Stat4-Dependent, T-Bet-Independent Regulation of IL-10 in NK Cells

ORIGINAL ARTICLE Stat4-dependent, T-bet-independent regulation of IL-10 in NK cells

LR Grant1, Z-J Yao2, CM Hedrich1, F Wang1, A Moorthy1, K Wilson2, D Ranatunga1 and JH Bream1 1W Harry Feinstone Department of Molecular Microbiology and Immunology, Johns Hopkins University, Bloomberg School of Public Health, Baltimore, MD, USA and 2Lymphocyte and Cell Biology Section, Molecular Immunology and Inflammation Branch, National Institute of Musculoskeletal and Skin Diseases, Bethesda, MD, USA

Interleukin-10 (IL-10) is intensely studied, yet little is known about the mechanisms that control IL-10 expression. We identified striking similarities between IL-10 and interferon-g (IFN-g) regulation in mouse natural killer (NK) cells. Like IFN-g, IL-10 expression is induced by IL-2 and IL-12 and IL-2 þ IL-12 stimulation is synergistic. Unlike IFN-g, neither IL-18 nor Ly-49D cross- linking induced IL-10 expression however. Additionally, the IL-12 homologs IL-23 and IL-27 also do not regulate NK cell- specific IL-10. We determined that a small population of NK cells accounts for IL-10 production. The induction of IL-10 by IL-2 þ IL-12 treatment in NK cells appears to be biphasic, with an initial burst of expression which diminishes by 12 h but spikes again at 18 h. We determined that much like IFN-g, Stat4 is largely required for IL-12-induced IL-10. Conversely, we observed normal induction of IL-10 in T-bet-deficient NK cells. We identified a Stat4-binding element in the fourth intron of the Il10 gene, which is completely conserved between mouse and human. This intronic Stat4 motif is within a conserved noncoding sequence, which is also a target for cytokine-induced histone acetylation. These findings highlight tissue- and receptor-specific IL-10 regulatory mechanisms, which may be part of an early feedback loop. Genes and Immunity (2008) 9, 316–327; doi:10.1038/gene.2008.20; published online 10 April 2008

Keywords: natural killer cells; cytokines; IL-10; Stat4; IL-12

Introduction cess results in severe pathology including tissue destruc- tion and death. The immunoregulatory cytokines IL-10 Maintaining a proper balance between inflammatory and and IFN-g play fundamental roles in this complex anti-inflammatory responses is essential for effective process. A key to understanding this delicate balance is immunity against infectious pathogens while repressing by evaluating the molecular and biochemical mechan- the onset of autoimmunity.1 The obligatory reaction of isms that govern the expression of these genes. Remark- organisms to foreign pathogens is to launch an inflam- ably, these processes remain poorly understood. matory response that is initiated by cells of the innate IL-10 is expressed by a variety of cell types including immune system, such as professional antigen-presenting macrophages, DC subsets, B cells, several T-cell sub- cells (APC) and natural killer (NK) cells. The inflamma- populations and T regulatory cells.3 Primarily, IL-10 tory process is driven largely by the secretion of functions as an anti-inflammatory molecule that inhibits proinflammatory cytokines, including macrophage- and macrophage/DC activation and maturation4,5 and antag- dendritic cell (DC)-derived interleukin-12 (IL-12), as well onizes the proinflammatory cytokines IL-1b, IL-6, IL-8, as NK cell-derived interferon-g (IFN-g). NK cells prime TNF-a and notably IL-12.6,7 Paradoxically, the biological macrophages and DC during primary responses by role of IL-10 is not limited to inactivation of APCs; IL-10 providing early sources of IFN-g and potentially other is also known to enhance B cell, granulocyte, mast cell cytokines. There is considerable cellular crosstalk be- and keratinocyte growth/differentiation, as well as NK tween these cell populations, which in turn influences cell and CD8 þ cytotoxic T-cell activation.7–11 In addition, the activation status of each population and provides the IL-10 has been shown to enhance NK cell proliferation microenvironmental conditions necessary to employ and/or production of IFN-g.12–14 adaptive immune measures, such as the differentiation Various cell types can be divided into distinct of T helper cell populations.2 Without appropriate subpopulations based on IL-10 expression patterns. feedback mechanisms, however, the inflammatory pro- In the myeloid compartment, alternatively activated macrophages, type II macrophages and some DC subsets are the principle sources of IL-10 production.15 Another Correspondence: Dr JH Bream, Department of Molecular Micro- potentially important, yet underappreciated, source biology and Immunology, Johns Hopkins University, Bloomberg of IL-10 is derived from the NK cell compartment. School of Public Health, Baltimore, MD 21205-1901, USA. E-mail: [email protected] NK cells and NKT cells are integral parts of the innate Received 20 November 2007; revised 11 February 2008; accepted 14 immune system, and are critical to host defense against February 2008; published online 10 April 2008 infectious pathogens, such as herpes viruses16,17 and Regulation of NK cell-specific IL-10 expression LR Grant et al 317 choriomeningitis virus.18 It is well known that NK cells provide early sources of IFN-g in vivo. Our group and 400 others have shown that IL-2, IL-12 and their combination are potent regulators of IFN-g expression in NK cells.19 300 Surprisingly, IL-2 and IL-12 also induce IL-10 expression in both mouse and human NK cells and the combination 200 of IL-2/IL-12 is synergistic.14,20 Interestingly, clinical administration of recombinant 100 human IL-12 to cancer patients resulted in elevated IL-10 (Fold Induction) plasma levels of IL-10 and decreased IFN-g.21 The 0 IL-2 IL-12 IL-18 Ly-49D 2+12 biological significance of IL-12 induction of IL-10 is not clear but has been suggested to be part of an important Stimulation feedback mechanism during infection to likely restrict inflammatory responses at specific tissues sites.22 There- 2000 fore, there is a link between IL-12 and regulation of IL-10, particularly in NK cells, however, the molecular mechan- 1600 isms have not been explored. Here we report unexpected similarities and some key 1200 disparities between regulation of IL-10 and IFN-g expression in NK cells. Like IFN-g, IL-2 and IL-12 800 synergistically induce IL-10 expression from mouse and IL-10 (pg/ml) human NK cells which is largely dependent on Stat4. 400 However, other known regulators of IFN-g (IL-18, Ly-49D and the transcription factor T-bet), do not control 0 IL-10 expression in NK cells. In searching for molecular NS IL-2 IL-12 IL-18 Ly-49D 2+12 targets for Stat4, we identified a conserved Stat4-like Stimulation element within a conserved noncoding sequence (CNS) Figure 1 Induction of interleukin-10 (IL-10) expression from in the fourth intron of the Il10 gene. This motif binds cultured natural killer (NK) cells in response to cytokine stimula- IL-12-induced Stat4 and the CNS site is a target for tion. (a) Cultured NK cells were stimulated for 6 h with the cytokine-induced histone acetylation. Thus, we provide indicated cytokines and mRNA expression was measured by real- molecular insight into a surprising parallel between time PCR. Real-time PCR data are represented as a fold increase IL-10 and IFN-g gene regulatory mechanisms. over nonstimulated cultured NK cells. The data are shown as the mean±s.d. from three independent experiments. (b) NK cell culture supernatants were analyzed for IL-10 production after 24 h stimulation by enzyme-linked immunosorbent assay (ELISA). The Results data are shown as the mean±s.d. from three independent experiments. Similarities and disparities in IL-10 and IFN-g induction in cultured NK cells In previous studies we established spleen-derived, IL-2- Freshly isolated splenic NK cells express IL-10 in response cultured mouse NK cells (cultured NK cells) as a to IL-2, IL-12 and IL-2 þ IL-12 costimulation representative model of IFN-g regulation as compared To confirm that the IL-10 expression patterns observed in to freshly isolated NK cells.19 We have identified notable cultured NK cells were also present in primary NK cells, similarities and stark differences between the induction we isolated NK1.1 þ CD3À cells from nylon wool-passed signals for IL-10 expression and IFN-g. Utilizing cultured spleens by negative flow cytometric sorting. The result- NK cells as a model, we found that IL-10 is regulated by ing populations were greater than 95% NK1.1 þ CD3À some of the same receptors, which control IFN-g pure. As we found in cultured NK cells, stimulation of expression. As shown in Figure 1, IL-2 and IL-12 freshly isolated NK cells with IL-2, IL-12 or IL-2 þ IL-12 treatment independently induce IL-10 mRNA and resulted in induction of IL-10 mRNA (Figure 2a) and protein production. In addition, as we and others have protein (Figure 2b). Likewise, IL-18 treatment did not reported for IFN-g,19,23,24 IL-2 þ IL-12 costimulation of induce the expression of IL-10 in fresh NK cells. Our cultured NK cells resulted in a highly synergistic group and others have reported that the response of IFN- induction of IL-10 mRNA (Figure 1a) and IL-10 secretion g expression is attenuated in fresh NK cells as compared (Figure 1b). The similarities in regulatory signals for IFN-g to highly activated, cultured NK cells. Similarly, we and IL-10 are not absolute, however, as other IFN-g- found that while the trends of IL-10 induction between inducing agents (IL-18 and Ly-49D cross-linking) have cultured NK cells and fresh NK cells are comparable, the no affect on IL-10 expression. Similarly, the IL-12 levels of IL-10 are lower and the kinetics are delayed. homologs IL-23 and IL-27 did not induce IL-10 indivi- Thus, both primary and cultured NK cells express IL-10 dually or in combination with one another, IL-2 or IL-12 in response to IL-2 and IL-12 and the combination of (data not shown). Induction of IFN-g was confirmed for IL-2 þ IL-12 exposure is synergistic. all experiments at the mRNA level by real-time PCR and by enzyme-linked immunosorbent assay (ELISA) in IL-10 is expressed by a small subset of cultured NK cells culture supernatants (data not shown). These data Having established that NK cells are a source of IL-2/ suggest common regulatory pathways for IL-10 and IL-12 induced IL-10, we wanted to determine the cellular IFN-g in NK cells and may also point to shared, tissue- relationship between IL-10-expressing and IFN-g-expres- specific epigenetic/genomic control mechanisms. sing NK cell populations. We used intracellular cytokine

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 318 75 Unexpectedly, there is another burst of expression 18 h after cytokine treatment that quickly subsides by 24 h and is nearly nondetectable at 48 h. Since large amounts 50 of IFN-g are in the culture supernatants by 12 h, we considered that IFN-g production by NK cells in culture may be inhibiting IL-10 expression at 12 h, or even possibly enhancing IL-10 mRNA profiles at 18 h. To 25 exclude this possibility, we either pretreated cultured NK

IL-10 (Fold Induction) cells for 4 h with a neutralizing IFN-g antibody or with rmIFN-g prior to IL-2/IL-12 stimulations and found no 0 significant difference in IL-10 mRNA or protein levels at IL-2 IL-12 IL-18 IL-2+12 6 and 24 h, respectively (data not shown). We also stimulated cultured NK cells derived from IfngÀ/À mice 300 and found no enhancement of IL-10 expression (data not 250 shown). While we cannot exclude the possibility that 200 other factors expressed by NK cells may contribute to IL-10 mRNA kinetic profiles at later time-points, it 150 appears that IFN-g does not play a role in IL-10 100 responses to IL-2/IL-12. IL-10 (pg/ml) 50 To determine the stability of IL-10 mRNA transcripts in NK cells, cytokine-stimulated NK cells were treated 0 with the transcriptional inhibitor actinomycin D (Act D) NS IL-2 IL-12 IL-18 IL-2+12 and IL-10 mRNA levels were assessed at 2 and 4 h post Stimulation Act D treatment (Figure 4b). As mentioned above, we Figure 2 Induction of interleukin-10 (IL-10) expression from determined that maximal IL-10 expression was achieved freshly isolated natural killer (NK) cells in response to cytokine by 3 h post stimulation and generally maintained for stimulation. (a) Primary NK cells (NK1.1 þ /CD3À) were stimulated 6–8 h. Thus, we choose the 3 h time point (control; 0 h) as for 24 h with the indicated cytokines and mRNA expression was a stable point of comparison for Act-treated cells. For measured by real-time PCR. Real-time PCR data are represented as a fold increase over nonstimulated, primary NK cells. (b) Primary transcriptional blockade, NK cells were stimulated with NK cell culture supernatants were analyzed for IL-10 production cytokines for 3 h, then Act D was added to cultures and after 48 h stimulation by enzyme-linked immunosorbent assay mRNA was harvested 2 or 4 h later. Results are shown as (ELISA). The data are representative of two separate experiments. a percent of control. Similar to previous reports of IFN-g stability in NK cells, IL-10 mRNA half-life is relatively short with over a staining (ICS) of nonstimulated (NS) and IL-2 þ IL-12 75% reduction in IL-10 mRNA levels in all cytokine treated cultured NK cells to determine cytokine expres- stimulation conditions after 2 h of Act D treatment.26 sion patterns (Figure 3). As shown in Figures 3a and b, Nevertheless, there is still detectible IL-10 mRNA up to IL-2 þ IL-12 stimulation induces a small population of 4 h following blockade of transcription. These data NK1.1 þ CD3À NK cells to express IL-10 (approximately indicate that IL-2/IL-12 induced IL-10 mRNA in NK 0.8% after subtraction of IL-10 þ cells in the NS control). cells is relatively unstable, regardless of the stimulation No IL-10-expressing NK1.1 þ CD3 þ NKT cells were conditions but may be sustained to a small degree for at detected (data not shown). Interestingly, Loza and least 4 h. Perussia25 recently described a similar small percentage of IL-10-expressing, human CD56 þ NK cells in response Partial involvement of the p38 MAP kinase pathway in IL-10 to IL-2 and IL-12 stimulation. induction As expected, when NK cells were also stained for IFN-g In addition to utilizing JAK/STAT signaling mechanisms, by ICS, we found a large induction of IFN-g-expressing IL-2 and IL-12 also activate the mitogen-activated protein cells after IL-2 þ IL-12 treatment within the NK1.1 þ kinase (MAPK) pathway. Our group and others have fraction (Figures 3c and d). Interestingly, we found that reported a role in particular for the MAPK p38 in within the population of IL-10-expressing NK1.1 þ cells, regulating IL-12 as well as IL-2 þ IL-12 induced IFN-g about 70% coexpress IFN-g. Albeit it a small fraction, expression.19,27,28 Therefore, we examined the effects of the majority of NK cells that produce IL-10 in response to p38-specific inhibitors on IL-10 induction by IL-2 and IL- IL-2 þ IL-12 stimulation also coexpress IFN-g. 12 (Figure 5). The synergistic effect of IL-2 þ IL-12 stimulation on IL-10 production was significantly Kinetics of IL-10 mRNA expression in cultured NK cells reduced by both p38 inhibitors as compared to is bi-phasic and IL-10 mRNA half-life is short in NK cells the dimethyl sulfoxide (DMSO) controls (SB202190, We know from previous studies in NK cells that Po0.01 and SB203580, Po0.05). SB202190 consistently induction of IFN-g mRNA by IL-2 þ IL-12 treatment is yielded a greater degree of inhibition as compared to rapid and peaks around 3–4 h post stimulation.23 Given another p38 inhibitor SB203580 (on average, 31% of the similarities between IFN-g and IL-10 induction in NK control for SB202190 compared to 44% of control for cells we examined the kinetics of IL-10 mRNA expres- SB203580). These findings suggest that the p38 pathway sion. Like IFN-g, IL-10 mRNA expression is induced has a role in IL-2 þ IL-12 regulation of IL-10 expression. maximally by 3 h and is sustained at 6–8 h following However, the fact that inhibition of the p38 MAPK cytokine stimulation (Figure 4a). However, IL-10 trans- pathway resulted in only partial blockage of IL-10 cripts are greatly diminished 12 h following stimulation. expression suggests multiple regulatory mechanisms

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 319 NS IL-2 + IL-12 104 104 0.01% 0.14% 0.15% 0.95%

103 103

102 102 mIL-10

101 101

93.39% 93.72% 100 100 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 NK 1.1

4 104 10 0.09% 0.01% 0.23% 0.68%

3 103 10

2 102 10 mIL-10

1 101 10

1.58% 46.53% 0 100 10 0 1 2 3 4 100 101 102 103 104 10 10 10 10 10 mIFN-γ

Figure 3 Intracellular staining of cultured natural killer (NK) cells for cytokine expression by flow cytometry. (a) Nonstimulated (NS) and (b) interleukin-2 (IL-2) þ IL-12-treated NK cells were stained with antibodies directed against NK1.1, CD3, IL-10, and IFN-g 6 h later. (c)NS and (d) IL-2 þ IL-12-treated NK cells were also stained with the same antibody cocktail 6 h later. Cells are gated on NK1.1 þ in Figures 4c and d. The data are representative of three separate experiments. may account for the synergistic effects of IL-2 and IL-12 largely, but not entirely responsible for the functional on IL-10 expression. synergy between IL-2 and IL-12 in regulating IL-10 expression. Since Stat4 does not completely explain IL-2 þ IL-12 regulation of IL-10 is mediated largely by Stat4 IL-2 þ IL-12 induction of IL-10 and the T-box transcrip- but not T-bet tion factor T-bet is required for optimal expression of IL-12 regulates IFN-g expression in Stat4-dependent IFN-g in NK cells, we asked if T-bet is involved in manner.29,30 Given the similarities in IL-10 and IFN-g regulation of IL-10 expression.19,31 Cultured NK cells regulation in NK cells, we hypothesized that Stat4 would were generated from T-bet-deficient mice and tested for be involved in IL-12-induction of IL-10, including the induction of IL-10. There was no effect of the absence of synergy induced by IL-2 þ IL-12 stimulation. To test this T-bet on induction of IL-10 by IL-2 or IL-12 alone or the hypothesis, we generated spleen-derived, cultured NK synergistic effect of IL-2 þ IL-12 (Figure 6c). Together cells from Balb/c Stat4À/À mice and stimulated them these data point to an important role of Stat4 in with cytokines as performed previously. As expected, regulating both the Il10 and Ifng genes, yet also indicate IL-2 induction of IL-10 is essentially intact in the absence a key disparity given that T-bet is required for optimal of Stat4, however, IL-12 fails to induce IL-10 mRNA IFN-g expression but dispensable for transcriptional expression in Stat4À/À NK cells (Figure 6a). In addition, regulation of IL-10 in NK cells. the synergistic activity of IL-2 þ IL-12 on IL-10 was dramatically reduced in the absence of Stat4, yet was not A conserved, intronic Stat4 element is located within a CNS completely ablated. In line with IL-10 mRNA expression, region (CNS þ 3.10) that is also a target for cytokine-induced IL-12 and IL-2 þ IL-12-induced IL-10 production was chromatin remodeling substantially reduced in Stat4-deficient NK cells In order to identify potential mechanisms of Stat4 in (Figure 6b). As observed in wild-type cells, we found regulating IL-10 expression, we scanned the Il10 gene no induction of IL-10 by IL-18 in Stat4 null NK cells. and surrounding sequences for potential STAT-binding Induction of IFN-g by IL-2 and IL-18 was intact in Stat4 elements. Based on a computer search string of common deficient NK cells (data not shown). These data suggest STAT-binding motifs, we identified a putative Stat4 that Stat4 is necessary for IL-12-induction of IL-10, and element in the fourth intron of the mouse Il10 gene

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 320 * 400 300 control ** IL-2 dmso sb202190 IL-12 300 sb203580 200 2+12

200 100 IL-10 (pg/ml) IL-10 mRNA (Fold Induction) 100 0 3h 6h 12h 18h 24h 48h

Time Point 0 NS IL-2 IL-12 IL-2+12 100 Stimulation Control (0h) 80 Figure 5 Effect of p38 MAPK inhibitors on cytokine-induced Act D (2h) interleukin-10 (IL-10) production from cultured natural killer (NK) Act D (4h) cells. Cultured NK cells were pretreated with p38 inhibitors 60 (sb202190 or sb203580), vehicle control (dimethyl sulfoxide; DMSO) or media (control) for 30 min prior to stimulation with cytokines. 40 NK cells were then stimulated overnight and NK cell culture supernatants were analyzed for IL-10 production by enzyme-linked 20 immunosorbent assay (ELISA). The IL-10 production data are shown as the mean±s.d. from three independent experiments. % of Control IL-10 mRNA Statistical significance is indicated by **(P 0.01) or *(P 0.05). 0 o o IL-2 IL-12 2+12 Stimulation Figure 4 Interleukin-10 (IL-10) mRNA kinetics and half-life in did not block or displace formation of the IL-12-induced cultured natural killer (NK) cells. (a) Cultured NK cells were complex (lane 7). Likewise, we excluded the possibility stimulated IL-2, IL-12 or IL-2 þ IL-12 for the indicated time points that the related molecule Stat5 was bound to this region and mRNA expression was measured by real-time PCR. IL-10 of the Il10 gene by incubating with anti-Stat5A and mRNA expression was measured by real-time PCR which is shown as a fold increase over the nonstimulated NK cell control. The data Stat5B antibodies which also did not block formation of are representative of two separate experiments. (b) Cultured NK the complex (lane 8). When an anti-Stat4 antibody was cells were treated with cytokines for 3 h and harvested for control used however, the IL-12-induced complex was blocked, mRNA levels (control; 0 h) or actinomycin D (Act D) at 5 mgmlÀ1, indicating that IL-12-induced Stat4 binds to the Stat4 was added 3 h after cytokine stimulation. Act D-treated cells were element in the fourth intron of the Il10 gene. harvested 2 h after treatment (Act D 2 h), and 4 h after Act D To confirm that IL-12-activated Stat4 interacts with the treatment (Act D 4 h). Data are represented as a percent of control IL-10 mRNA expression. Dashed line indicates 50% of control. endogenous Il10 gene in vivo, we employed chromatin immunoprecipitation (ChIP) analysis. As shown in Figure 7c, cultured NK cells that were NS (lane 3) or treated with IL-2 alone (lane 6) did not recruit Stat4 to the based on sequence identity. As shown in Figure 7a, the Il10 gene. However, when cells were stimulated either STAT-like sequence maps to a CNS þ 3.10 which has with IL-12 alone (lane 9) or IL-2 þ IL-12 (lane 12), Stat4 approximately 72% homology over more than 100 bp was immunoprecipitated, bound to the fourth intron of between mouse and human (http://genome.lbl.gov/ the Il10 gene. To control for nonspecific antibody vista/index.shtml). The STAT-motif identified in the interactions, normal rabbit immunoglobulin G (IgG) mouse Il10 gene is conserved in humans as we identified was used as an immunoprecipitation control but no a STAT-like element with 100% identity in the human Il10 evidence of nonspecific binding was observed (lanes 2, 5, gene, also located in the identified CNS region in fourth 8, 11). Thus, based on EMSA and ChIP analysis we intron. To determine if IL-12-induced Stat4 to bind to this conclude that IL-12 stimulation results in the recruitment element, we preformed electromobility shift assays of Stat4 to the mouse Il10 gene. (EMSA) using this region as a probe. When cultured Since the CNS þ 3.10 region is highly conserved and NK cells were treated with media alone (NS, lane 1), or recent reports identified a DNaseI hypersensitivity site in with IL-2 (lane 2), no unique complex(es) were formed the fourth intron of the mouse Il10 gene in T cells,32,33 we (Figure 7b). However, when NK cells were stimulated questioned if other chromatin-modifying mechanisms with IL-12 (lanes 3 and 4), a unique complex was formed. were targeted to this region in NK cells. Thus, we used To determine if Stat4 was bound in the complex, we used ChIP analysis to address the state of histone acetylation a Stat4-consensus sequence to compete for binding by following cytokine stimulation (Figures 8a and b). excess cold competition. The consensus Stat4 oligo Compared to nonstimulated NK cells (lane 3), we found blocked the formation of the IL-12-induced complex no evidence of enhanced histone H3 acetylation in IL-2- (lane 5), however an identical oligo with the Stat4 site treated cells (lane 6). Nonetheless, some degree of mutated failed to compete for binding (lane 6). To constitutive H3 acetylation was observed in NS and confirm Stat4 binding, we performed supershift analyses IL-2-stimulated cells (lanes 3 and 6). When NK cells were by first adding an irrelevant antibody against SP-1 which stimulated with IL-12 alone or in combination with IL-2,

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 321 140 amounts are represented as a relative ratio as compared to the input for each sample. Collectively, these data 120 WT suggest that the fourth intron of the mouse Il10 gene is a Stat4-/- 100 target for both cytokine-induced chromatin modifications and transcription factor recruitment. 80

60 Discussion 40

IL-10 (Fold Induction) A number of reports have accumulated over recent years 20 suggesting a surprising role for IL-12 in regulating IL-10 0 expression in the lymphoid compartment, often in IL-2 IL-12 IL-18 2+12 conjunction with IFN-g.22 It is well established that IL-2 Stimulation and IL-12 regulate IFN-g autonomously and together are highly synergistic in T and NK cells.34 Similarly, IL-2 and 700 IL-12 have been reported to induce IL-10 expression individually and together are synergistic in both T and 600 WT NK cells, but the mechanisms have not been ex- Stat4-/- plored.35,36 In mouse and human T cells, IL-12 has been 500 implicated in the differentiation of T-cell subpopulations 37,38 þ 400 that express both IL-10 and IFN-g. These IL10 IFN-g þ T cells are particularly well described in response 300 to infection with certain intracellular pathogens, and are IL-10 (pg/ml) thought to involve T-bet.39–42 Other reports indicate a 200 direct role for IL-12 induction of IL-10 expression in vivo. Clinical administrations of recombinant human IL-12 100 to cancer patients resulted in elevated plasma levels of 21 0 IL-10 but decreased IFN-g. The biological significance NS IL-2 IL-12 IL-18 IL-2+12 of IL-10 and IFN-g coexpression in NK or T-cell subsets Stimulation is not clear but has been suggested to be part of a feedback mechanism during infection to restrict inflam- 250 matory responses perhaps at specific tissues sites.22 Despite a clear link between IL-12 and regulation of IL-

200 WT 10, the molecular mechanisms are ill defined. Tbet-/- There are surprising parallels between the receptor systems and signaling pathways, which regulate IL-10 150 and IFN-g expression in NK cells (as well as T cells). Our results are in line with other reports indicating that IL-2 100 and IL-12 control IL-10 (and IFN-g) expression in NK cells (Figures 1 and 2). The similarities between regula-

IL-10 (Fold Inductionl) tion of IL-10 and IFN-g are not absolute however, 50 as other established regulators of IFN-g such as IL-18, Ly-49D signaling and T-bet, have no affect on NK 0 cell-specific IL-10 expression (Figures 1 and 6c, respec- IL-2 IL-12 2+12 tively). We also found that the IL-12 homologs IL-23 or Stimulation IL-27 do not regulate IL-10 (or IFN-g) expression in NK Figure 6 Role of Stat4 and T-bet in regulating induction of cells (data not shown). This was true when IL-23 or IL-27 interleukin-10 (IL-10) expression in mouse natural killer (NK) cells. was added to NK cultures alone or in combination with (a) Cultured NK cells were generated from Stat4À/À spleens as IL-2 or IL-12, indicating a nonredundant role for IL-12 in previously described and cells were stimulated for 6 h with the controlling IL-10 expression in mouse NK cells. This is indicated cytokines and mRNA expression was measured by real- in contrast to recent reports which demonstrate an time PCR. Real-time PCR data are represented as a fold increase À/À important role for IL-27 in the induction of IL-10 over nonstimulated, Stat4 NK cells. The data are representative 43 of three separate experiments. (b) Stat4À/À NK cell culture super- expression in mouse T-cell subsets. In addition, another natants were analyzed for IL-10 production after 24 h stimulation by previous report indicated that IL-23 induces IL-10 enzyme-linked immunosorbent assay (ELISA). (c) Cultured NK expression from human cord blood derived, polyclonally cells were generated from TbetÀ/À spleens and stimulated with the activated CD4 þ and CD8 þ T cells.44 The disparity in indicated cytokines for 6 h. mRNA expression was quantified by these data may suggest differences between IL-10 real-time PCR. Real-time PCR data are represented as a fold expression in mouse NK cells and T cells, and/or point increase over nonstimulated, TbetÀ/À NK cells. The data are representative of three separate experiments. to fundamental differences between mouse and human IL-10 regulation. In that regard, our group and others have shown that the levels of histone H3 acetylation were upregulated in primary human CD56 þ NK cells as well as the human within the CNS þ 3.10 site (lanes 9 and 12). To quantify NK cell line NK92, IL-2 and IL-12 stimulation also leads levels of H3 acetylation, ChIP gel images were examined to IL-10 mRNA and protein expression (data not by denisometry (Figure 8b). Histone H3 acetylation shown).35 In addition, in both mouse and human NK

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 322 Exon CNS CNS+3.10 Il10

70% 5 4 3 2 1

Mouse TGGCTTCTGGGAACT Human CGGTTTCTGGGAAAT

IL-12-Stimulated Mutated Stat4 SP-1 α Stat5 αStat4 NS IL-2 IL-12 2+12 Stat4 Lane number12 34 5 6 7 8 9

NS IL-2 IL-12 IL-2 + IL-12

Input IgG Stat4 Input IgG Stat4 Input IgG Stat4 Input IgG Stat4

123456789101112 Figure 7 Identification of a Stat4-binding element in a conserved noncoding sequence (CNS) region. (a) Comparative sequence analysis of the mouse Il10 gene for CNS regions between human and mouse Il10. Mouse is the base genome for comparison and the location of the CNS site (CNS þ 3.10 is relative to the start of exon 1). Gray histograms indicate CNS regions and black histograms indicate exons. The conserved Stat4 element sequence is indicated below. (b) Electromobility shift assays (EMSA) analysis on the Stat4 element with nuclear extracts from cultured mouse natural killer (NK) cells. Lanes 1–4 are cultured NK cells stimulated with the indicated cytokines. Lane 5 is a cold competition with a consensus Stat4 oligo, which blocks interleukin-12 (IL-12) induction of all complexes. Lane 6 is cold competition with mutated Stat4 oligo. Lane 7 is a cold competition with an irrelevant consensus SP-1 oligo. Lane 8 is a supershift with an anti-Stat5 antibody. Lane 9 is a supershift with an anti-Stat4 antibody. The IL-12-induced complex (indicated by the arrow) is blocked by addition of the anti-Stat4 antibody. (c) Chromatin immunoprecipitation (ChIP) analysis of Stat4 binding to the Il10 intronic CNS in cultured NK cells treated with media alone (NS), IL-2, IL-12 or IL-2 þ IL12 for 2 h.

cells, IL-18 potently stimulates IFN-g but does not (1.1%) of IL-10-expressing CD56high human NK cells regulate IL-10 (data not shown).19 Since the majority of following IL-2 þ IL-12 treatment.25 When we assessed work on IL-10 has been done in the mouse, it will be IL-10 and IFN-g accumulation in NK1.1 þ NK cells, important to determine tissue- and receptor-specific we observed that approximately 70% of all IL-10 þ NK regulatory mechanisms in both mouse and human cells also express IFN-g. The size of the IL-10 þ NK cell models. Nonetheless, the parallels between IL-10 and pool is comparatively small to the numbers of IFN-g- IFN-g regulation in both species suggest some conserved producing NK cells, however (about 46%). One group has genomic regulatory structures between the Il10 and Ifng suggested that acquiring the ability to produce IL-10 in genes that may harbor tissue-specific control elements. human NK cells is associated with dying NK cells and As mentioned, IL-10 and IFN-g coexpressing T cells thus may represent a stage of the NK cell life cycle.25 We have been reported previously, and we identified a small have not directly addressed this question here, but future subset of cultured NK cells that accounts for IL-10 studies will be directed at defining the functional expression in response to IL-2 þ IL-12 stimulation characteristics of IL-10 expressing mouse NK cell subsets. (Figure 3). These small numbers are in line with another Numerous signaling cascades have been implicated in report which also described a similar population size IFN-g regulation including the p38 MAPK pathway.27,45–47

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 323 NS IL-2 IL-12 IL-2 + IL-12 through the common g-chain or Stat5 (induced by IL-15) is sufficient to regulate IL-10, alone or in combination Input IgG AcH3 Input IgG AcH3 Input IgG AcH3 Input IgG AcH3 with IL-2 or IL-12. This is in contrast to our previous finding that IL-4 synergizes with IL-2 as well as IL-12 to regulate IFN-g in NK cells.19 It will be interesting to 123456789101112 determine if IL-2 regulates IL-10 through a Stat5- dependent mechanism which is the case for IFN-g.24 1.5 We observed another key disparity between the regulation of IL-10 and IFN-g in T-bet null NK cells. Cytokine induction of IL-10 was not affected in T-betÀ/À 1 NK cells despite the fact that T-bet is required for optimal expression of IFN-g in cultured NK cells.19 Conversely, another group recently reported increased IL-10 expression in T-bet-deficient mice infected with Mycobacterium 0.5 tuberculosis.53 The authors found there was enhanced

Histone H3 Acetylation (relative ratio to input) expression of IL-10 but impaired production of IFN-g by CD4 þ T cells in M. tuberculosis-infected mice. Further- 0 more, another recent report describes a T-bet þ Foxp3À NS IL-2 IL-12 IL-2+IL-12 Th1 cell population as the primary source of IL-10 during Stimulation Toxoplasma gondii infection.41 It will be interesting to Figure 8 The intronic þ 3.10 conserved noncoding sequence (CNS) determine if there are cell-specific requirements for T-bet region is also a target for cytokine-induced histone acetylation. in regulation of IL-10 outside of the NK lineage. (a) Chromatin immunoprecipitation (ChIP) analysis of acetyl We identified a conserved Stat4 DNA-binding element histone H3 levels in cultured natural killer cells treated with media alone (NS), interleukin-2 (IL-2), IL-12 or IL-2 þ IL12 for 6 h. (b) within the fourth intron of the Il10 gene that is located Denisometric analysis of histone H3 acetylation shown above. The within a CNS (CNS þ 3.10) of both the mouse and human data are expressed as a relative ratio to each respective input Il10 genes (Figures 8a and b). In the case of IFN-g,a sample. presumption has been that Stat4 regulates acute IFN-g transcription by binding Stat4 motifs in the Ifng gene. We have reported previously that up to 80% of IL-12- Stat4-binding elements have been identified in both the induced and 60% of IL-2 þ IL-12-induced IFN-g could be mouse and human Ifng genes, however, these are largely blocked by p38 inhibitors in NK cells.19 With respect to not conserved.30 In fact, multiple mechanisms of Stat4 IL-10, we determined here that up to 72% of IL-2 þ IL-12- regulation of IFN-g have emerged which also include the induced IL-10 could be blocked by inhibitors of the p38 recruitment of other transcription factors and chromatin- pathway (Figure 7). Interestingly, both the p38 and ERK modifying proteins to the Ifng gene.34 In that regard, we MAPK pathways have been linked to recruitment of Sp1 found that the CNS þ 3.10 region, which contains the to the IL-10 promoter.48–50 However, similar to the case of Stat4 site, is also a target for cytokine-induced chromatin p38 regulation of IFN-g, these data suggest an important, remodeling based on the enhancement of histone H3 but not absolute role for IL-2 þ IL-12-induced p38 acetylation levels following cytokine stimulation signaling in regulating IL-10. (Figure 8c). It will, therefore, be interesting to determine p38 is one of several kinases upstream of the if Stat4 (and this conserved Stat4-element) plays a role in transcription factor Stat4.51 Stat4 has an established role regulating IL-10 expression in other cell populations as in regulating IFN-g in T and NK cells, most likely via a well as in human cells. Stat4 element in the proximal promoter.52 Our findings Recent reports have identified intronic regions and/or indicate that Stat4 is required for IL-12 induction of IL-10 sites 30 to Il10 that are remodeled in T cells which express and the synergy between IL-2 and IL-12 is also largely IL-10.54,55 In fact, another study indicated the presence of dependent on the presence of Stat4 (Figures 6a and b). a Th1-specific HS site in the fourth intron of the mouse IL-2 induction of IL-10 is essentially intact in Stat4À/À NK Il10 gene (HSII) which is located within or adjacent to the cells, however, suggesting there is no ‘master regulatory’ þ 3.10CNS we characterized here.32 An even more function for Stat4 in the control IL-10 transcription. comprehensive analysis of the mouse Il10 gene structure IL-2 also induces IL-10 in NK cells and the combination and expression profiles was recently reported which of IL-2 and IL-12 is highly synergistic for IL-10 and IFN-g. focused on Th1/Th2 cells.33 The authors identified The concept of synergy between cytokine signaling multiple DNase I hypersensitivity sites between the pathways is a recurring question in the literature but Il10 and Il19 genes and described three distal enhancer there has been no clear resolution. We reported pre- elements. Of note, the authors reported the presence of a viously that IL-2 and IL-12 cotreatment results in non-Th-lineage-specific DNaseI hypersensitivity site in enhanced tyrosine phosphorylation of both Stat4 and the fourth intron at position þ 2.98 (HSS þ 2.98). Stat5.19 It is likely that other converging mechanisms play Although this is presumably the same HS site reported an important role in the synergistic effects of IL-2 and IL- by Im et al. (HSII),32 there are inconsistencies as to 12 on expression of target genes such as IL-10 and IFN-g. whether this HS is Th-lineage-specific. Nonetheless, Nevertheless, we questioned if there is a conserved these reports in T cells and our data in NK cells suggest component to IL-2 cytokine induction of IL-10 via the that the fourth intron of the mouse Il10 gene contains shared common g-chain. We stimulated cultured NK essential regulatory elements including a Stat4-binding cells with other cytokines which share the common site. g-receptor (IL-4 and IL-15) and found there was no Overall, very little is known about transcriptional induction of IL-10 which suggests that neither signaling control regions and the genomic/epigenetic require-

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 324 ments for signal- and cell-specific IL-10 expression. Our analyzed on a FACS Calibur (BD Bioscience). To purify report in NK cells and recent work from other groups NK1.1 þ CD3À cells, following nylon wool separation, provides an insight into tissue-specific IL-10 expression, cells were negatively sorted by MoFlo (Cytomation, Fort cis-acting factors, and chromatin structure within the Collins, CO, USA) using FITC-conjugated anti-mouse murine Il10 gene and flanking regions.48–50,56–60 It will be CD3 or PE-conjugated NK1.1. Single positive important to resolve the genomic boundaries, epigenetic NK1.1 þ CD3À cells were washed and directly stimulated modifications and transcriptional regulators in different with cytokines. IL-10-expressing cell populations in relation to the For intracellular staining (ICS) experiments, BD Cyto- human IL-10 in an attempt to consolidate and apply fix/Cytoperm kit was used according to the manufac- our understanding of murine IL-10 regulation and tures instructions (BD Bioscience). Analysis of ICS function to humans. experiments was performed on a FACSCalibur (BD Bioscience). Data were collected with CellQuest software and analyzed with FlowJo software. Live cells were Materials and methods gated according to their forward-scatter and side-scatter Cytokines and antibodies profiles. Cell counts were made using a hemocytometer, Recombinant human IL-2 was obtained from Hoffmann- and verified by FACS, when necessary, using a bead La Roche (Nutley, NJ, USA). Recombinant mouse IL-12 standard for calibration. and IL-18 was obtained from Peprotech (Rocky Hill, NJ, USA) and MBL (Watertown, MA, USA), respectively. mRNA analyses Recombinant IL-23, IL-27 and IFN-g, and IFN-g neutra- Total RNA was isolated by guanidinium-isothescyanate lizing antibodies were purchased from R&D Systems phenol/chloroform extraction method (Trizol; Invitrogen, (Minneapolis, MN, USA). Purified hamster anti-mouse Carlsbad, CA, USA) according to the manufacturer’s CD3e (145-2C11), NK1.1, IFN-g and IL-10 mAbs were protocol. RNA concentrations were quantified and purchased from BD Bioscience (San Diego, CA, USA). subsequently normalized before cDNA synthesis. cDNA Isotype-matched Ig for fluorescence-activated cell sorting was generated using a first strand cDNA synthesis kit (FACS) analyses was also purchased from BD Bioscience. (Roche, Indianapolis, IN, USA). Quantitative PCR was For ChIP assays, anti-histone H3 antibodies and normal performed using Taqman site-specific primers and rabbit IgG were purchased from Upstate Biotechnology probes (Applied Biosystems, Foster City, CA, USA) on (Temecula, CA, USA). The same anti-Stat4 was used in an ABI PRISM 7700 sequence detector. Results were EMSA and ChIP analysis, Santa Cruz Biotechnology normalized to b-2 microglobulin levels. For relative (Santa Cruz, CA, USA). Anti-Stat5 was purchased from comparisons, NS culture conditions were assigned an R&D Systems. The mAb 4E5 or an isotype control mAb, arbitrary value of one. 0 followed by goat anti-rat (F(ab )2) was used to cross-link the Ly-49D receptor as previously described.61 Cell culture and stimulations All tissue culture (unless otherwise mentioned) media Mice contained RPMI 1640 medium supplemented with 10% C57BL/6 mice were used in all experiments except heat-inactivated FBS, 2 mML-glutamine, 1 Â essential where indicated. Mice were maintained under specific amino acids, 1 mM sodium pyruvate, 10 mM 2-mercap- pathogen-free conditions and were used between 8 and toethanol, 100 U mlÀ1 penicillin and 100 mgmlÀ1 strepto- 14 weeks of age. All animal studies in this report were mycin. Cultured NK cells were washed two times in approved by the Johns Hopkins University Institutional 2–10% fetal calf serum RPMI 1640 and plated at 5 Â 106 Animal Care and Use Committee. cells per ml in six-well plates and incubated with the indicated cytokine(s) for 2–18 h depending on the NK cell isolation experiment. Although dose response curves were com- Murine splenic NK cells were isolated as previously pleted for all of the cytokines tested, optimal doses were described.62 Briefly, single cell suspensions were pre- determined for IL-10 production and used in each pared by passing the spleens through a wire mesh experiment as follows: IL-2 (100 U mlÀ1) and IL-12 screen. RBC were lysed with ACK reagent (Biowhittiker, (10 U mlÀ1). Fresh, highly purified NK cells were plated Walkerville, MD, USA) and cells were washed and at 1 Â106 cells per ml in 48-well plates. All cells were resuspended in 5% fetal bovine serum (FBS) RPMI 1640. rested for 2–3 h at 37 1C prior to stimulation (unless The spleen suspension was passed through a sterile, otherwise noted). For MAP kinase inhibition studies, p38 prewetted nylon wool column and following a 50 min inhibitors were used at 600 nM. SB203474 was used as a incubation at 37 1C, the nylon wool-nonadherent cells negative control and was also used at a concentration of were harvested, washed and counted. The nylon wool- 600 nM (Calbiochem, La Jolla, CA, USA). The inhibitor nonadherent cells were resuspended in 10% FBS RPMI concentrations were within the range of doses previously media supplemented with high dose IL-2 (1000 U mlÀ1) reported to have inhibitory effects on MAP kinases.63 The at a density of 2 Â 106 per ml. NK cells were cultured for inhibitors were dissolved in sterile DMSO and stored at 6–8 days and were typically between 75 and 90% DX5 þ À20 1C at a concentration of 20 mM. or NK1.1 þ before harvesting for experiments. Electrophoretic mobility shift assay Flow cytometry and NK cell purification Complementary single-stranded oligonucleotides were FACS was performed by staining cells with phycoery- commercially synthesized (Operon Biotechnologies Inc., thrin (PE)-conjugated NK1.1 and fluorescein isothyocya- Huntsville, AL, USA) to span approximately 10 bp on nate (FITC)-conjugated anti-mouse CD3 to determine NK either side of the putative Stat4 element: 50-AGGCCA cell purity in cultured NK cell cultures. Results were CATGGCTTCTGGGAACTAGGGTTG-30. The consensus

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 325 Stat4 oligo used in cold-competition assays was as precleared with salmon sperm DNA/protein A agarose follows: 50-GAGCCTGATTTCCCCGAAATGATGAGC for 30 min prior to immunoprecipitation with specified TAG-30. The mutant Stat4 oligo was identical to the antibodies. A proportion (2%) of the diluted supernatant consensus oligo except for a ‘CCC’ substitution of ‘TTT’ was kept as ‘input’ to quantify genomic DNA. After shown in bold. Briefly, complementary strands were immunoprecipitation, the protein–DNA complexes were annealed by combining 2 mg of each oligonucleotide and incubated with Protein A beads, and the protein–DNA

6 mlof10Â annealing buffer (500 mM Tris, 100 mM MgCl2 complexes were eluted in 1% SDS/0.1 M NaHCO3, and and 50 mM dithiothreitol) in a 60 ml reaction, denatured in cross-links were reversed at 65 1C. DNA was recovered a boiling water bath for 5 min and then allowed to cool to by phenol-chloroform extraction and ethanol precipita- room temperature.64 Double-stranded DNA overhangs tion, and then subjected to PCR and/or real-time PCR were labeled with Klenow enzyme by a fill-in reaction, analysis. PCR was carried out with ampliTaq Gold [a-32P]dCTP and free dNTPs (Amersham Biosciences (Applied Biosystems) for 35 cycles (45 s at 95 1C, 60 s at Corp., Piscataway, NJ, USA). The DNA-protein binding 59 1C, and 60 s at 72 1C), and the products were reaction was conducted in a 20 ml reaction mixture visualized by ethidium bromide staining. The immuno- consisting of 5–10 mg of nuclear protein extract, 1 mg precipitated DNA samples were normalized to ‘input’ poly(dI-dC) (Sigma, St Louis, MO, USA), 4 mlof5Â DNA samples by densitometry. PCR primers for binding buffer (60 mM Hepes, 7.5 mM MgCl2, 300 mM KCl, the fourth intron þ 3.01CNS site in the murine Il10 gene; 0 0 0 1mM ethylenediamine-tetraacetic acid, 2.5 mM dithio- 5 -CTCAGGTACATCATTG-3 and 5 -GAGTGTGTAGG threitol, 50% glycerol and 4-(2-aminoethyl)-benzenesul- CAGTC-30. fonyl flouride hydorochloride) and 1.5 Â 104 c.p.m. of 32P- labeled oligonucleotide probe. For supershift analysis to Cytokine secretion assays and statistical analysis identify Stat proteins interacting with the Il10 promoter, Cell-free supernatants were collected and assayed for m 1 l of antibody was added to the reaction prior to the IL-10 and IFN-g production by ELISA (R&D Systems). addition of the probes and incubated on ice for 20 min. In À The sensitivity limits for the assays were o2pgml 1. cold oligonucleotide competition experiments, 10- to 100- Comparative data were analyzed using the unpaired fold excess of unlabeled oligonucleotide probe was Student’s t-test. The software used to perform the added 10 min prior to adding the radiolabeled oligonucleo- statistical analysis was SigmaPlot 2000 for Windows tide probe. version 6.10. Nuclear extraction Nuclear extracts were prepared from cytokine-stimulated cultured NK cells that had been rested for 4 h in the Acknowledgements absence of IL-2 as described previously.65 Briefly, following NS or treatment with cytokines for 20–60 min, cell We thank Dr John J O’Shea for his valuable support of pellets were resuspended in lysis buffer (50 mM KCl, this work. In addition, we thank the Johns Hopkins 25 mM Hepes, pH 7.8, 0.5% Nonidet P-40, 1 mM phenyl- Becton Dickinson Immune Function Laboratory and Paul methylsulfonyl fluoride, 10 mgmlÀ1 leupeptin, 20 mgmlÀ1 Fallon for assistance with flow cytometry-related experi- aprotinin, 100 mM dithiothreitol) and subsequently incu- ments. We also thank Debbie Hodge and Howard Young for helpful discussions and critical review of this manu- bated on ice for 5 min. Cellular suspensions were À/À À/À collected by centrifugation at 2000 r.p.m. and the super- script and for providing Tbet and Ifng mice. This natant was harvested as the cytoplasmic protein fraction. work was supported by National Institutes of Health Nuclei were washed in wash buffer (lysis buffer without Grant AI070594 (to JHB). Nonidet P-40) and harvested by centrifugation at 2000 r.p.m. Nuclear pellets were resuspended in extraction buffer (500 mM KCl, 25 mM Hepes, pH 7.8, 5% References glycerol, 1 mM phenylmethylsulfonyl fluoride, 10 mgmlÀ1 leupeptin, 20 mgmlÀ1 aprotinin and 100 mM dithiothrei- 1 McCulloch CA, Downey GP, El-Gabalawy H. Signalling tol), frozen in dry ice, thawed slowly on ice and finally platforms that modulate the inflammatory response: new centrifuged at 14 000 r.p.m. for 10 min. The supernatant targets for drug development. Nat Rev Drug Discov 2006; 5: was harvested and nuclear proteins quantified with 864. the bicinchoninic acid protein assay reagent (Pierce, 2 Mocellin S, Marincola F, Rossi CR, Nitti D, Lise M. The multifaceted relationship between IL-10 and adaptive immu- Rockford, IL, USA). nity: putting together the pieces of a puzzle. Cytokine Growth Factor Rev 2004; 15: 61. Chromatin immunoprecipitation assays 3 Moore KW, O’Garra A, De Waal MR, Vieira P, Mosmann TR. ChIP assays were performed according to the manufac- Interleukin-10. Annu Rev Immunol 1993; 11: 165. turer’s instructions (Upstate Biotechnology). Briefly, 4 Mege JL, Meghari S, Honstettre A, Capo C, Raoult D. The two 2 Â 106 cells were fixed with 1% formaldehyde, washed faces of interleukin 10 in human infectious diseases. Lancet with cold phosphate-buffered saline, and lysed in buffer Infect Dis 2006; 6: 557. containing 10 mgmlÀ1 aprotinin (ICN Biomedicals, 5 Pestka S, Krause CD, Sarkar D, Walter MR, Shi Y, Fisher PB. Aurora, OH, USA), 10 mgmlÀ1 leupeptin (Bachem, Interleukin-10 and related cytokines and receptors. Annu Rev Immunol 2004; 22: 929. mM Torrance, CA, USA) and 2.5 NPGB (4-nitrophenyl 4- 6 D’Andrea A, Aste-Amezaga M, Valiante NM, Ma X, Kubin M, guanidinobenzoate hydrochloride, Sigma). Nuclei were Trinchieri G. Interleukin 10 (IL-10) inhibits human lymphocyte sonicated for a total of 60 s to shear DNA (Heat Systems, interferon gamma-production by suppressing natural killer Farmingdale, NY, USA), the lysates were pelleted, cell stimulatory factor/IL-12 synthesis in accessory cells. J Exp and supernatants were diluted. Diluted lysates were Med 1993; 178: 1041.

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 326 7 Fiorentino DF, Zlotnik A, Mosmann TR, Howard M, O’Garra mRNA stability in the synergistic interaction between NKSF A. IL-10 inhibits cytokine production by activated macro- and IL-2. J Immunol 1992; 148: 92. phages. J Immunol 1991; 147: 3815. 27 Rincon M, Enslen H, Raingeaud J, Recht M, Zapton T, Su MS 8 Mocellin S, Panelli MC, Wang E, Nagorsen D, Marincola FM. et al. Interferon-gamma expression by Th1 effector T cells The dual role of IL-10. Trends Immunol 2003; 24: 36. mediated by the p38 MAP kinase signaling pathway. EMBO J 9 Moore KW, De Waal MR, Coffman RL, O’Garra A. Interleukin- 1998; 17: 2817. 10 and the interleukin-10 receptor. Annu Rev Immunol 2001; 28 Morinobu A, Gadina M, Strober W, Visconti R, Fornace A, 19: 683. Montagna C et al. STAT4 serine phosphorylation is critical for 10 Mocellin S, Panelli M, Wang E, Rossi CR, Pilati P, Nitti D et al. IL-12-induced IFN-gamma production but not for cell IL-10 stimulatory effects on human NK cells explored by gene proliferation. Proc Natl Acad Sci USA 2002; 99: 12281. profile analysis. Genes Immun 2004; 5: 621. 29 Kaplan MH, Sun YL, Hoey T, Grusby MJ. Impaired IL-12 11 Boulland ML, Meignin V, Leroy-Viard K, Copie-Bergman C, responses and enhanced development of Th2 cells in Stat4- Briere J, Touitou R et al. Human interleukin-10 expression in deficient mice. Nature 1996; 382: 174. T/natural killer-cell lymphomas: association with anaplastic 30 Watford WT, Hissong BD, Bream JH, Kanno Y, Muul L, O’Shea large cell lymphomas and nasal natural killer-cell lymphomas. JJ. Signaling by IL-12 and IL-23 and the immunoregulatory Am J Pathol 1998; 153: 1229. roles of STAT4. Immunol Rev 2004; 202: 139. 12 Shibata Y, Foster LA, Kurimoto M, Okamura H, Nakamura 31 Townsend MJ, Weinmann AS, Matsuda JL, Salomon R, RM, Kawajiri K et al. Immunoregulatory roles of IL-10 in Farnham PJ, Biron CA et al. T-bet regulates the terminal innate immunity: IL-10 inhibits macrophage production of maturation and homeostasis of NK and Valpha14i NKT cells. IFN-gamma-inducing factors but enhances NK cell produc- Immunity 2004; 20: 477. tion of IFN-gamma. J Immunol 1998; 161: 4283. 32 Im SH, Hueber A, Monticelli S, Kang KH, Rao A. Chromatin- 13 Cai G, Kastelein RA, Hunter CA. IL-10 enhances NK cell level regulation of the IL10 gene in T cells. J Biol Chem 2004; proliferation, cytotoxicity and production of IFN-gamma 279: 46818. when combined with IL-18. Eur J Immunol 1999; 29: 2658. 33 Jones EA, Flavell RA. Distal enhancer elements transcribe 14 Chakir H, Lemay AM, Webb JR. Cytokine expression by intergenic RNA in the IL-10 family gene cluster. J Immunol murine DX5+ cells in response to IL-12, IL-18, or the 2005; 175: 7437. combination of IL-12 and IL-18. Cell Immunol 2001; 212: 71. 34 Young HA, Bream JH. IFN-gamma: recent advances in 15 Wakkach A, Fournier N, Brun V, Breittmayer JP, Cottrez F, understanding regulation of expression, biological functions, Groux H. Characterization of dendritic cells that induce and clinical applications. Curr Top Microbiol Immunol 2007; tolerance and T regulatory 1 cell differentiation in vivo. 316: 97. Immunity 2003; 18: 605. 35 Mehrotra PT, Donnelly RP, Wong S, Kanegane H, Geremew A, 16 Biron CA, Byron KS, Sullivan JL. Severe herpesvirus infections Mostowski HS et al. Production of IL-10 by human natural in an adolescent without natural killer cells. N Engl J Med 1989; killer cells stimulated with IL-2 and/or IL-12. J Immunol 1998; 320: 1731. 160: 2637. 17 Arase H, Mocarski ES, Campbell AE, Hill AB, Lanier LL. 36 Jeannin P, Delneste Y, Seveso M, Life P, Bonnefoy JY. IL-12 Direct recognition of cytomegalovirus by activating and synergizes with IL-2 and other stimuli in inducing IL-10 inhibitory NK cell receptors. Science 2002; 296: 1323. production by human T cells. J Immunol 1996; 156: 3159. 18 Cousens LP, Orange JS, Biron CA. Endogenous IL-2 con- 37 Gerosa F, Paganin C, Peritt D, Paiola F, Scupoli MT, tributes to T cell expansion and IFN-gamma production ste-Amezaga M et al. Interleukin-12 primes human CD4 and during lymphocytic choriomeningitis virus infection. JIm- CD8T cell clones for high production of both interferon- munol 1995; 155: 5690. gamma and interleukin-10. J Exp Med 1996; 183: 2559. 19 Bream JH, Curiel RE, Yu CR, Egwuagu CE, Grusby MJ, Aune 38 Windhagen A, Anderson DE, Carrizosa A, Williams RE, TM et al. IL-4 synergistically enhances both IL-2- and IL-12- Hafler DA. IL-12 induces human T cells secreting IL-10 with induced IFN-gamma expression in murine NK cells. Blood IFN-gamma. J Immunol 1996; 157: 1127. 2003; 102: 207. 39 Pohl-Koppe A, Balashov KE, Steere AC, Logigian EL, Hafler 20 Wolk K, Kunz S, Asadullah K, Sabat R. Cutting edge: immune DA. Identification of a T cell subset capable of both IFN- cells as sources and targets of the IL-10 family members? gamma and IL-10 secretion in patients with chronic Borrelia J Immunol 2002; 168: 5397. burgdorferi infection. J Immunol 1998; 160: 1804. 21 Portielje JE, Lamers CH, Kruit WH, Sparreboom A, Bolhuis 40 Gerosa F, Nisii C, Righetti S, Micciolo R, Marchesini M, RL, Stoter G et al. Repeated administrations of interleukin (IL)- Cazzadori A et al. CD4(+) T cell clones producing both 12 are associated with persistently elevated plasma levels of interferon-gamma and interleukin-10 predominate in bronch- IL-10 and declining IFN-gamma, tumor necrosis factor-alpha, oalveolar lavages of active pulmonary tuberculosis patients. IL-6, and IL-8 responses. Clin Cancer Res 2003; 9: 76. Clin Immunol 1999; 92: 224. 22 Trinchieri G. Regulatory role of T cells producing both 41 Jankovic D, Kullberg MC, Feng CG, Goldszmid RS, Collazo interferon gamma and interleukin 10 in persistent infection. CM, Wilson M et al. Conventional T-bet(+)Foxp3(À) Th1 cells J Exp Med 2001; 194: F53–F57. are the major source of host-protective regulatory IL-10 during 23 Chan SH, Perussia B, Gupta JW, Kobayashi M, Pospisil M, intracellular protozoan infection. J Exp Med 2007; 204: 273. Young HA et al. Induction of interferon gamma production by 42 Anderson CF, Oukka M, Kuchroo VJ, Sacks D. natural killer cell stimulatory factor: characterization of the CD4(+)CD25(À)Foxp3(À) Th1 cells are the source of IL-10- responder cells and synergy with other inducers. J Exp Med mediated immune suppression in chronic cutaneous leishma- 1991; 173: 869. niasis. J Exp Med 2007; 204: 285. 24 Bream JH, Hodge DL, Gonsky R, Spolski R, Leonard WJ, 43 Stumhofer JS, Silver JS, Laurence A, Porrett PM, Harris TH, Krebs S et al. A distal region in the interferon-gamma gene is a Turka LA et al. Interleukins 27 and 6 induce STAT3- site of epigenetic remodeling and transcriptional regulation by mediated T cell production of interleukin 10. Nat Immunol interleukin-2. J Biol Chem 2004; 279: 41249. 2007; 12: 1363. 25 Loza MJ, Perussia B. The IL-12 signature: NK cell terminal 44 Vanden ES, Goriely S, De WD, Willems F, Goldman M. CD56+high stage and effector functions. J Immunol 2004; IL-23 up-regulates IL-10 and induces IL-17 synthesis by 172: 88. polyclonally activated naive T cells in human. Eur J Immunol 26 Chan SH, Kobayashi M, Santoli D, Perussia B, Trinchieri G. 2005; 35: 469. Mechanisms of IFN-gamma induction by natural killer cell 45 Dong C, Davis RJ, Flavell RA. MAP kinases in the immune stimulatory factor (NKSF/IL-12). Role of transcription and response. Annu Rev Immunol 2002; 20: 55.

Genes and Immunity Regulation of NK cell-specific IL-10 expression LR Grant et al 327 46 Murphy KM, Reiner SL. The lineage decisions of helper T 56 Zhang X, Edwards JP, Mosser DM. Dynamic and Transient cells. Nat Rev Immunol 2002; 2: 933. Remodeling of the Macrophage IL-10 Promoter during 47 Agnello D, Lankford CS, Bream J, Morinobu A, Gadina M, Transcription. J Immunol 2006; 177: 1282. O’Shea JJ et al. Cytokines and transcription factors that 57 Brenner S, Prosch S, Schenke-Layland K, Riese U, Gausmann regulate T helper cell differentiation: new players and new U, Platzer C. cAMP-induced Interleukin-10 promoter insights. J Clin Immunol 2003; 23: 147. activation depends on CCAAT/enhancer-binding protein 48 Ma W, Lim W, Gee K, Aucoin S, Nandan D, Kozlowski M et al. expression and monocytic differentiation. J Biol Chem 2003; The p38 mitogen-activated kinase pathway regulates the 278: 5597. human interleukin-10 promoter via the activation of Sp1 58 Lin SC. Identification of an NF-Y/HMG-I(Y)-binding site in transcription factor in lipopolysaccharide-stimulated human the human IL-10 promoter. Mol Immunol 2006; 43: 1325. macrophages. J Biol Chem 2001; 276: 13664. 59 Cao S, Liu J, Song L, Ma X. The protooncogene c-Maf is an 49 Lucas M, Zhang X, Prasanna V, Mosser DM. ERK activation essential transcription factor for IL-10 gene expression in following macrophage FcgammaR ligation leads to chromatin macrophages. J Immunol 2005; 174: 3484. modifications at the IL-10 locus. J Immunol 2005; 175: 469. 60 Shoemaker J, Saraiva M, O’Garra A. GATA-3 directly 50 Vega MI, Huerta-Yepaz S, Garban H, Jazirehi A, Emmanoui- remodels the IL-10 locus independently of IL-4 in CD4+ lides C, Bonavida B. Rituximab inhibits p38 MAPK activity in T cells. J Immunol 2006; 176: 3470. 2F7 B NHL and decreases IL-10 transcription: pivotal role of 61 Mason LH, Willette-Brown J, Mason AT, McVicar D, Ortaldo p38 MAPK in drug resistance. Oncogene 2004; 23: 3530. JR. Interaction of Ly-49D+ NK cells with H-2Dd target cells 51 Zhang S, Kaplan MH. The p38 mitogen-activated protein leads to Dap-12 phosphorylation and IFN-gamma secretion. kinase is required for IL-12-induced IFN-gamma expression. J Immunol 2000; 164: 603. J Immunol 2000; 165: 1374. 62 Mason L, Giardina SL, Hecht T, Ortaldo J, Mathieson BJ. 52 Nguyen KB, Watford WT, Salomon R, Hofmann SR, Pien GC, LGL-1: a non-polymorphic antigen expressed on a major Morinobu A et al. Critical role for STAT4 activation by type 1 population of mouse natural killer cells. J Immunol 1988; interferons in the interferon-gamma response to viral infec- 140: 4403. tion. Science 2002; 297: 2063. 63 Kalina U, Kauschat D, Koyama N, Nuernberger H, Ballas K, 53 Sullivan BM, Jobe O, Lazarevic V, Vasquez K, Bronson R, Koschmieder S et al. IL-18 activates STAT3 in the natural killer Glimcher LH et al. Increased susceptibility of mice lacking T- cell line 92, augments cytotoxic activity, and mediates bet to infection with Mycobacterium tuberculosis correlates with IFN-gamma production by the stress kinase p38 and by increased IL-10 and decreased IFN-gamma production. the extracellular regulated kinases p44erk-1 and p42erk-21. J Immunol 2005; 175: 4593. J Immunol 2000; 165: 1307. 54 Wang ZY, Sato H, Kusam S, Sehra S, Toney LM, Dent AL. 64 Bream JH, Young HA, Rice N, Martin MP, Smith MW, Regulation of IL-10 gene expression in Th2 cells by Jun Carrington M et al. CCR5 Promoter Alleles and Specific proteins. J Immunol 2005; 174: 2098. DNA Binding Factors. Science 1999; 284: 223a. 55 Saraiva M, Christensen JR, Tsytsykova AV, Goldfeld AE, 65 Bream JH, Ping A, Zhang X, Winkler C, Young HA. A single Ley SC, Kioussis D et al. Identification of a macrophage- nucleotide polymorphism in the proximal IFN-gamma specific chromatin signature in the IL-10 locus. J Immunol 2005; promoter alters control of gene transcription. Genes Immun 175: 1041. 2002; 3: 165.

Genes and Immunity