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

Abstract

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-γ (IFN-γ) regulation in mouse natural killer (NK) cells. Like IFN-γ, IL-10 expression is induced by IL-2 and IL-12 and IL-2 + IL-12 stimulation is synergistic. Unlike IFN-γ, 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 + 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-γ, 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.

Introduction

Maintaining a proper balance between inflammatory and anti-inflammatory responses is essential for effective immunity against infectious pathogens while repressing the onset of autoimmunity.¹ The obligatory reaction of organisms to foreign pathogens is to launch an inflammatory response that is initiated by cells of the innate immune system, such as professional antigen-presenting cells (APC) and natural killer (NK) cells. The inflammatory process is driven largely by the secretion of proinflammatory cytokines, including macrophage- and dendritic cell (DC)-derived interleukin-12 (IL-12), as well as NK cell-derived interferon-γ (IFN-γ). NK cells prime macrophages and DC during primary responses by providing early sources of IFN-γ and potentially other cytokines. There is considerable cellular crosstalk between these cell populations, which in turn influences the activation status of each population and provides the microenvironmental conditions necessary to employ adaptive immune measures, such as the differentiation of T helper cell populations.² Without appropriate feedback mechanisms, however, the inflammatory process results in severe pathology including tissue destruction and death. The immunoregulatory cytokines IL-10 and IFN-γ play fundamental roles in this complex process. A key to understanding this delicate balance is by evaluating the molecular and biochemical mechanisms that govern the expression of these genes. Remarkably, these processes remain poorly understood.

IL-10 is expressed by a variety of cell types including macrophages, DC subsets, B cells, several T-cell subpopulations and T regulatory cells.³ Primarily, IL-10 functions as an anti-inflammatory molecule that inhibits macrophage/DC activation and maturation^4,5 and antagonizes the proinflammatory cytokines IL-1β, IL-6, IL-8, TNF-α and notably IL-12.^6,7 Paradoxically, the biological role of IL-10 is not limited to inactivation of APCs; IL-10 is also known to enhance B cell, granulocyte, mast cell and keratinocyte growth/differentiation, as well as NK cell and CD8⁺ cytotoxic T-cell activation.^7-11 In addition, IL-10 has been shown to enhance NK cell proliferation and/or production of IFN-γ.^12-14

Various cell types can be divided into distinct subpopulations based on IL-10 expression patterns. In the myeloid compartment, alternatively activated macrophages, type II macrophages and some DC subsets are the principle sources of IL-10 production.¹⁵ Another potentially important, yet underappreciated, source of IL-10 is derived from the NK cell compartment. NK cells and NKT cells are integral parts of the innate immune system, and are critical to host defense against infectious pathogens, such as herpes viruses^16,17 and choriomeningitis virus.¹⁸ It is well known that NK cells provide early sources of IFN-γ in vivo. Our group and others have shown that IL-2, IL-12 and their combination are potent regulators of IFN-γ expression in NK cells.¹⁹ Surprisingly, IL-2 and IL-12 also induce IL-10 expression in both mouse and human NK cells and the combination of IL-2/IL-12 is synergistic.^14,20

Interestingly, clinical administration of recombinant human IL-12 to cancer patients resulted in elevated plasma levels of IL-10 and decreased IFN-γ.²¹ The biological significance of IL-12 induction of IL-10 is not clear but has been suggested to be part of an important feedback mechanism during infection to likely restrict inflammatory responses at specific tissues sites.²² Therefore, there is a link between IL-12 and regulation of IL-10, particularly in NK cells, however, the molecular mechanisms have not been explored.

Here we report unexpected similarities and some key disparities between regulation of IL-10 and IFN-γ expression in NK cells. Like IFN-γ, IL-2 and IL-12 synergistically induce IL-10 expression from mouse and human NK cells which is largely dependent on Stat4. However, other known regulators of IFN-γ (IL-18, Ly-49D and the transcription factor T-bet), do not control IL-10 expression in NK cells. In searching for molecular targets for Stat4, we identified a conserved Stat4-like element within a conserved noncoding sequence (CNS) in the fourth intron of the Il10 gene. This motif binds IL-12-induced Stat4 and the CNS site is a target for cytokine-induced histone acetylation. Thus, we provide molecular insight into a surprising parallel between IL-10 and IFN-γ gene regulatory mechanisms.

Results

Similarities and disparities in IL-10 and IFN-γ induction in cultured NK cells

In previous studies we established spleen-derived, IL-2-cultured mouse NK cells (cultured NK cells) as a representative model of IFN-γ regulation as compared to freshly isolated NK cells.¹⁹ We have identified notable similarities and stark differences between the induction signals for IL-10 expression and IFN-γ. Utilizing cultured NK cells as a model, we found that IL-10 is regulated by some of the same receptors, which control IFN-γ expression. As shown in Figure 1, IL-2 and IL-12 treatment independently induce IL-10 mRNA and protein production. In addition, as we and others have reported for IFN-γ,^19,23,24 IL-2 + IL-12 costimulation of cultured NK cells resulted in a highly synergistic induction of IL-10 mRNA (Figure 1a) and IL-10 secretion (Figure 1b). The similarities in regulatory signals for IFN-γ and IL-10 are not absolute, however, as other IFN-γ-inducing agents (IL-18 and Ly-49D cross-linking) have no affect on IL-10 expression. Similarly, the IL-12 homologs IL-23 and IL-27 did not induce IL-10 individually or in combination with one another, IL-2 or IL-12 (data not shown). Induction of IFN-γ was confirmed for all experiments at the mRNA level by real-time PCR and by enzyme-linked immunosorbent assay (ELISA) in culture supernatants (data not shown). These data suggest common regulatory pathways for IL-10 and IFN-γ in NK cells and may also point to shared, tissue-specific epigenetic/genomic control mechanisms.

Figure 1.

Induction of interleukin-10 (IL-10) expression from cultured natural killer (NK) cells in response to cytokine stimulation. (a) Cultured NK cells were stimulated for 6 h with the indicated cytokines and mRNA expression was measured by real-time PCR. Real-time PCR data are represented as a fold increase 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 data are shown as the mean ± s.d. from three independent experiments.

Freshly isolated splenic NK cells express IL-10 in response to IL-2, IL-12 and IL-2 + IL-12 costimulation

To confirm that the IL-10 expression patterns observed in cultured NK cells were also present in primary NK cells, we isolated NK1.1⁺CD3⁻ cells from nylon wool-passed spleens by negative flow cytometric sorting. The resulting populations were greater than 95% NK1.1⁺CD3⁻ pure. As we found in cultured NK cells, stimulation of freshly isolated NK cells with IL-2, IL-12 or IL-2 + IL-12 resulted in induction of IL-10 mRNA (Figure 2a) and protein (Figure 2b). Likewise, IL-18 treatment did not induce the expression of IL-10 in fresh NK cells. Our group and others have reported that the response of IFN-γ expression is attenuated in fresh NK cells as compared to highly activated, cultured NK cells. Similarly, we found that while the trends of IL-10 induction between cultured NK cells and fresh NK cells are comparable, the levels of IL-10 are lower and the kinetics are delayed. Thus, both primary and cultured NK cells express IL-10 in response to IL-2 and IL-12 and the combination of IL-2 + IL-12 exposure is synergistic.

Figure 2.

Induction of interleukin-10 (IL-10) expression from freshly isolated natural killer (NK) cells in response to cytokine stimulation. (a) Primary NK cells (NK1.1⁺/CD3⁻) were stimulated for 24 h with the indicated cytokines and mRNA expression was measured by real-time PCR. Real-time PCR data are represented as a fold increase over nonstimulated, primary NK cells. (b) Primary NK cell culture supernatants were analyzed for IL-10 production after 48 h stimulation by enzyme-linked immunosorbent assay (ELISA). The data are representative of two separate experiments.

IL-10 is expressed by a small subset of cultured NK cells

Having established that NK cells are a source of IL-2/IL-12 induced IL-10, we wanted to determine the cellular relationship between IL-10-expressing and IFN-γ-expressing NK cell populations. We used intracellular cytokine staining (ICS) of nonstimulated (NS) and IL-2 + IL-12 treated cultured NK cells to determine cytokine expression patterns (Figure 3). As shown in Figures 3a and b, IL-2 + IL-12 stimulation induces a small population of NK1.1⁺CD3⁻ NK cells to express IL-10 (approximately 0.8% after subtraction of IL-10⁺ cells in the NS control). No IL-10-expressing NK1.1⁺CD3⁺ NKT cells were detected (data not shown). Interestingly, Loza and Perussia²⁵ recently described a similar small percentage of IL-10-expressing, human CD56⁺ NK cells in response to IL-2 and IL-12 stimulation.

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) + IL-12-treated NK cells were stained with antibodies directed against NK1.1, CD3, IL-10, and IFN-γ 6 h later. (c) NS and (d) IL-2 + IL-12-treated NK cells were also stained with the same antibody cocktail 6 h later. Cells are gated on NK1.1⁺ in Figures 4c and d. The data are representative of three separate experiments.

As expected, when NK cells were also stained for IFN-γ by ICS, we found a large induction of IFN-γ-expressing cells after IL-2 + IL-12 treatment within the NK1.1⁺ fraction (Figures 3c and d). Interestingly, we found that within the population of IL-10-expressing NK1.1⁺ cells, about 70% coexpress IFN-γ. Albeit it a small fraction, the majority of NK cells that produce IL-10 in response to IL-2 + IL-12 stimulation also coexpress IFN-γ.

Kinetics of IL-10 mRNA expression in cultured NK cells is bi-phasic and IL-10 mRNA half-life is short in NK cells

We know from previous studies in NK cells that induction of IFN-γ mRNA by IL-2 + IL-12 treatment is rapid and peaks around 3−4 h post stimulation.²³ Given the similarities between IFN-γ and IL-10 induction in NK cells we examined the kinetics of IL-10 mRNA expression. Like IFN-γ, IL-10 mRNA expression is induced maximally by 3 h and is sustained at 6−8 h following cytokine stimulation (Figure 4a). However, IL-10 transcripts are greatly diminished 12 h following stimulation.

Figure 4.

Interleukin-10 (IL-10) mRNA kinetics and half-life in cultured natural killer (NK) cells. (a) Cultured NK cells were stimulated IL-2, IL-12 or IL-2 + IL-12 for the indicated time points and mRNA expression was measured by real-time PCR. IL-10 mRNA expression was measured by real-time PCR which is shown as a fold increase over the nonstimulated NK cell control. The data are representative of two separate experiments. (b) Cultured NK cells were treated with cytokines for 3 h and harvested for control mRNA levels (control; 0 h) or actinomycin D (Act D) at 5 μg ml⁻¹, was added 3 h after cytokine stimulation. Act D-treated cells were harvested 2 h after treatment (Act D 2 h), and 4 h after Act D treatment (Act D 4 h). Data are represented as a percent of control IL-10 mRNA expression. Dashed line indicates 50% of control.

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 of IFN-γ are in the culture supernatants by 12 h, we considered that IFN-γ 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 exclude this possibility, we either pretreated cultured NK cells for 4 h with a neutralizing IFN-γ antibody or with rmIFN-γ prior to IL-2/IL-12 stimulations and found no significant difference in IL-10 mRNA or protein levels at 6 and 24 h, respectively (data not shown). We also stimulated cultured NK cells derived from Ifng^−/− mice and found no enhancement of IL-10 expression (data not shown). While we cannot exclude the possibility that other factors expressed by NK cells may contribute to IL-10 mRNA kinetic profiles at later time-points, it appears that IFN-γ does not play a role in IL-10 responses to IL-2/IL-12.

To determine the stability of IL-10 mRNA transcripts in NK cells, cytokine-stimulated NK cells were treated with the transcriptional inhibitor actinomycin D (Act D) and IL-10 mRNA levels were assessed at 2 and 4 h post Act D treatment (Figure 4b). As mentioned above, we determined that maximal IL-10 expression was achieved by 3 h post stimulation and generally maintained for 6−8 h. Thus, we choose the 3 h time point (control; 0 h) as a stable point of comparison for Act-treated cells. For transcriptional blockade, NK cells were stimulated with cytokines for 3 h, then Act D was added to cultures and mRNA was harvested 2 or 4 h later. Results are shown as a percent of control.

Similar to previous reports of IFN-γ stability in NK cells, IL-10 mRNA half-life is relatively short with over a 75% reduction in IL-10 mRNA levels in all cytokine stimulation conditions after 2 h of Act D treatment.²⁶ Nevertheless, there is still detectible IL-10 mRNA up to 4 h following blockade of transcription. These data indicate that IL-2/IL-12 induced IL-10 mRNA in NK cells is relatively unstable, regardless of the stimulation conditions but may be sustained to a small degree for at least 4 h.

Partial involvement of the p38 MAP kinase pathway in IL-10 induction

In addition to utilizing JAK/STAT signaling mechanisms, IL-2 and IL-12 also activate the mitogen-activated protein kinase (MAPK) pathway. Our group and others have reported a role in particular for the MAPK p38 in regulating IL-12 as well as IL-2 + IL-12 induced IFN-γ expression.^19,27,28 Therefore, we examined the effects of p38-specific inhibitors on IL-10 induction by IL-2 and IL-12 (Figure 5). The synergistic effect of IL-2 + IL-12 stimulation on IL-10 production was significantly reduced by both p38 inhibitors as compared to the dimethyl sulfoxide (DMSO) controls (SB202190, P<0.01 and SB203580, P<0.05). SB202190 consistently yielded a greater degree of inhibition as compared to another p38 inhibitor SB203580 (on average, 31% of control for SB202190 compared to 44% of control for SB203580). These findings suggest that the p38 pathway has a role in IL-2 + IL-12 regulation of IL-10 expression. However, the fact that inhibition of the p38 MAPK pathway resulted in only partial blockage of IL-10 expression suggests multiple regulatory mechanisms may account for the synergistic effects of IL-2 and IL-12 on IL-10 expression.

Figure 5.

Effect of p38 MAPK inhibitors on cytokine-induced interleukin-10 (IL-10) production from cultured natural killer (NK) cells. Cultured NK cells were pretreated with p38 inhibitors (sb202190 or sb203580), vehicle control (dimethyl sulfoxide; DMSO) or media (control) for 30 min prior to stimulation with cytokines. NK cells were then stimulated overnight and NK cell culture supernatants were analyzed for IL-10 production by enzyme-linked immunosorbent assay (ELISA). The IL-10 production data are shown as the mean ± s.d. from three independent experiments. Statistical significance is indicated by **(P<0.01) or *(P<0.05).

IL-2 + IL-12 regulation of IL-10 is mediated largely by Stat4 but not T-bet

IL-12 regulates IFN-γ expression in Stat4-dependent manner.^29,30 Given the similarities in IL-10 and IFN-γ regulation in NK cells, we hypothesized that Stat4 would be involved in IL-12-induction of IL-10, including the synergy induced by IL-2 + IL-12 stimulation. To test this hypothesis, we generated spleen-derived, cultured NK cells from Balb/c Stat4^−/− mice and stimulated them with cytokines as performed previously. As expected, IL-2 induction of IL-10 is essentially intact in the absence of Stat4, however, IL-12 fails to induce IL-10 mRNA expression in Stat4^−/− NK cells (Figure 6a). In addition, the synergistic activity of IL-2 + IL-12 on IL-10 was dramatically reduced in the absence of Stat4, yet was not completely ablated. In line with IL-10 mRNA expression, IL-12 and IL-2 + IL-12-induced IL-10 production was substantially reduced in Stat4-deficient NK cells (Figure 6b). As observed in wild-type cells, we found no induction of IL-10 by IL-18 in Stat4 null NK cells. Induction of IFN-γ by IL-2 and IL-18 was intact in Stat4 deficient NK cells (data not shown). These data suggest that Stat4 is necessary for IL-12-induction of IL-10, and largely, but not entirely responsible for the functional synergy between IL-2 and IL-12 in regulating IL-10 expression. Since Stat4 does not completely explain IL-2 + IL-12 induction of IL-10 and the T-box transcription factor T-bet is required for optimal expression of IFN-γ in NK cells, we asked if T-bet is involved in regulation of IL-10 expression.^19,31 Cultured NK cells were generated from T-bet-deficient mice and tested for induction of IL-10. There was no effect of the absence of T-bet on induction of IL-10 by IL-2 or IL-12 alone or the synergistic effect of IL-2 + IL-12 (Figure 6c). Together these data point to an important role of Stat4 in regulating both the Il10 and Ifng genes, yet also indicate a key disparity given that T-bet is required for optimal IFN-γ expression but dispensable for transcriptional regulation of IL-10 in NK cells.

Figure 6.

Role of Stat4 and T-bet in regulating induction of interleukin-10 (IL-10) expression in mouse natural killer (NK) cells. (a) Cultured NK cells were generated from Stat4^−/− spleens as previously described and cells were stimulated for 6 h with the indicated cytokines and mRNA expression was measured by real-time PCR. Real-time PCR data are represented as a fold increase over nonstimulated, Stat4^−/− NK cells. The data are representative of three separate experiments. (b) Stat4^−/− NK cell culture supernatants were analyzed for IL-10 production after 24 h stimulation by enzyme-linked immunosorbent assay (ELISA). (c) Cultured NK cells were generated from Tbet^−/− spleens and stimulated with the indicated cytokines for 6 h. mRNA expression was quantified by real-time PCR. Real-time PCR data are represented as a fold increase over nonstimulated, Tbet^−/− NK cells. The data are representative of three separate experiments.

A conserved, intronic Stat4 element is located within a CNS region (CNS + 3.10) that is also a target for cytokine-induced chromatin remodeling

In order to identify potential mechanisms of Stat4 in regulating IL-10 expression, we scanned the Il10 gene and surrounding sequences for potential STAT-binding elements. Based on a computer search string of common STAT-binding motifs, we identified a putative Stat4 element in the fourth intron of the mouse Il10 gene based on sequence identity. As shown in Figure 7a, the STAT-like sequence maps to a CNS + 3.10 which has approximately 72% homology over more than 100 bp between mouse and human (http://genome.lbl.gov/vista/index.shtml). The STAT-motif identified in the mouse Il10 gene is conserved in humans as we identified a STAT-like element with 100% identity in the human Il10 gene, also located in the identified CNS region in fourth intron. To determine if IL-12-induced Stat4 to bind to this element, we preformed electromobility shift assays (EMSA) using this region as a probe. When cultured NK cells were treated with media alone (NS, lane 1), or with IL-2 (lane 2), no unique complex(es) were formed (Figure 7b). However, when NK cells were stimulated with IL-12 (lanes 3 and 4), a unique complex was formed. To determine if Stat4 was bound in the complex, we used a Stat4-consensus sequence to compete for binding by excess cold competition. The consensus Stat4 oligo blocked the formation of the IL-12-induced complex (lane 5), however an identical oligo with the Stat4 site mutated failed to compete for binding (lane 6). To confirm Stat4 binding, we performed supershift analyses by first adding an irrelevant antibody against SP-1 which did not block or displace formation of the IL-12-induced complex (lane 7). Likewise, we excluded the possibility that the related molecule Stat5 was bound to this region of the Il10 gene by incubating with anti-Stat5A and Stat5B antibodies which also did not block formation of the complex (lane 8). When an anti-Stat4 antibody was used however, the IL-12-induced complex was blocked, indicating that IL-12-induced Stat4 binds to the Stat4 element in the fourth intron of the Il10 gene.

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 + 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 + IL12 for 2 h.

To confirm that IL-12-activated Stat4 interacts with the 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 Il10 gene. However, when cells were stimulated either with IL-12 alone (lane 9) or IL-2 + IL-12 (lane 12), Stat4 was immunoprecipitated, bound to the fourth intron of the Il10 gene. To control for nonspecific antibody interactions, normal rabbit immunoglobulin G (IgG) was used as an immunoprecipitation control but no evidence of nonspecific binding was observed (lanes 2, 5, 8, 11). Thus, based on EMSA and ChIP analysis we conclude that IL-12 stimulation results in the recruitment of Stat4 to the mouse Il10 gene.

Since the CNS + 3.10 region is highly conserved and recent reports identified a DNaseI hypersensitivity site in the fourth intron of the mouse Il10 gene in T cells,^32,33 we questioned if other chromatin-modifying mechanisms were targeted to this region in NK cells. Thus, we used ChIP analysis to address the state of histone acetylation following cytokine stimulation (Figures 8a and b). Compared to nonstimulated NK cells (lane 3), we found no evidence of enhanced histone H3 acetylation in IL-2-treated cells (lane 6). Nonetheless, some degree of constitutive H3 acetylation was observed in NS and IL-2-stimulated cells (lanes 3 and 6). When NK cells were stimulated with IL-12 alone or in combination with IL-2, the levels of histone H3 acetylation were upregulated within the CNS + 3.10 site (lanes 9 and 12). To quantify levels of H3 acetylation, ChIP gel images were examined by denisometry (Figure 8b). Histone H3 acetylation amounts are represented as a relative ratio as compared to the input for each sample. Collectively, these data suggest that the fourth intron of the mouse Il10 gene is a target for both cytokine-induced chromatin modifications and transcription factor recruitment.

Figure 8.

The intronic +3.10 conserved noncoding sequence (CNS) region is also a target for cytokine-induced histone acetylation. (a) Chromatin immunoprecipitation (ChIP) analysis of acetyl histone H3 levels in cultured natural killer cells treated with media alone (NS), interleukin-2 (IL-2), IL-12 or IL-2 + IL12 for 6 h. (b) Denisometric analysis of histone H3 acetylation shown above. The data are expressed as a relative ratio to each respective input sample.

Discussion

A number of reports have accumulated over recent years suggesting a surprising role for IL-12 in regulating IL-10 expression in the lymphoid compartment, often in conjunction with IFN-γ.²² It is well established that IL-2 and IL-12 regulate IFN-γ autonomously and together are highly synergistic in T and NK cells.³⁴ Similarly, IL-2 and IL-12 have been reported to induce IL-10 expression individually and together are synergistic in both T and NK cells, but the mechanisms have not been explored.^35,36 In mouse and human T cells, IL-12 has been implicated in the differentiation of T-cell subpopulations that express both IL-10 and IFN-γ.^37,38 These IL10⁺ IFN-γ⁺ T cells are particularly well described in response to infection with certain intracellular pathogens, and are thought to involve T-bet.^39-42 Other reports indicate a direct role for IL-12 induction of IL-10 expression in vivo. Clinical administrations of recombinant human IL-12 to cancer patients resulted in elevated plasma levels of IL-10 but decreased IFN-γ.²¹ The biological significance of IL-10 and IFN-γ coexpression in NK or T-cell subsets is not clear but has been suggested to be part of a feedback mechanism during infection to restrict inflammatory responses perhaps at specific tissues sites.²² Despite a clear link between IL-12 and regulation of IL-10, the molecular mechanisms are ill defined.

There are surprising parallels between the receptor systems and signaling pathways, which regulate IL-10 and IFN-γ expression in NK cells (as well as T cells). Our results are in line with other reports indicating that IL-2 and IL-12 control IL-10 (and IFN-γ) expression in NK cells (Figures 1 and 2). The similarities between regulation of IL-10 and IFN-γ are not absolute however, as other established regulators of IFN-γ such as IL-18, Ly-49D signaling and T-bet, have no affect on NK cell-specific IL-10 expression (Figures 1 and 6c, respectively). We also found that the IL-12 homologs IL-23 or IL-27 do not regulate IL-10 (or IFN-γ) expression in NK cells (data not shown). This was true when IL-23 or IL-27 was added to NK cultures alone or in combination with IL-2 or IL-12, indicating a nonredundant role for IL-12 in controlling IL-10 expression in mouse NK cells. This is in contrast to recent reports which demonstrate an important role for IL-27 in the induction of IL-10 expression in mouse T-cell subsets.⁴³ In addition, another previous report indicated that IL-23 induces IL-10 expression from human cord blood derived, polyclonally activated CD4⁺ and CD8⁺ T cells.⁴⁴ The disparity in these data may suggest differences between IL-10 expression in mouse NK cells and T cells, and/or point to fundamental differences between mouse and human IL-10 regulation.

In that regard, our group and others have shown that in primary human CD56⁺ NK cells as well as the human NK cell line NK92, IL-2 and IL-12 stimulation also leads to IL-10 mRNA and protein expression (data not shown).³⁵ In addition, in both mouse and human NK cells, IL-18 potently stimulates IFN-γ but does not regulate IL-10 (data not shown).¹⁹ Since the majority of work on IL-10 has been done in the mouse, it will be important to determine tissue- and receptor-specific regulatory mechanisms in both mouse and human models. Nonetheless, the parallels between IL-10 and IFN-γ regulation in both species suggest some conserved genomic regulatory structures between the Il10 and Ifng genes that may harbor tissue-specific control elements.

As mentioned, IL-10 and IFN-γ coexpressing T cells have been reported previously, and we identified a small subset of cultured NK cells that accounts for IL-10 expression in response to IL-2 + IL-12 stimulation (Figure 3). These small numbers are in line with another report which also described a similar population size (1.1%) of IL-10-expressing CD56^high human NK cells following IL-2 + IL-12 treatment.²⁵ When we assessed IL-10 and IFN-γ accumulation in NK1.1⁺ NK cells, we observed that approximately 70% of all IL-10⁺ NK cells also express IFN-γ. The size of the IL-10⁺ NK cell pool is comparatively small to the numbers of IFN-γ-producing NK cells, however (about 46%). One group has suggested that acquiring the ability to produce IL-10 in human NK cells is associated with dying NK cells and thus may represent a stage of the NK cell life cycle.²⁵ We have not directly addressed this question here, but future studies will be directed at defining the functional characteristics of IL-10 expressing mouse NK cell subsets.

Numerous signaling cascades have been implicated in IFN-γ regulation including the p38 MAPK pathway.^27,45-47 We have reported previously that up to 80% of IL-12-induced and 60% of IL-2 + IL-12-induced IFN-γ could be blocked by p38 inhibitors in NK cells.¹⁹ With respect to IL-10, we determined here that up to 72% of IL-2 + IL-12-induced IL-10 could be blocked by inhibitors of the p38 pathway (Figure 7). Interestingly, both the p38 and ERK MAPK pathways have been linked to recruitment of Sp1 to the IL-10 promoter.^48-50 However, similar to the case of p38 regulation of IFN-γ, these data suggest an important, but not absolute role for IL-2 + IL-12-induced p38 signaling in regulating IL-10.

p38 is one of several kinases upstream of the transcription factor Stat4.⁵¹ Stat4 has an established role in regulating IFN-γ in T and NK cells, most likely via a Stat4 element in the proximal promoter.⁵² Our findings indicate that Stat4 is required for IL-12 induction of IL-10 and the synergy between IL-2 and IL-12 is also largely dependent on the presence of Stat4 (Figures 6a and b). IL-2 induction of IL-10 is essentially intact in Stat4^−/− NK cells, however, suggesting there is no ‘master regulatory’ function for Stat4 in the control IL-10 transcription.

IL-2 also induces IL-10 in NK cells and the combination of IL-2 and IL-12 is highly synergistic for IL-10 and IFN-γ. The concept of synergy between cytokine signaling pathways is a recurring question in the literature but there has been no clear resolution. We reported previously that IL-2 and IL-12 cotreatment results in enhanced tyrosine phosphorylation of both Stat4 and Stat5.¹⁹ It is likely that other converging mechanisms play an important role in the synergistic effects of IL-2 and IL-12 on expression of target genes such as IL-10 and IFN-γ.

Nevertheless, we questioned if there is a conserved component to IL-2 cytokine induction of IL-10 via the shared common γ-chain. We stimulated cultured NK cells with other cytokines which share the common γ-receptor (IL-4 and IL-15) and found there was no induction of IL-10 which suggests that neither signaling through the common γ-chain or Stat5 (induced by IL-15) is sufficient to regulate IL-10, alone or in combination 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-γ in NK cells.¹⁹ It will be interesting to determine if IL-2 regulates IL-10 through a Stat5-dependent mechanism which is the case for IFN-γ.²⁴

We observed another key disparity between the regulation of IL-10 and IFN-γ in T-bet null NK cells. Cytokine induction of IL-10 was not affected in T-bet^−/− NK cells despite the fact that T-bet is required for optimal expression of IFN-γ in cultured NK cells.¹⁹ Conversely, another group recently reported increased IL-10 expression in T-bet-deficient mice infected with Mycobacterium tuberculosis.⁵³ The authors found there was enhanced expression of IL-10 but impaired production of IFN-γ by CD4⁺ T cells in M. tuberculosis-infected mice. Furthermore, another recent report describes a T-bet⁺Foxp3⁻ Th1 cell population as the primary source of IL-10 during Toxoplasma gondii infection.⁴¹ It will be interesting to determine if there are cell-specific requirements for T-bet in regulation of IL-10 outside of the NK lineage.

We identified a conserved Stat4 DNA-binding element within the fourth intron of the Il10 gene that is located within a CNS (CNS + 3.10) of both the mouse and human Il10 genes (Figures 8a and b). In the case of IFN-γ, a presumption has been that Stat4 regulates acute IFN-γ transcription by binding Stat4 motifs in the Ifng gene. Stat4-binding elements have been identified in both the mouse and human Ifng genes, however, these are largely not conserved.³⁰ In fact, multiple mechanisms of Stat4 regulation of IFN-γ have emerged which also include the recruitment of other transcription factors and chromatin-modifying proteins to the Ifng gene.³⁴ In that regard, we found that the CNS + 3.10 region, which contains the Stat4 site, is also a target for cytokine-induced chromatin remodeling based on the enhancement of histone H3 acetylation levels following cytokine stimulation (Figure 8c). It will, therefore, be interesting to determine if Stat4 (and this conserved Stat4-element) plays a role in regulating IL-10 expression in other cell populations as well as in human cells.

Recent reports have identified intronic regions and/or sites 3′ to Il10 that are remodeled in T cells which express IL-10.^54,55 In fact, another study indicated the presence of a Th1-specific HS site in the fourth intron of the mouse Il10 gene (HSII) which is located within or adjacent to the +3.10CNS we characterized here.³² An even more comprehensive analysis of the mouse Il10 gene structure and expression profiles was recently reported which focused on Th1/Th2 cells.³³ The authors identified multiple DNase I hypersensitivity sites between the Il10 and Il19 genes and described three distal enhancer elements. Of note, the authors reported the presence of a non-Th-lineage-specific DNaseI hypersensitivity site in the fourth intron at position +2.98 (HSS + 2.98). Although this is presumably the same HS site reported by Im et al. (HSII),³² there are inconsistencies as to whether this HS is Th-lineage-specific. Nonetheless, these reports in T cells and our data in NK cells suggest that the fourth intron of the mouse Il10 gene contains essential regulatory elements including a Stat4-binding site.

Overall, very little is known about transcriptional control regions and the genomic/epigenetic requirements for signal- and cell-specific IL-10 expression. Our report in NK cells and recent work from other groups provides an insight into tissue-specific IL-10 expression, cis-acting factors, and chromatin structure within the murine Il10 gene and flanking regions.^48-50,56-60 It will be important to resolve the genomic boundaries, epigenetic modifications and transcriptional regulators in different IL-10-expressing cell populations in relation to the human IL-10 in an attempt to consolidate and apply our understanding of murine IL-10 regulation and function to humans.

Materials and methods

Cytokines and antibodies

Recombinant human IL-2 was obtained from Hoffmann-La Roche (Nutley, NJ, USA). Recombinant mouse IL-12 and IL-18 was obtained from Peprotech (Rocky Hill, NJ, USA) and MBL (Watertown, MA, USA), respectively. Recombinant IL-23, IL-27 and IFN-γ, and IFN-γ neutralizing antibodies were purchased from R&D Systems (Minneapolis, MN, USA). Purified hamster anti-mouse CD3ε (145−2C11), NK1.1, IFN-γ and IL-10 mAbs were purchased from BD Bioscience (San Diego, CA, USA). Isotype-matched Ig for fluorescence-activated cell sorting (FACS) analyses was also purchased from BD Bioscience. For ChIP assays, anti-histone H3 antibodies and normal rabbit IgG were purchased from Upstate Biotechnology (Temecula, CA, USA). The same anti-Stat4 was used in EMSA and ChIP analysis, Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-Stat5 was purchased from R&D Systems. The mAb 4E5 or an isotype control mAb, followed by goat anti-rat (F(ab′)₂) was used to cross-link the Ly-49D receptor as previously described.⁶¹

Mice

C57BL/6 mice were used in all experiments except where indicated. Mice were maintained under specific pathogen-free conditions and were used between 8 and 14 weeks of age. All animal studies in this report were approved by the Johns Hopkins University Institutional Animal Care and Use Committee.

NK cell isolation

Murine splenic NK cells were isolated as previously described.⁶² Briefly, single cell suspensions were prepared by passing the spleens through a wire mesh screen. RBC were lysed with ACK reagent (Biowhittiker, Walkerville, MD, USA) and cells were washed and resuspended in 5% fetal bovine serum (FBS) RPMI 1640. The spleen suspension was passed through a sterile, prewetted nylon wool column and following a 50 min incubation at 37 °C, the nylon wool-nonadherent cells were harvested, washed and counted. The nylon wool-nonadherent cells were resuspended in 10% FBS RPMI media supplemented with high dose IL-2 (1000 U ml⁻¹) at a density of 2 × 10⁶ per ml. NK cells were cultured for 6−8 days and were typically between 75 and 90% DX5⁺ or NK1.1⁺ before harvesting for experiments.

Flow cytometry and NK cell purification

FACS was performed by staining cells with phycoerythrin (PE)-conjugated NK1.1 and fluorescein isothyocyanate (FITC)-conjugated anti-mouse CD3 to determine NK cell purity in cultured NK cell cultures. Results were analyzed on a FACS Calibur (BD Bioscience). To purify NK1.1⁺CD3⁻ cells, following nylon wool separation, cells were negatively sorted by MoFlo (Cytomation, Fort Collins, CO, USA) using FITC-conjugated anti-mouse CD3 or PE-conjugated NK1.1. Single positive NK1.1⁺CD3⁻ cells were washed and directly stimulated with cytokines.

For intracellular staining (ICS) experiments, BD Cytofix/Cytoperm kit was used according to the manufactures instructions (BD Bioscience). Analysis of ICS experiments was performed on a FACSCalibur (BD Bioscience). Data were collected with CellQuest software and analyzed with FlowJo software. Live cells were gated according to their forward-scatter and side-scatter profiles. Cell counts were made using a hemocytometer, and verified by FACS, when necessary, using a bead standard for calibration.

mRNA analyses

Total RNA was isolated by guanidinium-isothescyanate phenol/chloroform extraction method (Trizol; Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. RNA concentrations were quantified and subsequently normalized before cDNA synthesis. cDNA was generated using a first strand cDNA synthesis kit (Roche, Indianapolis, IN, USA). Quantitative PCR was performed using Taqman site-specific primers and probes (Applied Biosystems, Foster City, CA, USA) on an ABI PRISM 7700 sequence detector. Results were normalized to β-2 microglobulin levels. For relative comparisons, NS culture conditions were assigned an arbitrary value of one.

Cell culture and stimulations

All tissue culture (unless otherwise mentioned) media contained RPMI 1640 medium supplemented with 10% heat-inactivated FBS, 2 μm l-glutamine, 1 × essential amino acids, 1 mm sodium pyruvate, 10 mm 2-mercaptoethanol, 100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin. Cultured NK cells were washed two times in 2−10% fetal calf serum RPMI 1640 and plated at 5 × 10⁶ cells per ml in six-well plates and incubated with the indicated cytokine(s) for 2−18 h depending on the experiment. Although dose response curves were completed for all of the cytokines tested, optimal doses were determined for IL-10 production and used in each experiment as follows: IL-2 (100 U ml⁻¹) and IL-12 (10 U ml⁻¹). Fresh, highly purified NK cells were plated at 1 × 10⁶ cells per ml in 48-well plates. All cells were rested for 2−3 h at 37 °C prior to stimulation (unless otherwise noted). For MAP kinase inhibition studies, p38 inhibitors were used at 600 nm. SB203474 was used as a negative control and was also used at a concentration of 600 nm (Calbiochem, La Jolla, CA, USA). The inhibitor concentrations were within the range of doses previously reported to have inhibitory effects on MAP kinases.⁶³ The inhibitors were dissolved in sterile DMSO and stored at −20 °C at a concentration of 20 mm.

Electrophoretic mobility shift assay

Complementary single-stranded oligonucleotides were commercially synthesized (Operon Biotechnologies Inc., Huntsville, AL, USA) to span approximately 10 bp on either side of the putative Stat4 element: 5′-AGGCCACATGGCTTCTGGGAACTAGGGTTG-3′. The consensus Stat4 oligo used in cold-competition assays was as follows: 5′-GAGCCTGATTTCCCCGAAATGATGAGCTAG-3′. The mutant Stat4 oligo was identical to the consensus oligo except for a ‘CCC’ substitution of ‘TTT’ shown in bold. Briefly, complementary strands were annealed by combining 2 μg of each oligonucleotide and 6 μl of 10 × annealing buffer (500 mm Tris, 100 mm MgCl₂ and 50 mm dithiothreitol) in a 60 μl reaction, denatured in a boiling water bath for 5 min and then allowed to cool to room temperature.⁶⁴ Double-stranded DNA overhangs were labeled with Klenow enzyme by a fill-in reaction, [α-³²P]dCTP and free dNTPs (Amersham Biosciences Corp., Piscataway, NJ, USA). The DNA-protein binding reaction was conducted in a 20 μl reaction mixture consisting of 5−10 μg of nuclear protein extract, 1 μg poly(dI-dC) (Sigma, St Louis, MO, USA), 4 μl of 5 × binding buffer (60 mm Hepes, 7.5 mm MgCl₂, 300 mm KCl, 1 mm ethylenediamine-tetraacetic acid, 2.5 mm dithiothreitol, 50% glycerol and 4-(2-aminoethyl)-benzenesulfonyl flouride hydorochloride) and 1.5 × 10⁴ c.p.m. of ³²P-labeled oligonucleotide probe. For supershift analysis to identify Stat proteins interacting with the Il10 promoter, 1 μl of antibody was added to the reaction prior to the addition of the probes and incubated on ice for 20 min. In cold oligonucleotide competition experiments, 10- to 100-fold excess of unlabeled oligonucleotide probe was added 10 min prior to adding the radiolabeled oligonucleotide probe.

Nuclear extraction

Nuclear extracts were prepared from cytokine-stimulated cultured NK cells that had been rested for 4 h in the absence of IL-2 as described previously.⁶⁵ Briefly, following NS or treatment with cytokines for 20−60 min, cell pellets were resuspended in lysis buffer (50 mm KCl, 25 mm Hepes, pH 7.8, 0.5% Nonidet P-40, 1 mm phenylmethylsulfonyl fluoride, 10 μg ml⁻¹ leupeptin, 20 μg ml⁻¹ aprotinin, 100 μm dithiothreitol) and subsequently incubated on ice for 5 min. Cellular suspensions were collected by centrifugation at 2000 r.p.m. and the supernatant was harvested as the cytoplasmic protein fraction. Nuclei were washed in wash buffer (lysis buffer without 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% glycerol, 1 mm phenylmethylsulfonyl fluoride, 10 μg ml⁻¹ leupeptin, 20 μg ml⁻¹ aprotinin and 100 μm dithiothreitol), frozen in dry ice, thawed slowly on ice and finally centrifuged at 14 000 r.p.m. for 10 min. The supernatant was harvested and nuclear proteins quantified with the bicinchoninic acid protein assay reagent (Pierce, Rockford, IL, USA).

Chromatin immunoprecipitation assays

ChIP assays were performed according to the manufacturer's instructions (Upstate Biotechnology). Briefly, 2 × 10⁶ cells were fixed with 1% formaldehyde, washed with cold phosphate-buffered saline, and lysed in buffer containing 10 μg ml⁻¹ aprotinin (ICN Biomedicals, Aurora, OH, USA), 10 μg ml⁻¹ leupeptin (Bachem, Torrance, CA, USA) and 2.5 μm NPGB (4-nitrophenyl 4-guanidinobenzoate hydrochloride, Sigma). Nuclei were sonicated for a total of 60 s to shear DNA (Heat Systems, Farmingdale, NY, USA), the lysates were pelleted, and supernatants were diluted. Diluted lysates were precleared with salmon sperm DNA/protein A agarose for 30 min prior to immunoprecipitation with specified antibodies. A proportion (2%) of the diluted supernatant was kept as ‘input’ to quantify genomic DNA. After immunoprecipitation, the protein–DNA complexes were incubated with Protein A beads, and the protein–DNA complexes were eluted in 1% SDS/0.1 M NaHCO₃, and cross-links were reversed at 65 °C. DNA was recovered by phenol-chloroform extraction and ethanol precipitation, and then subjected to PCR and/or real-time PCR analysis. PCR was carried out with ampliTaq Gold (Applied Biosystems) for 35 cycles (45 s at 95 °C, 60 s at 59 °C, and 60 s at 72 °C), and the products were visualized by ethidium bromide staining. The immunoprecipitated DNA samples were normalized to ‘input’ DNA samples by densitometry. PCR primers for the fourth intron +3.01CNS site in the murine Il10 gene; 5′-CTCAGGTACATCATTG-3′ and 5′-GAGTGTGTAGGCAGTC-3′.

Cytokine secretion assays and statistical analysis

Cell-free supernatants were collected and assayed for IL-10 and IFN-γ production by ELISA (R&D Systems). The sensitivity limits for the assays were <2 pg ml⁻¹. Comparative data were analyzed using the unpaired Student's t-test. The software used to perform the statistical analysis was SigmaPlot 2000 for Windows version 6.10.