced by IL-4.

To define the regulatory role of IL-10-producing CD8 ϩ T cells in vivo, a TNBS-induced colitis model was adapted with different populations of CD8 ϩ T cells transferred at different time-points of disease induction, as described in Materials and Methods. Mice injected with GFP ϩ CD8 ϩ T cells recovered their body weight more rapidly than mice treated with GFP Ϫ CD8 ϩ T cells or PBS at the later stage of colitis (Days 4 and 5; Fig. 6A ). The reduced inflammation in the GFP ϩ CD8 ϩ T cell-treated group was also confirmed by colon histopathology (Fig. 6B) . With the use of a standard scoring system, the GFP ϩ CD8 ϩ T cell-treated group reduced the severity of the disease significantly (Fig. 6C) . Finally, colon lengths of GFP ϩ CD8 ϩ T cell-treated mice were also obviously longer than those in the GFP Ϫ CD8 ϩ T cells or PBS-injected groups and had no significant differences from those in the control group (Fig. 6D) . These results indicate that IL-10-producing CD8 ϩ T cells play a protective role in the TNBS-induced murine colitis model.

To investigate further the mechanisms underlying the protective role of IL-10-producing CD8 ϩ T cells in TNBSinduced colitis model, we analyzed IFN-␥ mRNA expression levels in colon tissues among different groups. The results showed that IFN-␥ mRNA levels in the GFP ϩ CD8 ϩ T celltreated group were significantly lower than that of GF-P Ϫ CD8 ϩ T cells or PBS groups, which were similar to the control group and naïve mice (Fig. 6E) . Taken together, IL-10-producing CD8 ϩ T cells possibly protect mice from TNBS-induced colitis through inhibition of IFN-␥ production. 

CD8 ϩ T cells play an important role in protective immune responses against intracellular viral, bacterial infection, and tumor cells [1, 2, [40] [41] [42] . Similar to CD4 ϩ T cells, CD8 ϩ T cells have divergent, functional subsets, according to their cytokine profiles, including Tc1, Tc2, Tc17, and Tregs [43, 44] . However, less attention has been paid to the study of differentiation and transcriptional-controlling mechanisms of CD8 ϩ Tregs. In this study, we observed that CD8 ϩ T cells were induced to produce IL-10 by IL-4 and controlled by the cell cycle regulatory protein Cdkn2a. Mice were killed at Day 5 post-TNBS induction. Colon tissues from four groups of mice were fixed for H&E staining to determine disease severity. Representative sections were shown in B, and histopathological scores were shown in C (meanϮsem; ***PϽ0.001). Tissue sections were observed by Leica DM3000, and photos were obtained by the Leica Application Suite V3 software. Data were representative of three independent experiments (original magnification, ϫ200; original scale barsϭ 100 m). (D) Colon lengths were compared among different groups at Day 5 after treatment. One representative experiment of three was shown (meanϩsem; **PϽ0.01). (E) RNA was extracted from colon tissues, and the cDNA was then prepared for analyzing the expression levels of ifn-␥ among different groups. One representative experiment of three was shown (meanϩsem; *PϽ0.05).

IL-4 is a critical cytokine in the induction of CD4 ϩ T cell Th2 differentiation, and these CD4 ϩ T cells produce IL-4, IL-5, IL-13, and IL-10 cytokines [45, 46] . It has also been reported that together with IL-12, IL-4 also induces IL-10 production by CD8 ϩ T cells [47] . Interestingly, we demonstrated that together with IL-2, IL-4 was sufficient to induce naïve CD8 ϩ T cells to differentiate into IL-10-producing cells (Fig.  1) . Either IL-2 alone or IL-4 alone failed to induce significant levels of IL-10 (data not shown). In our preliminary experiments, cells cultured in the absence of IL-2 failed to survive and grow. We therefore speculated that IL-2 served as the growth and surviving factor in our system, and IL-4 played a critical role in inducing IL-10 production by CD8 ϩ T cells. Also, this process was dependent on TCR signaling-mediated activation, as even cytokine alone in the absence of T cell activation could lead to cell death (data not shown). Furthermore, this process was not affected by the presence of endogenous IFN-␥, as neutralizing IFN-␥ did not change the expression level of IL-10 ( Fig. 1 ). Our results thus provide an additional pathway for inducing IL-10 production by CD8 ϩ T cells.

One of the key findings in our study was the identification of the specific role of Cdkn2a in CD8 ϩ T cell differentiation, especially in IL-4-induced IL-10 secretion. Although quite a few signaling factors were up-regulated upon IL-4 treatment, only knockdown of cdkn2a and blimp1 suppressed IL-10 expression significantly (Fig. 4) , and it was evident that these two pathways had no direct interaction (Fig. 4) . The critical role of Cdkn2a was confirmed further using Cdkn2a Ϫ/Ϫ mice, and the effect was restricted to CD8 ϩ T cells but was not identified in CD4 ϩ T cells or APCs (Fig. 5) . Interestingly, knockdown of gata3 failed to affect IL-4-induced IL-10 expression (Fig. 4) , implying that Cdkn2a is not downstream of GATA3. It is unclear whether Cdkn2a is downstream of phosphorylated Stat6 or directly downstream of TCR signaling. To the best of our knowledge, this is the first report to define the role of this cell cycle regulatory protein in CD8 ϩ T cell differentiation, and further studies are required to explore the function of Cdkn2a in other cytokine-induced differentiation of CD8 ϩ T cells and their detailed signaling pathways.

Defining the function of IL-10-producing T cells is a challenge. Recently, several reporter mice have been made to use GFP or Thy1.1 to indicate IL-10-expressing cells, such as tiger mice [10] and 10BiT mice [35] . By taking advantage of IL-10 GFP or Thy1.1 reporter mice, we demonstrated that IL-10-producing CD8 ϩ T cells possessed a strong, suppressive function in vitro via IL-10-and cell contact-dependent mechanisms (Fig.  3 ) and in vivo through suppressing IFN-␥ production (Fig. 6) . Interestingly, these IL-4-induced, IL-10-producing CD8 ϩ T cells expressed suppressing molecules (PD-1, CTLA-4), as well as a high level of Tc2 cytokines (Fig. 2) . Several subsets of CD8 ϩ Tregs have been identified, such as CD8 ϩ CD122 ϩ Tregs [6] , CD8 ϩ CD28 Ϫ T cells [7] , and Qa-1-restricted CD8 ϩ T cells [48] . It is interesting to define whether this population of IL-10-producing CD8 ϩ T cells is a new subset of CD8 ϩ Tregs. Further validation would be needed to study the gene expression profiles and functions of these IL-4-induced, IL-10-producing CD8 ϩ T cells.

It has been shown that IL-10 suppresses Th1-mediated inflammation in several animal models [49 -52] . TNBS-induced colitis is a typical Th1-mediated inflammation [53] . We demonstrated a protective role of these IL-10-producing CD8 ϩ T cells in this colitis model (Fig. 6) . Interestingly, colon tissue in mice injected with IL-10 ϩ CD8 ϩ T cells showed a significantly decreased level of IFN-␥ (Fig. 6) , implying a possible mechanism of these IL-10 ϩ CD8 ϩ T cells in TNBS-induced colitis through suppressing the Th1 response. Further studies are in progress to test the detailed molecular mechanisms of this population of CD8 ϩ Tregs in the TNBS-induced colitis model and its function in other inflammatory models.

In summary, we have defined a population of CD8 ϩ Tregs that was induced by IL-4 and controlled by the cell cycle regulatory protein Cdkn2a, which is the first time it was reported to be involved in T cell differentiation. These IL-10-producing CD8 ϩ T cells played a protective role in TNBS-induced colitis. Further studies to understand the pathophysiological function of this subset of CD8 ϩ T cells will enable us to facilitate the effective CD8 ϩ T cells in protective immune responses against pathogens and tumors. 

Tc1 and Tc2 effector cell therapy elicit long-term tumor immunity by contrasting mechanisms that result in complementary endogenous type 1 antitumor responses

Migration kinetics and final destination of type 1 and type 2 CD8 effector cells predict protection against pulmonary virus infection

Allergen immunotherapy in intermittent allergic rhinitis reduces the intracellular expression of IL-4 by CD8ϩ T cells

Adoptive transfer of tumor-specific Tc17 effector T cells controls the growth of B16 melanoma in mice

Anomalous type 17 response to viral infection by CD8ϩ T cells lacking T-bet and eomesodermin

CD8ϩCD122ϩ regulatory T cells (Tregs) and CD4ϩ Tregs cooperatively prevent and cure CD4ϩ cell-induced colitis

Regulatory functions of CD8ϩCD28Ϫ T cells in an autoimmune disease model

Highly activated cytotoxic CD8 T cells express protective IL-10 at the peak of coronavirus-induced encephalitis

Effector T cells control lung inflammation during acute influenza virus infection by producing IL-10

Expression of interleukin-10 in intestinal lymphocytes detected by an interleukin-10 reporter knockin tiger mouse

HIV-specific IL-10-positive CD8ϩ T cells are increased in advanced disease and are associated with decreased HIV-specific cytolysis

Intrahepatic virus-specific IL-10-producing CD8 T cells prevent liver damage during chronic hepatitis C virus infection

IL-10: the master regulator of immunity to infection

CD4ϩCD25ϩ regulatory T cells cure murine colitis: the role of IL-10, TGF-␤, and CTLA4

Tolllike receptor 2 suppresses immunity against Candida albicans through induction of IL-10 and regulatory T cells

Induction of IL-10-producing CD4ϩCD25ϩ T cells in animal model of collageninduced arthritis by oral administration of type II collagen

Interleukin 10 acts on regulatory T cells to maintain expression of the transcription factor Foxp3 and suppressive function in mice with colitis

IL-10 inhibits cytokine production by activated macrophages

IL-10 acts on the antigen-presenting cell to inhibit cytokine production by Th1 cells

Th17 cells express interleukin-10 receptor and are controlled by Foxp3 and Foxp3ϩ regulatory CD4ϩ T cells in an interleukin-10-dependent manner

Memory/effector (CD45RB(lo)) CD4 T cells are controlled directly by IL-10 and cause IL-22-dependent intestinal pathology

Interleukin-10-secreting type 1 regulatory T cells in rodents and humans

GATA-3 directly remodels the IL-10 locus independently of IL-4 in CD4ϩ T cells

The transcription factors Blimp-1 and IRF4 jointly control the differentiation and function of effector regulatory T cells

The protooncogene c-Maf is an essential transcription factor for IL-10 gene expression in macrophages

STAT-1-mediated repression of monocyte interleukin-10 gene expression in vivo

Cutting edge: involvement of the type I IFN production and signaling pathway in lipopolysaccharide-induced IL-10 production

IL-10 gene expression is controlled by the transcription factors Sp1 and Sp3

The regulation of INK4/ARF in cancer and aging

Expression, deletion [was deleton] and mutation of p16 gene in human gastric cancer

Abnormalities of tumor suppressor gene p16 in pancreatic carcinoma: immunohistochemical and genetic findings compared with clinicopathological parameters

Role of the INK4a locus in tumor suppression and cell mortality

Hapten-induced model of murine inflammatory bowel disease

The interrelated roles of TGF-␤ and IL-10 in the regulation of experimental colitis

Regulatory T cells expressing interleukin 10 develop from Foxp3ϩ and Foxp3Ϫ precursor cells in the absence of interleukin 10

Burn-induced immunosuppression: attenuated T cell signaling independent of IFN-␥-and nitric oxide-mediated pathways

Conversion of peripheral CD4ϩCD25Ϫ naive T cells to CD4ϩCD25ϩ regulatory T cells by TGF-␤ induction of transcription factor Foxp3

All-trans retinoic acid down-regulates inflammatory responses by shifting the Treg/Th17 profile in human ulcerative and murine colitis

Interferon-␥ is causatively involved in experimental inflammatory bowel disease in mice

Anti-viral CD8 T cells and the cytokines that they love

CD8 T cell help for innate antitumor immunity

CD8 T cells can protect against an intracellular bacterium in an interferon ␥-independent fashion

Tc17, a unique subset of CD8 T cells that can protect against lethal influenza challenge

Functions of CD8 T-cell subsets secreting different cytokine patterns

Differentiation of effector CD4 T cell populations

IL-4 induces differentiation and expansion of Th2 cytokine-producing eosinophils

Cytokine-induced IL-10-secreting CD8 T cells represent a phenotypically distinct suppressor T-cell lineage

Inhibition of follicular T-helper cells by CD8(ϩ) regulatory T cells is essential for self tolerance

A critical role for IL-10 in limiting inflammation during toxoplasmic encephalitis

Bacteria-triggered CD4(ϩ) T regulatory cells suppress Helicobacter hepaticus-induced colitis

Interleukin-10 modulates susceptibility in experimental cerebral malaria

Murine cytomegalovirus infection down-regulates MHC class II expression on macrophages by induction of IL-10

Treatment of murine Th1-and Th2-mediated inflammatory bowel disease with NF-B decoy oligonucleotides

This work was supported by grants from the National Science Foundation of China (31170858 and 30007101), International S&T Cooperation Program of China (2010DFB34000) and was supported by 111 project (B08011). We thank Dr. Mark Bartlam for manuscript revision.

The authors have no financial or commercial conflicts of interest.