B cell lymphoma-6 (Bcl-6) is a transcriptional repressor that is required for the differentiation of T follicular helper (TFH) cell populations. Currently, the molecular mechanisms underlying the transcriptional regulation of Bcl-6 expression are unclear. In this study, we have identified the Ikaros zinc finger transcription factors Aiolos and Ikaros as novel regulators of Bcl-6. We found that increased expression of Bcl-6 in CD4+ Th cell populations correlated with enhanced enrichment of Aiolos and Ikaros at the Bcl6 promoter. Furthermore, overexpression of Aiolos or Ikaros, but not the related family member Eos, was sufficient to induce Bcl6 promoter activity. Intriguingly, STAT3, a known Bcl-6 transcriptional regulator, physically interacted with Aiolos to form a transcription factor complex capable of inducing the expression of Bcl6 and the TFH-associated cytokine receptor Il6ra. Importantly, in vivo studies revealed that the expression of Aiolos was elevated in Ag-specific TFH cells compared with that observed in non-TFH effector Th cells generated in response to influenza infection. Collectively, these data describe a novel regulatory mechanism through which STAT3 and the Ikaros zinc finger transcription factors Aiolos and Ikaros cooperate to regulate Bcl-6 expression.
CD4+ Th cells are responsible for coordinating a wide array of immune responses. Upon activation, naive CD4+ T cells differentiate into specific Th cell subtypes that are critical for coordinating individual activities as part of a pathogen-specific immune response. These include TH1, TH2, TH17, TH9, TH22, and T follicular helper (TFH) cell populations (1–4). The armamentarium provided by these subsets is diverse, ranging from the TH1-mediated secretion of proinflammatory cytokines, such as IFN-γ, to the critical role of TFH cells in promoting the generation of pathogen-neutralizing Abs by B cells. This level of CD4+ T cell subtype specialization depends upon unique lineage-defining transcription factors that direct Th cell development by activating cell-specific gene expression programs and repressing alternative Th cell fates (5–8).
One such example is the transcriptional repressor B cell lymphoma-6 (Bcl-6). Bcl-6 is a member of the broad-complex, tramtrack and bric-à-brac–zinc finger family of proteins and has been identified as a lineage-defining transcription factor required for TFH cell differentiation and the formation of germinal centers (9–13). Additionally, Bcl-6 is important for numerous aspects of B cell development and function, as well as the differentiation of CD4+ and CD8+ memory T cell populations (5, 14–16). A conserved role for Bcl-6 in the generation of these populations is to repress the expression of a second repressor, B lymphocyte–induced maturation protein-1, a direct antagonist of TFH cell– and memory cell–associated genes (5). Other Bcl-6 target genes include those that encode the TH1 and TH2 cell lineage-defining transcription factors T-bet and Gata3, as well as genes associated with cell cycle and metabolic regulation (10–12, 14, 17). Thus, through its ability to modulate a litany of developmental and regulatory pathways, Bcl-6 has emerged as a key driver of immune cell differentiation and function.
As with other transcriptional regulators, the expression and activity of Bcl-6 are regulated by cell-intrinsic signaling cascades that are initiated by extracellular cytokine signals. For example, it is recognized that the cytokines IL-6, IL-12, and IL-21 promote Bcl-6 expression in CD4+ T cells (18–24). In contrast, signaling cascades initiated downstream of IL-2 and IL-7 negatively regulate Bcl-6 (25–30). The differential effects of these cytokines are propagated through the activation of specific STAT factors known to associate with regulatory regions within the Bcl6 gene locus. Specifically, STAT1, STAT3, and STAT4 have been shown to positively regulate Bcl-6 expression, whereas STAT5 is a demonstrated repressor of Bcl-6 (21, 31). Beyond STAT factors, additional transcriptional regulators, including Batf and Tcf-1, have been shown to induce Bcl-6 expression (32–35). Despite these important insights, many questions remain regarding the identity of the transcriptional network that regulates Bcl-6 expression in CD4+ T cell populations.
Similar to STAT factors, the five members of the Ikaros zinc finger (IkZF) family of transcription factors have been implicated in the differentiation of numerous immune cell types, including Th cell subsets (36–39). In the current study, we found that the expression patterns of two IkZF factors, Aiolos and Ikaros, correlated with the expression of Bcl-6 in in vitro–generated TFH-like and in vivo–generated TFH cell populations. Mechanistically, we found that Aiolos and Ikaros were enriched at the Bcl6 promoter and that their association was coincident with chromatin remodeling events, consistent with gene activation, including increased histone acetylation and histone 3 lysine 4 trimethylation (H3K4Me3). Surprisingly, we found that Aiolos physically interacted with the known Bcl-6 activator STAT3 to form a novel transcription factor complex capable of inducing Bcl-6 expression. Importantly, we found that Aiolos expression was elevated in Ag-specific TFH cells compared with non-TFH effector cells that are generated in response to influenza infection. Collectively, our findings identify Aiolos as a novel regulator of Bcl-6 expression and uncover an unexpected cooperative relationship between IkZF and STAT transcription factors that may be an important regulatory feature in the specification of Th cell–differentiation programs.
Materials and Methods
Primary cells, cell culture, and nucleofection
Naive CD4+ T cells were isolated from the spleens and lymph nodes of 5–8-wk-old age- and sex-matched C57BL/6 mice using the MagCellect CD4+ T cell isolation kit (R&D Systems), per the manufacturer’s instructions. Cells were plated at a density of ∼5.0 × 105 cells per well in complete IMDM (IMDM [Life Technologies], 10% FBS [catalog number (cat. no.) 26140079, Life Technologies], 1% Penicillin-Streptomycin [Life Technologies], and 0.05% 2-ME [Sigma-Aldrich]) and stimulated using plate-bound anti-CD3ε (5 μg/ml) and anti-CD28 (10 μg/ml; both from BD Biosciences) in the presence of TH1-polarizing conditions: 5 ng/ml IL-12 (R&D Systems), 5 μg/ml anti–IL-4 (11B11; BioLegend), and 20 ng/ml IL-2 (PeproTech). After 72 h, cells were removed from stimulation and expanded to plate at 5 × 105 cells per well in TH1-polarizing conditions containing high IL-2 (20 ng/ml) or low IL-2 (0.8 ng/ml) for an additional 2 d to generate TH1 or TFH-like cell populations, respectively (Supplemental Fig. 1A) (28, 30). To generate TFH-like cells as described by Awe et al. (40), CD4+ T cells were cultured as previously described. Briefly, naive CD4+ T cells were isolated as described above and stimulated on plate-bound anti-CD3ε (5 μg/ml) and anti-CD28 (10 μg/ml) for 72 h in complete IMDM in the presence of TFH-polarizing conditions: 10 μg/ml anti–IFN-γ (XMG1.2; BioLegend), 10 μg/ml anti–TGF-β (1D11; Bio X Cell), 10 μg/ml anti–IL-2 (JES-1A12; eBioscience), 10 μg/ml anti–IL-4, 100 ng/ml recombinant mouse (rm)IL-6 (R&D Systems), and 50 ng/ml rmIL-21 (PeproTech). Subsequently, cells were expanded and placed in fresh media containing 0.8 ng/ml IL-2, 100 ng/ml rmIL-6, and 50 ng/ml rmIL-21 for an additional 48 h. TH2 cells were generated under the following polarizing conditions: 10 ng/ml IL-4, 20 ng/ml IL-2, and 10 μg/ml anti–IFN-γ). All studies involving mice were conducted in accordance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of Virginia Tech and the University of Alabama at Birmingham.
The murine EL4 T cell line (TIB-39; American Type Culture Collection) was cultured as previously described (28, 41). EL4 cell nucleofection was performed using the Lonza 4D Nucleofection system (Program CM-120; Buffer SF). Overexpression of proteins was confirmed via immunoblot, and endogenous gene expression changes in response to overexpressed proteins were assessed using quantitative RT-PCR (qRT-PCR) analysis. Expression vectors were generated by cloning the coding sequence of genes of interest into the pcDNA3.1/V5-His-TOPO vector (cat. no. K4800; Life Technologies). Sequences were confirmed by sequencing with T7 and BGH primers, followed by transfer of the coding sequence into the pEF1/V5-His vector (cat. no. V920; Life Technologies). The constitutively active STAT3 (STAT3CA) expression vector was generated using the methods above in conjunction with the Agilent QuikChange Site-Directed Mutagenesis Kit (part number 200519), as previously described (42). Expression of each protein was confirmed via immunoblot using V5- and protein-specific Abs.
RNA purification and qRT-PCR
RNA was isolated using the NucleoSpin RNA Kit (MACHEREY-NAGEL), and cDNA was generated using the SuperScript IV First-Strand Synthesis System (Life Technologies), according to the manufacturers’ instructions. cDNA was used at a concentration of 20 ng per qRT-PCR reaction with gene-specific primers (Rps18 forward: 5′-GGA GAA CTC ACG GAG GAT GAG-3′, Rps18 reverse: 5′-CGC AGC TTG TTG TCT AGA CCG-3′; Bcl6 forward: 5′-CCA ACC TGA AGA CCC ACA CTC-3′, Bcl6 reverse: 5′-GCG CAG ATG GCT CTT CAG AGT C-3′; Ikzf1 forward: 5′-ACG CAC TCC GTT GGT AAG CCT C-3′, Ikzf1 reverse: 5′-TGC ACA GGT CTT CTG CCA TCT CG-3′; Ikzf2 forward: 5′-ACG CTC TCA CAG GAC ACC TCA G-3′, Ikzf2 reverse: 5′-GGC AGC CTC CAT GCT GAC ATT C-3′; Ikzf3 forward: 5′-GCT GCA AGT GTG GAG GCA AGA C-3′, Ikzf3 reverse: 5′-GTT GGC ATC GAA GCA GTG CCG-3′; Ikzf4 forward: 5′-GAC GCA CTC ACT GGC CAC CTC C-3′, Ikzf4 reverse: 5′-GGC ACC TCT CCT TGT GCT CCT CC-3′; Ikzf5 forward: 5′-TCG GTA CTG CAA CTA TGC CAG C-3′, Ikzf5 reverse: 5′-AGG TGG CGC TCG TAA GCA GAT G-3′; Il6ra forward: 5′-CCA CAT AGT GTC ACT GTG CG-3′, Il6ra reverse: 5′-GGT ATC GAA GCT GGA ACT GC-3′; and Il2ra forward: 5′- CCA CAA CAG ACA TGC AGA AGC C -3′, Il2ra reverse: 5′-GCA GGA CCT CTC TGT AGA GCC TTG-3′) and SYBR Select Master Mix for CFX (Life Technologies). All samples were normalized to Rps18 as a control and are represented relative to Rps18 expression or relative to the indicated control sample.
Small interfering RNA nucleofection
Day-5 primary murine TFH-like cells were nucleofected with siGENOME SMARTpool small interfering RNA (siRNA) (D-051247, D-064214; Dharmacon) targeting Ikzf1, Ikzf3, or both, using the Lonza 4D-Nucleofector system and buffer P3, per the manufacturer’s instructions. siGENOME nontargeting siRNA was used as a control (D-001210-01; Dharmacon). Following nucleofection, cells recovered in TH1-polarizing conditions containing low IL-2 (TFH-like–polarizing conditions) for 48 h. RNA was isolated, and changes in gene expression were analyzed via qRT-PCR, including Ikzf1 and Ikzf3 to establish knockdown efficiency.
Immunoblot analysis of endogenous and overexpressed proteins was performed using standard procedures. Briefly, cell pellets were lysed in 1× SDS-PAGE buffer (50 mM Tris [pH 6.8], 100 mM DTT, 2% SDS, 0.1% bromophenol blue, 10% glycerol) and immediately boiled for 15 min. Separation of lysates from an equivalent number of cells by SDS-PAGE was followed by immunoblot analysis on a 0.45-μm nitrocellulose membrane, which had been blocked using 2% instant nonfat dry milk in TBS-T (10 mM Tris [pH 8], 150 mM NaCl, 0.05% Tween-20). Abs used to detect proteins of interest were as follows: Aiolos (1:500; cat. no. 39657; Active Motif; 1:5000; cat. no. sc-18683X; Santa Cruz Biotechnology), Ikaros (1:5,000; cat. no. sc-13039X; Santa Cruz Biotechnology), Eos (1:1,000; cat. no. ABE1331; Millipore), Bcl-6 (1:500; cat. no. 561520; BD Pharmingen), and V5 (1:20,000; code R960; Invitrogen). β-actin (1:15,000; cat. no. A00730; GenScript) expression was used as a control to ensure equivalent protein loading.
A Bcl6 promoter-reporter construct (pGL3-Bcl6) was generated by cloning the regulatory region of Bcl6 (positions −1573 to 0 bp) into the pGL3-Basic vector (Promega). EL4 T cells were nucleofected with expression vectors for Ikaros, Aiolos, Eos, or the indicated mutants, in conjunction with pGL3-Bcl6 and a SV40-Renilla vector as a control for transfection efficiency. Following 20–24 h of recovery, samples were harvested, and luciferase expression was analyzed using the Dual-Luciferase Reporter system, according to the manufacturer’s instructions (Promega). Abundance of overexpressed proteins was assessed via immunoblot.
Chromatin immunoprecipitation assay
Chromatin immunoprecipitation (ChIP) assays were performed as published (28, 30). In brief, chromatin was harvested from TH1 and TFH-like cells and immunoprecipitated with Abs to Aiolos (cat. no. sc-18683X; 5 μg i.p.), Ikaros (cat. no. sc-9859X; 5 μg i.p.), STAT3 (cat. no. sc-482X; 5 μg i.p.; all from Santa Cruz Biotechnology), H3K4Me3 (cat. no. 39160, 1 μl i.p.; Active Motif), histone 4 acetylation (H4Ac; cat. no. 39926, 1 μl i.p.; Active Motif), histone 3 lysine 27 acetylation (H3K27Ac; cat. no. 07-360; 1 μl i.p.; Millipore), histone 3 lysine 9 acetylation (H3K9Ac; cat. no. 06-942; 1 μl i.p.; Millipore), or IgG control (cat. no. ab6709; 5 μg i.p.; Abcam). Precipitated DNA was analyzed by quantitative PCR with gene-specific primers (Bcl6 “A” forward: 5′-GTA CTC CAA CAA CAG CAC AGC-3′, and reverse, 5′-GTG GCT CGT TAA ATC ACA GAG G-3′; Bcl6 “B” forward: 5′-CGA CCT TGA AAC GAA CCC AG-3′, and reverse: 5′-GTG TGG GTA CGT GTA ATG TTT GCC-3′; Bcl6 “C” forward: 5′-CGA GTT TAT GGG TAG GAG AGG-3′, and reverse: 5′-GTC TTC GCT GTA GCA AAG CTC G-3′; Bcl6 “D” forward: 5′-GCG GAG CAA TGG TAA AGC CC-3′, and reverse: 5′-CTG GTG TCC GGC CTT TCC TAG-3′; and Il2 forward: 5′-CTG CCA CAC AGG TAG ACT C-3′, and reverse: 5′-GGT CAC TGT GAG GAG TGA TTA GC-3′). Samples were normalized to a standardized total input DNA control, followed by subtraction of the IgG Ab as a control for the nonspecific binding. The final value represents the percentage enrichment of Aiolos, Ikaros, STAT3, H3K4Me3, H4Ac, H3K27Ac, and H3K9Ac.
Coimmunoprecipitation (Co-IP) assays were performed as previously described (17, 30). Briefly, lysates from primary murine TFH-like cells or EL4 cells overexpressing the indicated proteins were immunoprecipitated with an experiment-specific Ab or Ab control. Lysates were incubated at 4°C in the presence of Protein A or Protein G Sepharose beads (cat nos. 16157 and 16201; Millipore) for 1–2 h. Immunoprecipitated proteins were detected by subsequent immunoblot analysis. The following Abs were used for immunoprecipitation at 5 μg i.p., for overexpression and primary T cell Co-IP analyses: STAT3 (cat. no. sc-482X) and Ikaros (cat. no. sc-9859X; both from Santa Cruz Biotechnology). Abs used to detect immunoprecipitated proteins were as follows: Aiolos (cat. no. 39293; Active Motif), Aiolos (cat. no. sc-18683X), STAT3 (cat. no. sc-8019), Ikaros (cat. no. sc-13039X; all from Santa Cruz Biotechnology), and V5 (cat. no. R960-25; Invitrogen).
Influenza virus infections and in vivo analysis
Influenza virus infections were performed intranasally with 6500 VFU A/PR8/34 in 100 μl of PBS. Cell suspensions from mediastinal lymph nodes (mLNs) were prepared by passing tissues through nylon mesh. Cells from mLNs were resuspended in 150 mM NH4Cl, 10 mM KHCO3, and 0.1 mM EDTA for 5 min to lyse red cells. Cell suspensions were then filtered through a 70-μm nylon cell strainer (BD Biosciences), washed, and resuspended in PBS with 5% donor calf serum and 10 μg/ml Fc Block (2.4G2; Bio X Cell) for 10 min on ice before staining with fluorochrome-conjugated Abs or tetramer reagents. Fluorochrome-labeled anti–Bcl-6 (clone K112-91; dilution 1:50), anti-Cxcr5 (clone 2G-8; dilution 1:50), anti-Aiolos (clone S48-791; dilution 1:50), and anti-CD4 (clone RM4-5; dilution 1:200) were from BD Biosciences. The IAbNP311–325 MHC class II tetramer (dilution 1:100) was obtained from the National Institutes of Health Tetramer Core Facility. Intracellular staining for Bcl-6 and Aiolos was performed using the Mouse Regulatory T Cell Staining Kit (eBioscience), according to the manufacturer’s instructions. Flow cytometry was performed using an Attune NxT Flow Cytometer (Thermo Fisher Scientific), and data were analyzed using FlowJo software.
Listeria monocytogenes infections and analysis
All studies used 6- to 10-wk-old age- and sex-matched C57B6/J mice that were maintained in specific pathogen–free facilities. Listeria monocytogenes was obtained from the American Type Culture Collection (ATCC BAA-679) and cultured in brain heart infusion agar/broth, following product guidelines. Prior to in vivo inoculation, bacteria density was estimated in liquid broth culture by measuring the absorbance at 600 nm (0.4 OD600 = 7.9 × 108 bacteria/ml). Cultures of L. monocytogenes were pelleted, washed twice in PBS, and resuspended in PBS. Mice were inoculated via i.v. injection with 5000 CFU L. monocytogenes in 50 μl of PBS. L. monocytogenes inoculation dosages were confirmed for each experiment by plating an aliquot of homogenized liver from randomly selected mice 24 h after inoculation and calculating the resultant CFU/mg of tissue.
At 6 d after infection, primary CD4+ T cells were isolated from the spleens of L. monocytogenes–infected mice using the BioLegend MojoSort Isolation Kit (cat. no. 480033), per the manufacturer’s instructions. For subsequent flow cytometry analysis, isolated cells were stained with the following fluorochrome-labeled Abs: anti-CD44 (cat. no. 560452; BD Biosciences), anti–PD-1 (cat. no. 11-985; eBioscience), anti-Cxcr5 (cat. no. FAB6198P; R&D Systems), anti-CD4 (cat. no. 56-0031; eBioscience), and the respective isotype controls (R&D). Briefly, harvested cells were pelleted, washed, and stained with the indicated fluorochrome-conjugated Abs. Following staining, cells were washed three times with FACS buffer and then resuspended for analysis and sorting on a Sony SH800 flow cytometer. All of the obtained data were analyzed using FlowJo software. CD4+CD44+Cxcr5hiPD1+ TFH cell transcript analysis was conducted as described above in RNA purification and qRT-PCR.
All data represent at least three independent experiments. Error bars represent the SEM or SD, as indicated. For statistical analyses, an unpaired t test or one-way ANOVA with the Tukey multiple-comparison test were performed to assess statistical significance, as appropriate for a given experiment. The p values < 0.05 were considered statistically significant.
Expression of Aiolos and Ikaros correlate with that of Bcl-6
Although it was once widely believed that initial commitment to an individual effector T cell subset was a terminal event, a large body of literature has emerged demonstrating that a substantial degree of flexibility exists between effector Th cell populations (43–47). It has been postulated that this phenomenon, termed Th cell plasticity, allows Th cell subsets to respond to the cellular microenvironment in real-time to provide a more efficient and effective immune response. For example, many studies have demonstrated that TH1 cells maintain flexibility with the TFH cell type (19, 30, 48). More recently, there have also been reports of TH1-biased TFH cell populations (49, 50). Corroborating these studies, our laboratory has previously demonstrated that TH1 cells are capable of upregulating a Bcl-6–dependent TFH-like cell state in response to altered IL-2 signaling (28, 30). Mechanistically, in TH1 cells exposed to low IL-2 environments, increased Bcl-6 expression allows for the repression of the TFH antagonist B lymphocyte–induced maturation protein-1 and the subsequent expression of a TFH-like cell program (Fig. 1A) (28, 30). Because the induction of Bcl-6 expression is a critical step for the transition from the TH1 to the TFH-like state, our current study aimed to identify the IL-2–sensitive transcription factors that are responsible for regulating Bcl-6 expression.
Members of the IkZF transcription factor family have previously been implicated in the differentiation of specific effector Th cell subsets, including TH1, TH17, and regulatory T cell populations (37, 51–54). As such, we began by assessing whether these factors were differentially expressed in in vitro–generated TH1 versus TFH-like cells (Supplemental Fig. 1A). We found that two IkZF factors, Ikaros (Ikzf1) and Aiolos (Ikzf3), were upregulated in TFH-like cells, whereas a third IkZF family member, Eos (Ikzf4), was more highly expressed in TH1 cells (Fig. 1B). Although two additional IkZF family members, Helios (Ikzf2) and Pegasus (Ikzf5), also displayed altered expression, their overall transcript levels were much lower than those of the other IkZF factors (Fig. 1B). A time course encompassing the naive CD4+ T cell transition to differentiated TH1 or TFH-like cell populations further demonstrated that increased Aiolos expression correlated with the highest expression of Bcl-6, whereas Eos expression was elevated in day-3 and day-5 TH1 cells (Fig. 1C, Supplemental Fig. 1A). Although there was a slight increase in Ikaros expression in TFH-like cells, the expression was relatively abundant across CD4+ T cell populations, perhaps implying a broader role for this factor in Th cell differentiation (Fig. 1C). In addition to our in vitro–generated TFH-like cells, other laboratories have generated TFH-like cells using alternative in vitro culturing conditions (40, 48). Importantly, analysis of Ikaros and Aiolos expression revealed a similar trend among the TFH-like populations compared with that observed for in vitro–generated TH1 or TH2 cells (Supplemental Fig. 1B). Furthermore, similar expression patterns were observed in TFH cells generated in vivo in response to L. monocytogenes infection (Supplemental Fig. 1C). Collectively, these gene-expression analyses indicated that a positive correlation exists among Aiolos (Ikzf3), Ikaros (Ikzf1), and Bcl6 expression.
To determine whether protein expression correlated with the observed changes in transcript, we performed immunoblot analyses of Aiolos, Ikaros, Eos, and Bcl-6 expression in TH1 and TFH-like cells. Indeed, we observed a sharp increase in Aiolos expression that was consistent with increased expression of Bcl-6 in TFH-like cells (Fig. 1D). As with the transcript data, Ikaros expression was moderately elevated in TFH-like cells, whereas Eos protein expression inversely correlated with that of Bcl-6 and the other two IkZF family members (Fig. 1D–F). Given the positive correlation among Aiolos, Ikaros, and Bcl-6, we focused on elucidating possible mechanisms by which these IkZF factors may contribute to the induction of Bcl-6 expression.
Knockdown of Aiolos or Ikaros results in reduced Bcl6 expression
To further assess whether Aiolos or Ikaros may be involved in promoting Bcl-6 expression, we used an siRNA-knockdown approach to determine whether reduced expression of either factor affected Bcl-6 expression in TFH-like cells. Indeed, upon knockdown of Aiolos (Ikzf3) or Ikaros (Ikzf1) individually, we observed decreased expression of Bcl6 transcript, whereas combined knockdown of Ikzf1 and Ikzf3 resulted in a further reduction in Bcl6 expression (Fig. 1G). Importantly, Eos (Ikzf4) levels remained unchanged, demonstrating the specificity of the Aiolos- and Ikaros-dependent effects upon Bcl-6 expression. These data also demonstrate the specificity of the siRNAs for the intended Ikzf1 and Ikzf3 targets, an important consideration given the high degree of conservation among IkZF family members (36). Collectively, these data support a role for Aiolos and Ikaros in the positive regulation of Bcl-6 expression.
The ZF DNA-binding domains of Aiolos and Ikaros are required to induce Bcl6 promoter activity
To determine whether Aiolos and Ikaros may interact with regulatory regions of the Bcl6 locus, we performed an in silico analysis of the Bcl6 promoter and located several predicted binding sites containing the core IkZF DNA-binding motif “GGGAA” (Fig. 2A). To test the functionality of these sites, we created a Bcl6 promoter-reporter construct encompassing the predicted sites to examine the effect of Aiolos, Ikaros, or Eos overexpression on Bcl6 promoter activity. Importantly, upon Aiolos or Ikaros overexpression, we observed a significant increase in Bcl6 promoter activity (Fig. 2A). Conversely, as a control, there was no increase in Bcl6 promoter activity in the presence of Eos. These data suggest that Aiolos and Ikaros may induce Bcl-6 expression via effects upon the Bcl6 promoter.
Aiolos and Ikaros contain N-terminal zinc finger (ZF) domains, which mediate their DNA-binding capabilities. Of the four ZFs that make up the N-terminal domain, ZF2 and ZF3 are required for DNA binding, whereas ZF1 and ZF4 appear to modulate sequence specificity (55, 56). To determine whether ZF-mediated DNA binding was required for Aiolos- and Ikaros-dependent Bcl6 promoter activation, we constructed expression vectors with point mutations in select N-terminal ZFs (Aiolos: AiΔZF1, AiΔZF1,2, and AiΔZF4; Ikaros: IkΔZF1, IkΔZF1,2, and IkΔZF4) (Fig. 2B, 2C). We then compared the ability of wild-type Aiolos or Ikaros and their corresponding ZF mutants to induce Bcl6 promoter activity. As with our previous data, Bcl6 promoter activity was readily induced by wild-type Aiolos or Ikaros (Fig. 2B, 2C). However, we did not observe increases in Bcl6 promoter activity with the AiΔZF1,2 or IkΔZF1,2 mutants, suggesting that direct DNA binding by Aiolos or Ikaros may be required for Bcl-6 induction (Fig. 2B, 2C). Interestingly, overexpression of a subset of the Aiolos and Ikaros mutants with a single ZF mutation in either the first or fourth ZF (AiΔZF1, IkΔZF1, and IkΔZF4) resulted in only a modest increase in Bcl6 promoter activity, suggesting that there may be differential requirements for individual ZFs in mediating promoter activation. Taken together, these results suggest that the N-terminal ZF DNA-binding domain is of functional importance in the Aiolos- and Ikaros-dependent induction of Bcl6 promoter activity.
Aiolos and Ikaros associate with the Bcl6 promoter
Because our previous data suggested that Aiolos and Ikaros may induce the activation of Bcl-6 expression and that direct DNA binding may be required, we performed ChIP assays to determine whether either factor directly associates with the endogenous Bcl6 promoter in TFH-like cells (Fig. 3A). Indeed, we observed association of Aiolos and Ikaros with a region of the Bcl6 promoter proximal to the transcriptional start site (TSS) (Fig. 3B, 3C). Importantly, this association was significantly enriched in TFH-like cells compared with that observed for TH1 cells (Fig. 3B, 3C). Furthermore, the binding observed proximal to the TSS was specific, because enrichment of these factors was markedly reduced at a distal location from the Bcl6 locus. Interestingly, further ChIP analysis for the known Bcl-6 regulator STAT3 revealed a similar enrichment pattern to that observed for Aiolos and Ikaros in TFH-like cells, suggesting the intriguing possibility that these factors may collaboratively regulate Bcl-6 expression (Fig. 3D).
Quintana et al. (37) recently demonstrated that Aiolos directly represses IL-2 production during TH17 cell differentiation. Because IL-2 signaling negatively regulates Bcl-6 expression, we considered the possibility that Aiolos could contribute to increased levels of Bcl-6 indirectly in a similar IL-2–dependent fashion. To address this, we examined Aiolos association with the IL-2 (Il2) locus, as described by Quintana et al. (37). Interestingly, although there was a slight increase in Aiolos binding at the Il2 locus in TFH-like cells, overall enrichment was much less than that observed at the Bcl6 locus, suggesting that the effect of Aiolos on Bcl6 expression is mediated by a direct mechanism rather than indirectly via an IL-2–dependent effect (Fig. 3B–D). Collectively, these data support a model in which Aiolos, Ikaros, and STAT3 associate with the Bcl6 promoter to induce Bcl-6 expression in TFH-like cells.
IkZF/STAT factor association correlates with increased histone modification of the Bcl6 promoter
IkZF and STAT factors are known to exert their effects, at least in part, through their association with chromatin-remodeling complexes that are capable of directing histone modifications indicative of gene activation and repression. To assess whether Aiolos and Ikaros binding was associated with changes to chromatin structure at the Bcl6 promoter, we performed a ChIP assay to examine alterations in histone acetylation and methylation. Consistent with a role in activating Bcl-6 expression, we observed increased H3K4Me3, H3K9Ac, H3K27Ac, and H4Ac at the Bcl6 promoter in TFH-like cells compared with TH1 cells (Fig. 3E–H). Importantly, significant increases in H3K4Me3, H3K9Ac, and H3K27Ac were detected at regions where Aiolos, Ikaros, and STAT3 were most highly enriched (Fig. 3E–G). Collectively, these data indicate that association of Aiolos, Ikaros, and STAT3 with the Bcl6 promoter correlates with changes in chromatin structure consistent with gene activation.
Aiolos interacts with STAT3 in TFH-like cells
Given the similar enrichment patterns observed for Aiolos, Ikaros, and STAT3 at the Bcl6 promoter, we considered the possibility that they may cooperate to induce Bcl6 expression. To this end, we performed Co-IP assays to determine whether these factors interact in TFH-like cells. It is well established that Aiolos and Ikaros form heterodimers via their C-terminal ZF domains (36, 56). Indeed, Co-IP analysis demonstrated that Aiolos and Ikaros form heterodimeric complexes in TFH-like cells (Fig. 4A). Intriguingly, additional Co-IP assays revealed the presence of novel Aiolos/STAT3 complexes in TFH-like cells (Fig. 4B). Despite the large degree of amino acid conservation between Aiolos and Ikaros, we did not detect the presence of Ikaros/STAT3 complexes in TFH-like cells. Whether this was due to the absence of such complexes or was a limitation of the Co-IP analysis itself remains unresolved. Still, these findings support the existence of a novel Aiolos/STAT3 protein complex in TFH-like cells.
To examine whether Aiolos and STAT3 may cooperate to induce Bcl-6 expression, we overexpressed STAT3CA alone, Aiolos alone, or Aiolos in combination with increasing amounts of STAT3CA (Fig. 4C). Interestingly, expression of Aiolos or STAT3CA alone resulted in only slight increases in the Bcl6 transcript. However, when Aiolos was expressed in combination with increasing amounts of STAT3CA, we observed significant increases in Bcl6 expression compared with the control sample or the samples in which Aiolos or STAT3CA were expressed individually (Fig. 4C).
N- and C-terminal ZF domains of Aiolos are required for induction of Bcl-6
In addition to the N-terminal ZF DNA-binding domain discussed previously, members of the IkZF family contain a C-terminal ZF domain that mediates homo- or heterodimerization with other IkZF proteins (36, 56). To determine whether either ZF domain may be required for interaction between Aiolos and STAT3, we coexpressed STAT3CA with wild-type Aiolos or with Aiolos mutants harboring disruptions to the N- or C-terminal ZF domains and performed Co-IP analysis (Fig. 5A). As with the TFH-like cells, wild-type Aiolos and STAT3 interactions were readily detected. Similarly, we also detected interactions between STAT3 and the AiΔZF1,2 mutant. However, STAT3 was unable to interact with the Aiolos mutant lacking the C-terminal ZF dimerization domain (AiolosΔC), suggesting that this domain is required for the interaction between Aiolos and STAT3 (Fig. 5A). To determine the functional impact of the C-terminal mutation, we overexpressed STAT3CA with wild-type Aiolos or AiolosΔC and assessed the impact on Bcl6 expression. Indeed, compared with coexpression of wild-type Aiolos and STAT3CA, Bcl6 expression was significantly diminished when STAT3CA was coexpressed with the AiolosΔC mutant (Fig. 5B).
Although the AiΔZF1,2 mutant was able to interact with STAT3, the Bcl6-reporter data suggested that the N-terminal ZF DNA-binding domain was required to induce promoter activity. To determine the functional impact of the N-terminal mutation, we overexpressed STAT3CA with wild-type Aiolos or AiolosΔZF1,2 and assessed the impact on Bcl6 expression. Indeed, similar to the results observed for the AiolosΔC mutant, the combination of STAT3CA and AiolosΔZF1,2 was unable to induce Bcl-6 expression, suggesting that both the N- and C-terminal ZF domains are required for Aiolos-dependent activation of Bcl-6 expression (Fig. 5C).
Taken together, these data suggest that STAT3 and Aiolos form a transcription factor complex via the Aiolos C-terminal protein–protein interaction domain and that this novel complex is capable of promoting Bcl-6 expression. It is important to note that, although interactions between Ikaros and STAT3 were not detected, our data do not preclude the possibility that such interactions, or interactions between some combination of STAT3, Aiolos, and Ikaros, may play important roles in regulating Bcl-6 expression.
Aiolos and STAT3 cooperatively regulate cytokine receptor expression
Our findings suggest that Aiolos and Ikaros are direct regulators of Bcl-6 expression in Th cells. As discussed previously, signals from environmental cytokines are key determinants of Th cell differentiation. These include IL-6 and IL-2, which have been shown to positively and negatively influence TFH development, respectively (21, 57, 58). Therefore, we wanted to assess whether Aiolos or Ikaros may play a broader role in promoting TFH cell differentiation by regulating the expression of the receptors for these cytokines. We began by performing Aiolos (Ikzf3) and Ikaros (Ikzf1) siRNA-knockdown experiments to assess the effect on Il6ra and Il2ra expression. Indeed, the expression of Il6ra decreased significantly upon Ikzf3 and Ikzf1 knockdown (Fig. 6A). Importantly, this decrease in expression was specific to Il6ra, because the expression of Il2ra was unaffected (Fig. 6A). To determine whether Aiolos and STAT3 may cooperate to induce Il6ra (as with Bcl-6), we overexpressed STAT3CA alone, Aiolos alone, or Aiolos in combination with increasing amounts of STAT3CA and examined Il6ra expression. Importantly, the coexpression of Aiolos and STAT3CA resulted in a significant increase in Il6ra expression, whereas the expression of Il2ra was unchanged (Fig. 6B). Collectively, these data suggest that the interplay between Aiolos, Ikaros, and STAT3 may play a broader role in regulating TFH differentiation, perhaps through induction of the cytokine receptor Il6ra.
Aiolos expression is increased in Ag-specific TFH cells after influenza infection
Our mechanistic findings implicate Aiolos in the positive regulation of Bcl-6, and possibly the TFH differentiation program, via a cooperative mechanism with the known TFH regulator STAT3. To extend these findings, we sought to determine whether Aiolos was preferentially expressed in in vivo–generated TFH cells, as opposed to non-TFH effector Th (TEFF) cells, in response to infection. To this end, we infected mice with influenza and assessed Aiolos expression in Ag-specific TFH (Bcl-6hiCxcr5hi) and TEFF (Bcl-6loCxcr5lo) populations (Fig. 7A). Importantly, at the peak of infection (12 d after infection), nucleoprotein (NP)-specific TFH cells expressed significantly more Aiolos than that observed in the Bcl-6lo TEFF population (Fig. 7B). Collectively, these in vivo data, in combination with our in vitro findings, are supportive of a role for Aiolos in promoting Bcl-6 expression and TFH cell differentiation.
Bcl-6 has well-established roles in the development of a multitude of immune cell types, including TFH cells, memory T cell populations, and B cells. As such, there has been an ongoing interest in identifying the molecular mechanisms involved in the transcriptional regulation of Bcl-6 expression. In this article, we describe previously unidentified roles for the IkZF family members Aiolos and Ikaros in the induction of Bcl-6. Perhaps most intriguingly, our data have begun to define a novel cooperative relationship between STAT3, a known positive regulator of Bcl-6 expression, and the IkZF factor Aiolos.
The interplay between opposing STAT factors (e.g., STAT3, STAT5) at the Bcl6 promoter is an important contributor to the regulation of Bcl-6 expression (21, 57, 58). Unexpectedly, our findings demonstrate that STAT3 physically interacts with Aiolos. Thus, our data support the possibility that a primary role for STAT3 may be to recruit Aiolos to the Bcl6 locus. Taken together with the increase in Aiolos expression observed in in vitro– and in vivo–derived TFH cell populations, these data suggest that this novel IkZF/STAT protein complex may be an important driver of Bcl-6 expression and, perhaps, TFH cell fate. Indeed, it will be of interest to establish whether the Aiolos/STAT3 complex regulates additional TFH genes beyond Bcl6 and Il6ra. Likewise, given that Aiolos and STAT3 are members of transcription factor families that are both widely expressed and highly conserved, it will be of considerable interest to determine whether STAT and IkZF interactions regulate the differentiation and function of other immune cell populations, including additional Th cell subsets. For example, Aiolos has been shown to influence TH17 differentiation through the direct repression of IL-2 expression (37). Interestingly, TH17 and TFH cells share a number of regulatory features during their development, including sensitivity to the IL-2/STAT5 signaling axis and dependence on STAT3 activity (26, 29, 30, 59, 60). Thus, an intriguing possibility is that the STAT3/Aiolos complex identified in this study may also play a role in TH17 differentiation.
The precise molecular mechanisms by which STAT3, Aiolos, and Ikaros cooperate to regulate Bcl-6 expression remain unclear. The activity of IkZF factors has been attributed primarily to their association with chromatin-modifying enzymes, including the switch/sucrose nonfermenting and Mi-2/nucleosome remodeling and deacetylase complexes (56, 61). Indeed, our results demonstrate that the association of Aiolos and Ikaros with the Bcl6 promoter correlates with alterations to the chromatin structure surrounding this region, including increased histone acetylation and methylation. These chromatin modifications are indicative of an accessible chromatin structure and consistent with an actively transcribing gene. Our observations are not without precedence, because IkZF factors have been implicated in gene activation prior to the current study (36, 56, 62, 63). It is also possible that Aiolos and Ikaros may act to remodel the chromatin structure of regulatory regions located proximal to the Bcl6 promoter. Intriguingly, there are predicted CCCTC-binding factor binding sites surrounding this region, suggesting the presence of insulator or silencing elements. Another, nonmutually exclusive, possibility is that Aiolos and Ikaros association with the Bcl6 promoter near the TSS contributes to the assembly and/or activation of the transcriptional initiation complex. In this regard, Ikaros has been shown to physically interact with, and alter the activity of, the RNA Pol II complex (64). It is also possible that Aiolos and Ikaros may be required to mediate interactions between the Bcl6 promoter and distal enhancers. In support of this possibility, Ikaros has been implicated in promoting long-range chromatin interactions between regulatory elements at other genetic loci (65, 66). Further experimentation will be required to address these possibilities and to understand whether established IkZF-remodeling activities are involved or whether a novel mechanism may be responsible for promoting Bcl-6 expression.
Future studies will also be necessary to comprehensively assess the distinct contributions of Aiolos and Ikaros to the regulation of Bcl-6 expression. Thus far, our data support the existence of STAT3/Aiolos complexes but not those made up of STAT3 and Ikaros, because we could not detect the latter. Still, ChIP analysis of the Bcl6 promoter clearly demonstrates an increase in Aiolos and Ikaros association at the Bcl6 locus, and we detect the presence of Aiolos/Ikaros complexes in TFH-like cells. Therefore, it is possible that Ikaros and Aiolos could be recruited independently of STAT3 to the Bcl6 locus. However, the signals responsible for Aiolos/Ikaros recruitment to the Bcl6 promoter remain unclear. Based upon our observation that Ikaros is expressed at moderately high levels in naive and TH1 cells, we propose a model in which a basal level of Ikaros is bound to the Bcl6 locus, perhaps allowing this gene to remain in a poised state during Th cell differentiation. Indeed, this would be consistent with the established role of Ikaros as a broad regulator of T cell differentiation and our observation that Ikaros is expressed at moderately high levels in naive and TH1 cells prior to the transition to the TFH-like cell state (36, 61). In our proposed model, we hypothesize that the association of Aiolos/STAT3 complexes with the Bcl6 promoter, in the absence of IL-2/STAT5 signaling, leads to chromatin-remodeling activities that result in the activation of Bcl-6 expression. Additionally, because IkZF factors are known to homo- and heterodimerize upon binding to DNA, we also propose that IkZF proteins could mediate interactions between distal regulatory elements (65, 66). Further elucidation of the exact contribution of Aiolos and Ikaros to the induction of Bcl-6 will be of interest and may serve to shed light on how STAT3, Aiolos, and Ikaros cooperate to regulate the expression of additional target genes involved in TFH cell differentiation.
The appropriate temporal expression of Bcl-6 is a molecular linchpin that regulates the differentiation and function of many cell types that are critical to the promotion of an effective immune response. The importance of understanding Bcl-6 regulation is further highlighted when considering the numerous human diseases that have been linked to aberrant expression and function of this transcriptional repressor (14, 67). Our findings identify novel regulators and provide insight into the mechanisms by which they promote Bcl-6. Future work will be required to fully elucidate the complex network of signals and factors that regulate Bcl-6 expression. In doing so, we may enhance the potential to design more efficacious vaccines and develop novel immunotherapeutic approaches as a result of the wide-ranging importance of this transcriptional regulator.
We thank Sheryl Coutermarsh-Ott and Daniel Rothschild for technical assistance with the in vivo studies and Dr. James Smyth and members of the Oestreich laboratory for insightful discussions.
This work was supported by start-up funds from the Virginia Tech Carilion Research Institute, a seed grant from the Virginia-Maryland College of Veterinary Medicine (to I.C.A. and K.J.O.), and a Careers in Immunology Fellowship from the American Association of Immunologists (to M.D.P. and K.J.O.). Support for I.C.A. was provided by Virginia Tech and National Institutes of Health Grant DK105975. Support for A.B.-T. was provided by the University of Alabama Birmingham and National Institutes of Health Grant R01AI110480.
The online version of this article contains supplemental material.
Abbreviations used in this article:
B cell lymphoma-6
- cat. no.
histone 4 acetylation
histone 3 lysine 9 acetylation
histone 3 lysine 27 acetylation
histone 3 lysine 4 trimethylation
Ikaros zinc finger
mediastinal lymph node
small interfering RNA
constitutively active STAT3
non-TFH effector Th
T follicular helper
transcriptional start site
The authors have no financial conflicts of interest.