STAT4-mediated transcriptional repression of the IL5 gene in human memory Th2 cells

Abstract

Type I Interferon (IFN-α/β) plays a critical role in suppressing viral replication by driving the transcription of hundreds of interferon sensitive genes (ISGs). While many ISGs are transcriptionally activated by the ISGF3 complex, the significance of other signaling intermediates in IFN-α/β-mediated gene regulation remains elusive, particularly in rare cases of gene silencing. In human Th2 cells, IFN-α/β signaling suppressed IL5 and IL13 mRNA expression during recall responses to T cell receptor (TCR) activation. This suppression occurred through a rapid reduction in the rate of nascent transcription, independent of de novo expression of ISGs. Further, IFN-α/β-mediated STAT4 activation was required for repressing the human IL5 gene, and disrupting STAT4 dimerization reversed this effect. This is the first demonstration of STAT4 acting as a transcriptional repressor in response to IFN-α/β signaling and highlights the unique activity of this cytokine to acutely block the expression of an inflammatory cytokine in human T cells.

Introduction

Cytokines such as IFN-α/β, play an important role in coordinating cellular responses to external stimuli, particularly viral pathogens. In addition, IFN-α/β plays a central role in regulating adaptive T cell responses by driving both effector and memory cell development [1–5]. Conversely, our recent studies demonstrated that IFN-α/β blocked human Th2 development by selectively inhibiting IL-4-mediated expression of GATA3 [6, 7]. IFN-α/β promoted epigenetic modifications and repression of the GATA3 gene, which correlated with a reduction in IL-4-mediated GATA3 binding to an upstream enhancer [6]. Thus, as T cells divide and develop in response to IL-4, IFN-α/β dominantly suppresses Th2 programming. Distinct from this mechanism is the observation that IFN-α/β suppresses the production of IL-5 and IL-13, but not IL-4 in human CD4⁺ T cells and PBMCs in response to acute TCR stimulation [8, 9]. For example, house dust-mite-induced IL-5 and IL-13 production by human PBMCs was suppressed by IFN-β treatment [10]. These observations suggest an important role for acute IFN-α/β signaling in counter-regulating the secretion of inflammatory Th2 cytokines upon antigen- or allergen-mediated stimulation.

In the present study, we demonstrate that IFN-α/β signaling blocks IL-5 and IL-13 secretion from purified human memory Th2 cells, IFN-α/β induced STAT4 binding within the human IL5 promoter, which correlated with suppressed nascent transcription. Here, we show that STAT4 is sufficient for IFN-α/β-mediated repression of the human IL5 promoter, and highlights an unprecedented role for STAT molecules in transcriptional repression.

Results and Discussion

IFN-α/β selectively suppresses IL-5 and IL-13 expression

Early studies observed that IFN-α/β could suppress IL-5 and IL-13 expression from human CD4⁺ T cells and PBMCs [8, 9]. However, it was unclear whether this effect was direct or regulated through additional mediators. We addressed this issue by determining whether IFN-α/β could directly suppress cytokine expression from a highly purified population of human memory Th2 cells (CRTH2⁺CD4⁺ T cells), which are enhanced in allergic diseases [11]. As expected, CRTH2⁺ cells secreted the majority of IL-4, IL-5, and IL-13, while the CRTH2⁻ cells secreted very little Th2 cytokines. Further, the CRTH2⁻ cells secreted high levels of IFN-γ, which was undetectable in CRTH2⁺ cells (Figure 1A). IFN-α/β signaling suppressed TCR-mediated IL-5 and IL-13 expression from CRTH2⁺ cells, while enhancing IL-4 production (Figure 1A, top panel). Likewise, IFN-α/β treatment suppressed IL5 and IL13 mRNA expression in memory CD4⁺CD45RO⁺ T cells, which include CRTH2⁺CD4⁺ memory Th2 cells. In each donor, IFN-α/β suppressed TCR-induced IL5 and IL13 gene expression, while IL4 expression was unaffected (Figure 1B). TCR-mediated IFNG mRNA was enhanced in response to IFN-α/β treatment (Figure 1B), which has been previously reported [9, 12]. Because of the heterogeneous nature of total CD4⁺CD45RO⁺ T cells, it was possible that the effect was induced by other cytokines within these cultures. However, regardless of whether IL-4, IL-12 and IFN-γ were neutralized (“Neut”, Figure 1B, upper panels) or left intact (“Drift”, Figure 1B, lower panels), IFN-α/β suppressed IL5 and IL13 gene expression. Further, IFN-α(2a), IFN-β and IFN-ω similarly suppressed TCR-mediated IL5 and IL13 gene expression while enhancing IFNG expression in a dose-dependent manner (Supporting Information Figure 1A). In addition to IFN-α/β, type II interferon (IFN-γ) and type III interferons (IFN-λ1, IFN-λ2, IFN-λ3) are expressed in response to both bacterial and viral infections. However, only IFN-α/β suppressed IL-5 and IL-13 protein and mRNA expression, while IFN-γ and IFN-λ1 failed to suppress these cytokines (Supporting Information Figure 2A and B). IFN-λ1 treatment modestly enhanced MX1 expression compared to the potent induction by IFN-α/β (Supporting Information Figure 2C), strongly suggesting that despite similarities between IFN-α/β and IFN-λ signaling, memory CD4⁺ T cells fail to respond to IFN-λ. Collectively, these findings demonstrate that IFN-α/β suppresses TCR-induced IL5 and IL13 gene expression, but leaves IL4 expression intact.

Figure 1. IFN-α/β suppresses IL5 and IL13 expression in human memory Th2 cells.

(A) Human memory Th2 cells (CD4⁺CRTH2⁺) were stimulated with anti-hCD3 and rhIL-2 +/− rhIFN-α(2a) The concentration of different cytokines on cell supernatants was measured by ELISA. Data are shown as mean +SD/SEM (n=3) and are representative of 3 independent experiments. (B) Purified human CD4⁺CD45RO⁺ T cells were stimulated with anti-hCD3, rhIL-2, anti-hIFN-γ, anti-IL-4, and +/− rhIFN-α(2a). ‘Neut,’ ‘Drift,’ and ‘IFN-α’ treatments are expressed relative to a ‘No TCR’ control, and are referenced to PPIA. Each line represents an individual experiment (n= 3–8). * p≥ 0.05, **p ≥ 0.01, ns: not significant; Student’s t-test.

Nascent IL5 and IL13 transcription are rapidly suppressed by IFN-α/β

Because cytokine-driven STAT activation occurs within minutes, the kinetics of IL5 mRNA suppression was measured by stimulating human lymphocytes in the presence of IFN-α/β at the time of TCR stimulation or 2 hrs post-TCR stimulation (Figure 2A). TCR stimulation enhanced IL5 and IFNG gene expression, and IFN-α/β added at ‘0-hr’ suppressed IL5 while enhancing IFNG. Interestingly, IL5 expression in cells that received IFN-α/β treatment at ‘2-hrs’ was similarly inhibited compared to cells that received IFN-α/β at ‘0-hr’. This potent block of IL5 expression was confirmed by measuring IFNG, in which enhanced gene expression in response to IFN-α/β treatment was delayed (Figure 2A), suggesting that IFN-α/β signaling suppressed IL5 expression minutes after receptor ligation. Nonetheless, it was possible that repression was being mediated indirectly through the induction of an ISG with repressor activity. This possibility was tested by determining whether the inhibition of de novo protein synthesis altered IFN-α/β-mediated suppression of IL5 and IL13 mRNA expression (Figure 2B). IFN-α/β suppressed both IL5 and IL13 mRNA expression despite blocking protein synthesis with cycloheximide, leaving TCR-mediated IFNG mRNA induction by TCR stimulation intact. These findings suggest that the initial signaling events that suppress IL5 and IL13 gene expression occur directly in response to IFN-α/β signaling, and do not require the expression of ISGs.

Figure 2. IFN-α/β suppresses acute IL5 and IL13 expression by directly affecting the rate of nascent transcription.

(A) Total human lymphocytes were stimulated with anti-hCD3 and rhIL-2, and IFN-α(2a) was added either with TCR stimulation (white), or 2 hrs post-TCR stimulation (orange). Each treatment is relative to its own 0-hr time point. (B) Human CD4⁺CD45RO⁺ T cells were TCR-stimulated as in (A) and then treated with IFN-α +/− cycloheximide or DMSO. Vehicle or CHX samples are analyzed relative to their respective ‘No TCR’ controls. (C) Total human lymphocytes were stimulated as in (A), and IFN-α(2a) and/or actinomycin D was added at 2 hrs post-TCR stimulation (arrow). Each treatment is relative to its own 0-hr time point. (D) Polarized human Th2 cells were restimulated as in (A) +/− IFN-α(2a) or actinomycin D (10 µg/ml). Nuclei were subjected to in vitro transcription to label nascent transcripts with biotin-UTP, and qPCR was used to analyze the rate of nascent transcription. Each treatment is relative to the ‘No TCR’ control. All qPCR samples were referenced to PPIA. (A–D) Data are shown as mean +- SEM (n=3) and are representative of at least 2 independent experiments. (B and D) Two-way ANOVA in (B), and one-way ANOVA in (D) with a Bonferroni post-hoc test were used to determine statistical significance. *p ≥ 0.05, **p ≥ 0.01, *** p≥ 0.001, ****p ≥ 0.0001, ns: not significant.

IFN-α/β could reduce the expression of IL5 transcripts either by enhanced mRNA decay or by suppressing TCR-mediated nascent transcription. To address this, experiments were performed to assess IL5 mRNA stability versus alterations in the rate of de novo transcription. Actinomycin D-mediated suppression of nascent transcription revealed no enhancement in the rate of IL5 mRNA decay by IFN-α/β treatment (Figure 2C, open squares versus open triangles). However, nuclear run-on assays demonstrated that IFN-α/β signaling reduced the rate of TCR-mediated nascent transcription of IL5 and IL13 (Figure 2D). TCR-stimulation enhanced the rate of nascent transcription of all cytokine genes measured, while IFN-α treatment reduced the rate of nascent transcription of IL5 and IL13, while enhancing IFNG and leaving IL4 unaffected. Taken together, these data demonstrate that IFN-α/β potently reduces the expression of IL5 and IL13 transcription by directly inhibiting the rate of nascent transcription.

STAT4 regulates IFN-α/β-mediated gene suppression

In addition to the ISGF3 complex, multiple STAT dimers become activated by IFN-α/β, including STAT4, and have the ability to bind compound GAS elements. An in silico analysis of the human IL5 promoter was performed using Jaspar [13], which identified 146 putative STAT binding elements within the 2 kb human IL5 promoter. Notably, within the -300 bp proximal promoter, we identified an overlapping GAS/ISRE consensus motif (Supporting Information Figure 3). To determine whether this GAS/ISRE element was bound by IFN-α/β-induced STAT proteins, we probed human PBMC nuclear lysates with the GAS/ISRE DNA probe (red line, Supporting Information Figure 3) in the presence of STAT-specific antibodies (Figure 3A). The EαY box factor served as the loading control [6]. As demonstrated by supershift analysis, STAT1 and STAT4, but not STAT2 or STAT3, bound the GAS/ISRE probe in response to IFN-α/β treatment (Figure 3A). This site within the proximal promoter is a single example of other sites upstream with the potential to be bound by IFN-α/β-induced STATs.

Figure 3. STAT4 binds to the human IL5 promoter, and is required for IFN-α/β-mediated gene suppression.

(A) Human PBMCs were stimulated with anti-hCD3 and rhIL-2 for 2 hrs in the presence (α; lanes 3 – 8) or absence (ctrl; lane 2) of IFN-α(2a). Nuclear lysates were probed with EαY (a loading control) or hIL-5pro dsDNA probe +/− specific STAT antibodies (lanes 5–8) or a cold competitor probe (CC, lane 4). The arrow indicates super-shifted proteins. Blots are representative of 3 independent experiments. (B) Polarized human Th2 cells from 4 donors were restimulated with anti-hCD3 and rhIL-2 for 2 hrs +/− IFN-α(2a), and ChIP lysates were probed for STAT2 or STAT4 binding. qPCR was used to analyze the human IL5 proximal promoter, the human MX2 promoter (STAT2 site), and human IL2RA intron 1 (STAT4 site). (C) SV40 Jurkat cells were co-transfected with the listed plasmids and rested 24 hrs in antibiotic-free media. Cells were stimulated with soluble anti-hCD3 and rhIL-2 +/− IFN-α(2a) for 6 hrs, and firefly and renilla luciferase were quantified. (B and C) Data are shown as mean +/−SEM (n=3) and are representative of three independent experiments. * p≥ 0.05, ** p≥ 0.01, *** p≥ 0.001, **** p≥ 0.0001, ns: not significant; (B) Student’s t-test Or (C) two-way ANOVA with a Bonferroni post-hoc test.

To verify that STAT4 binds to this region in human Th2 cells, we performed ChIP and assessed the human IL5 promoter (blue line indicates ChIP-qPCR amplicon, Supporting Information Figure 3). It was important to distinguish the activities of STAT4 and STAT2, as both are exclusively activated by IFN-α/β, but not by IFN-γ. The myxovirus resistance 2 (MX2) promoter and IL2RA intron 1 were probed for STAT2 or STAT4 binding. IFN-α/β treatment enhanced respective STAT binding to these sites (Figure 3B). Further, STAT2 binding to the proximal human IL5 promoter was not consistently altered in response to IFN-α/β treatment, while STAT4 binding to this site was significantly enhanced in all subjects (Figure 3B). These results demonstrate that IFN-α/β-activated STAT4 binds within the proximal human IL5 promoter in human Th2 cells.

To determine the functional role of STAT4 in repressing the IL5 promoter, reporter assays were performed with the -300 bp sequence of the human IL5 promoter driving luciferase (hIL-5pro-Fluc). Jurkat T cells transfected with a hIL-5pro-Fluc construct failed to upregulate luciferase expression in response to TCR stimulation due to low levels of GATA3, which is required for IL5 transcription (Figure 3C) [14]. These cells also lack significant expression of STAT4. Transcriptional activation was rescued by co-transfection with a GATA3-GFP construct, however, the hIL-5pro-Fluc plasmid remained insensitive to IFN-α/β-mediated suppression (Figure 3C). IL-5 promoter activity was suppressed by IFN-α/β only when cells were co-transfected with a vector expressing STAT4, but not STAT1, indicating a unique role for STAT4 in repressing IL-5. Further, co-transfecting cells with an activation-defective STAT4-GFP point-mutant (ST4-R599K) [15] reversed IFN-α-mediated hIL-5pro-Fluc suppression. This mutation prevents STAT4 dimerization, thus preventing translocation to the nucleus. Although the suppression of the IL-5 promoter could occur through STAT4-mediated reduction in GATA3, we did not observe GATA3 mRNA suppression in response to IFN-α/β treatment during acute TCR activation (Supporting Information Figure 4A). Further, neither the expression of STAT4 nor treatment with IFN-α/β altered GATA3 protein expression in transfected Jurkat cells (Supporting Information Figure 4B). Thus, IFN-α/β-mediated activation of STAT4 was sufficient for IFN-α/β-mediated suppression of the human IL5 promoter, which correlated with STAT4 binding to the human IL5 promoter by ChIP.

This report is the first to demonstrate a direct role for IFN-α/β-mediated STAT4 activation in gene repression. Studies of transcriptome profiling suggest that both STAT4 and STAT6 are involved in suppressing a subset of genes during CD4⁺ T cell lineage commitment [16]. However, very little is understood regarding the consequences of IFN-α/β-mediated STAT4 activation in human CD4⁺ T cells. As human Th2 cells fail to respond to IL-12 signaling due to lack of expression of the IL-12Rβ2 subunit [17], IFN-α/β-induced STAT4 activation is unique in its ability to regulate Th2 cytokine production in fully differentiated human Th2 cells. A number of possible mechanisms could account for STAT4-mediated repression of IL5, including the recruitment of a direct repressor or a chromatin modifying enzyme. However, treatment of cells with Trichostatin A failed to block the effects of IFN-α/β in suppressing IL5 mRNA expression (data not shown). Alternatively, STAT4 could simply act to displace a direct positive regulator, such as GATA3 [18]. Indeed, GATA3 is important for inducing the expression of the Th2 cytokine genes, especially IL5 [14]. However, IFN-α did not significantly reduce GATA3 binding within the proximal IL5 promoter due to the low efficiency of GATA3 binding to this site as compared to the isotype control (Supporting Information Figure 5). However, we were able to detect GATA3 binding to the IL4 conserved intronic regulatory element (CIRE) element in polarized Th2 cells (Supporting Information Figure 5B), which agrees with our previous studies demonstrating that this site is bound by GATA3 [6]. Since GATA3 binds to the IL-5 promoter at such low levels, it was not possible to determine whether STAT4 activation in response to IFN-α/β signaling blocked GATA3 from binding at this region. Another possibility is that IFN-α/β signaling via STAT4 could be preventing other transcriptional enhancers from binding to the promoter, thus inhibiting TCR-mediated nascent transcription. It is unclear why IFN-α/β does not negatively regulate IL-4 expression, however, discordant regulation of the Th2 cytokines has been described previously [19, 20].

Our findings strongly implicate STAT4 in IFN-α-mediated suppression of two key Th2 cytokines, IL-5 and IL-13. Given the prominent role that these cytokines play in allergic diseases, this discovery underscores the potential for the IFN-α/β signaling pathway in controlling acute inflammation driven by Th2 effector cells, particularly in the context of viral infections [21, 22]. Recently, a single nucleotide polymorphism (SNP) located within intron 11 was found to be associated with high IgE levels in Korean allergic asthma patients sensitive to dust mite [23]. How this polymorphism affects STAT4 expression or translation is unknown. Other SNPs were identified in this study, particularly within the STAT4 promoter region, yet none of them associated with any disease risk.

The STAT4 protein exists in two forms, α and β, due to alternative splicing of the last exon encoding the C-terminal transactivation domain [24]. In Th1 cells, both isoforms participate equally in Th1 commitment. However, only the α isoform efficiently drives IFN-γ secretion in response to IL-12 + IL-18 stimulation, indicating a selective role for the α isoform in regulating cytokine-mediated gene transcription. Moreover, the ratio of these isoforms was found to be altered within PBMCs and colonic specimens from patients with Crohn’s disease and ulcerative colitis [25]. It is interesting to speculate how these two isoforms may either be differentially expressed or utilized in T cells from allergic subjects. If the β isoform is selectively mobilized by IFN-α/β signaling, it may block the suppressive activity of IFN-α on Th2 cytokine expression, leading to a preferential skewing toward Th2 development in conjunction with viral infections.

The current study demonstrates a negative regulatory pathway induced by IFN-α/β signaling that acts through STAT4 to suppress IL-5 and IL-13 expression. IL-5 and IL-13 are highly inflammatory cytokines in the context of allergic inflammation, inducing eosinophil activation and survival as well as enhancing mucus production, respectively [26, 27]. This signaling pathway could be targeted therapeutically for the treatment of atopic diseases, especially in cases in which conventional therapies are not adequate to control disease symptoms.

Materials and Methods

Human Donors

Peripheral blood was collected by venipuncture from healthy adults. Informed consent was obtained all donors in accordance with the Internal Review Board guidelines (UT Southwestern Medical Center, Dallas TX).

Antibodies, Cytokines, and Reagents

Recombinant human IL-4 (rhIL-4), rhIFN-γ, anti-human IL-4 (anti-hIL-4), and mouse IL-4 (rmIL-4) were purchased from R&D. rhIFN-α(A), rhIFN-ω, anti-hIFN-α/β receptor (IFNAR2) antibody were purchased from PBL Laboratories. rhIL-2 was obtained from the NIAID Resources for Researchers. Rabbit polyclonal antisera against STAT1, (H-119), STAT2 (C-20 and L-20), and STAT4 (C-20 and E-23) were purchased from Santa Cruz Biotechnology, Inc. Rabbit polyclonal antisera against STAT3 was purchased from Cell Signaling Technologies. Purified rabbit polyclonal antibody against GATA3 (H-48) was purchased from Santa Cruz.

T Cell Isolation and Culture

CD4⁺CD45RA⁺, CD4⁺CD45RO⁺, and CD4⁺CRTH2⁺ T cells (≥ 90%) were enriched by FACS or by bead kit as described previously [1]. Cells were stimulated with plate-bound anti-hCD3 (3 µg/ml) in the presence of rhIL-2 (50 U/ml).

Quantitative PCR

Quantitative PCR was conducted as described previously [6]. Brilliant II SYBR Green Master Mix (Agilent Technologies), Maxima SYBR Green Master Mix (Thermo Scientific) or TaqMan Gene Expression Mix (Applied Biosystems) was used to assess gene expression. Primers are listed in Supporting Information Table 1. The relative gene expression of specific cytokine genes were assessed using the 2^−ΔΔCt approach [28].

Nuclear Run-On

Nuclear run-on was performed as described previously [29]. Polarized human Th2 cells were restimulated for 2 hrs +/− IFN-α or actinomycin D. Nuclei were prepared and nascent transcripts were labeled with reaction buffer containing Biotin-rUTP. Transcripts were captured with streptavidin magnetic beads (Invitrogen). cDNA was synthesized from the beads followed by qPCR to measure the relative rate of nascent transcription of human IL5, IL13, IFNG and IL4 genes compared to PPIA. Primers are listed under ‘qPCR’ in Supporting Information Table 1.

Chromatin Immunoprecipitation (ChIP)

Human polarized Th2 cells were restimulated for 2 hrs, and cells were harvested for ChIP as previously described [6]. Sheered chromatin was immunoprecipitated, and DNA was eluted and purified using the QIAgen minElute PCR Kit. Samples were compared relative to its input DNA using the following formula: % ChIP Efficiency = 2^{(Input Ct − ChIP Ct)} × dilution factor × 100. Primers are listed in Supporting Information Table 1.

EMSA

EMSA was performed as described previously [6]. Clarified nuclear lysates from human PBMCs were incubated at room temperature with 3’-biotin-labeled dsDNA probe in the presence of STAT1, STAT2, STAT3 or STAT4 antibodies, then resolved on nondenaturing gels. Complexes were transferred to a Hybond-N⁺ membrane (Amersham Biosciences) and detected using the Chemiluminescent Nucleic Acid Detection Module (Thermo Scientific). Probe sequences are listed in Supporting Information Table 1.

Cloning, Transfection, and Reporter Assay

The human IL-5 promoter (312 bp) was amplified from a BAC containing the human IL5 gene (clone 729C24, Empire Genomics). The human IL5 promoter was cloned into pGL3-Basic plasmid (Promega). The MSCV2.2-GFP plasmids containing human GATA3 or murine STAT4 and STAT1 have been described previously [15]. Jurkat cells expressing the SV40 Large T-Antigen were electroporated (220 V and 960 µF) with 10 – 12 µg hIL-5pro-Fluc, 7 µg GATA3-GFP, 7 µg STAT1-GFP, STAT4-GFP, or STAT4-R599K-GFP, and 4 µg CMV-renilla (total 28 – 30 µg DNA). Cells were stimulated with anti-hCD3 (M305.2) and rhIL-2 +/− IFN-α for 6 hrs. The Dual-Glo Luciferase Reporter Assay was used to quantify firefly luciferase relative to renilla (Promega).

Statistical analysis

All data are shown as mean ± SEM. Statistical analysis was performed as described in the figure legends with GraphPad Prism software. All p values ≤ 0.05 were considered significant.