PLZF Regulates CCR6 and is Critical for the Acquisition and Maintenance of the Th17 Phenotype in Human Cells

Abstract

Th17 cells, which all express the chemokine receptor CCR6, are implicated in many immune-mediated disorders such as psoriasis and multiple sclerosis. We found that expression levels of CCR6 on human effector/memory CD4⁺ T cells reflect a continuum of Th17 differentiation. By evaluating the transcriptome in cells with increasing CCR6, we detected progressive up-regulation of ZBTB16, which encodes the BTB-ZF transcription factor PLZF. Using ChIP for modified histones, p300, and PLZF, we identified enhancer-like sites at −9/−10 kb and −13/−14 kb from the upstream transcription start site of CCR6 that bind PLZF in CCR6⁺ cells. For Th cells from adult blood, both in the CCR6⁺ memory population and in naïve cells activated ex vivo, knockdown of ZBTB16 down-regulated CCR6 and other Th17-associated genes. ZBTB16 and RORC (which encodes the “master regulator” RORγt) cross-regulate each other, and PLZF binds at the RORC promoter in CCR6⁺ cells. In naive Th cells from cord blood, ZBTB16 expression was confined to the CD161⁺ cells, which are the Th17 cell precursors. ZBTB16 was not expressed in mouse Th17 cells and Th17 cells could be made from luxoid mice, which harbor an inactivating mutation in Zbtb16. These studies demonstrate a role for PLZF as an activator of transcription important both for Th17 differentiation and the maintenance of the Th17 phenotype in human cells, expand the role of PLZF as a critical regulator in the human adaptive immune system, and identify a novel, essential element in a regulatory network that is of significant therapeutic interest.

Introduction

Th17 cells contribute to host defense against extracellular bacteria and fungi at mucosal sites, and have been implicated in autoimmune and inflammatory diseases (1, 2). The potential for developing novel therapies has driven the interest in understanding the mechanisms whereby Th17 cells are produced, and how their phenotypes are regulated (1, 3, 4). Both cytokines in the Th17 pathway, such as IL-23, IL-17, and IL-22, and the transcription factors that control Th17 cell differentiation are targets of these therapies (5-7). Data on the regulation of Th17 cell transcription have come primarily from work in mice (3), and the best studied lineage-specific transcription factor is RORγt, the so-called “master regulator” of Th17 differentiation (8).

Studies of human cells revealed that all T cells that are able to produce IL-17A express the chemokine receptor CCR6 (9-12). CCR6 is the sole receptor for CCL20, which is inducible in many cells in response to inflammatory stimuli (13). In mice, CCR6 is critical for formation of components of the gut-associated lymphoid tissue and antibody responses to intestinal pathogens, and important for pathogenesis in models of a number of immune-mediated diseases (14-19). CCR6 is an RORγt-responsive gene in mouse and human CD4⁺ T cells (15, 20), and stable expression of CCR6 in human T cells is associated with promoter de-methylation (21). Little else is known about how CCR6 is regulated and how its regulation is similar to or differs from the regulation of other genes that are part of the Th17 program. In addition to IL-17A-producing cells, virtually all mouse and human Th cells that can produce IL-17F, IL-22, and CCL20 and express RORC and IL23R are found within the CCR6⁺ subset (15, 22), and S.P.S. and J.M.F., unpublished data, and see below) suggesting that CCR6 may be controlled by factors that are shared broadly with the genes that characterize the Th17-phenotype and/or that are important in initiating a regulatory pathway that, as it is further modified and arborizes, gives rise to Th17 cells and associated cell types.

In the work described below, we found that CCR6 and other Th17-associated genes are regulated by the Broad complex, Tramtrack, Bric a brac-zinc finger (BTB-ZF) transcription factor promyelocytic leukemia zinc finger protein PLZF, encoded by the gene ZBTB16. PLZF was discovered as part of a fusion protein with retinoic acid receptor-alpha in a case of acute promyelocytic leukemia (23) and has been recognized both as a tumor suppressor (24), and as a driver of leukemogenesis (25). PLZF, like other BTB-ZF factors, has generally been described as a repressor of transcription through its binding to histone deacetylases, polycomb group proteins, and other co-repressors (24, 26). In normal hematopoiesis, PLZF serves as a switch to balance the maintenance of undifferentiated myeloid progenitors with demands for expanded numbers of mature cells under stress (27), and PLZF has a critical role in establishing the phenotype of NKT cells and innate-like γ/δ T cells as well as in the development of innate lymphoid cells in mice (28-31). Our data demonstrate a new role for ZBTB16/PLZF in cells of the human adaptive immune system, contributing both to Th17 cell differentiation and the maintenance of the Th17 phenotype in effector/memory cells.

Materials and Methods

Human samples

Human umbilical cord blood was obtained from term placentas following the delivery of healthy newborns at Shady Grove Adventist Hospital, Gaithersburg, MD, and provided without identifiers under a protocol approved and designated “exempt” by the hospital's Institutional Review Board. Elutriated lymphocytes were obtained from healthy donors by the Department of Transfusion Medicine, NIH, under an Institutional Review Board-approved protocol.

Mice

C57BL/6J mice were purchased from The Jackson Laboratory and B6.C3-Zbtb16^lu/J (luxoid) were a kind gift from Robert Braun of The Jackson Laboratory. All mice were housed in sterile cages in pathogen free conditions and used at 6-8 weeks of age. Animal protocols were approved by the Animal Care and Use Committee of the NIAID, NIH.

Purification of lymphocyte subsets from blood and flow cytometry

Before purification by cell sorting, CD4⁺ T cells were isolated from approximately 2 × 10⁹ elutriated lymphocytes by negative selection using RosetteSep CD4⁺ T cell enrichment cocktail (StemCell Technologies). Naïve CD4⁺ T cells from cord blood were sorted based on a phenotype of CD4⁺CD45RA⁺CD62L⁺CD25⁻CXCR3⁻CCR4⁻. In some experiments, cells were also purified based on staining with anti-CD161. For cells from adult blood, staining always included anti-CD4 plus anti-CD45RO. In addition, cells were stained with anti-CCR6 or anti-CXCR3 or anti-CCR4 or anti-CXCR5. All cell sorting was done using a FACS Aria flow cytometer (BD Biosciences). By post-sort analysis, the purity of populations was 95-99%.

Isolation of RNA, and real-time fluorogenic RT-PCR

Cells were activated with 20 ng/ml PMA and 1 mM ionomycin for 4 h and total cellular RNA was purified using PureLink™ RNA kit (Life Technology). Reverse transcription and real-time PCR was performed in duplicate (10-50 ng RNA per sample) using SuperScript One Step RT-PCR kit (Life Technology). Primer and probe sets (FAM/MGB-labeled) were purchased from Applied Biosystems. Results were normalized based on the values for GAPDH mRNA, detected using TaqMan GAPDH Control reagents (Applied Biosystems).

Chromatin immunoprecipitation (ChIP) assays

ChIP experiments were performed using the Magna ChIP™ A/G kit from Millipore with antibodies against the modified histones H3K4me2, H3K4me3 or H3K27ac, or against p300 (Abcam), PLZF (Active Motif), or RNA polymerase II (Millipore). For analyzing promoter regions of CCR6 and RORC by ChIP we used custom-made plates with wells containing primers spanning the regions of CCR6 or RORC as noted in the figure legends (SABiosciences). Real-time PCR was performed using the RT² SYBR Green/ROX qPCR master mix (SABiosciences). Primers matching sequences within an intergenic region (human IGX1A primers, SABiosciences) were used as a negative control. Results of ChIP assays are expressed as percent input enrichment, calculated using ChIP PCR array data analysis software from SABiosciences.

Knockdown of ZBTB16 and RORC by siRNAs

SMARTpool control siRNAs and SMARTpool ZBTB16 and RORC siRNAs were obtained from Dharmacon, which was also the source for single ZBTB16 siRNAs that were not present in the SMARTpool. Two million CD4⁺ T cells were transfected with 200-300 pmol of siRNAs for ZBTB16, RORC or non-targeting control alone or in combination using Human T Cell Nucleofector Kit with the amaxa nucleofector (Lonza). In order to check the siRNA transfection efficiency, cells were transfected with siGLO (Dharmacon). Transfection efficiency in three representative experiments ranged from 78-87% (data not shown). Transfected cells were re-suspended in RPMI 1640 medium supplemented with 10% FBS, and 50 units/ml IL-2 and incubated for 72 h before being harvested. Mean viability at the time of harvesting after transfection was 85.57 ± 1.24% for 10 representative samples (data not shown).

Mouse T cell isolation and differentiation in vitro

Naïve T cells from the spleens of wild-type and luxoid mice were isolated as CD4⁺CD25⁻CD62L^hiCD44^lo cells using a FACS Aria flow cytometer. In addition, NKT cells were isolated from spleens of C57BL/6 wild-type mice based on a phenotype of CD3⁺CD8⁻CD24⁺CD44^loNK1.1⁻. One x 10⁶ naïve cells/well were cultured in 24 well plates at 37^oC and 5% CO2 in RPMI 1640 supplemented with 10% heat-inactivated FBS and 50 μM β-mercaptoethenol. Cells were activated with anti-CD3/CD28 coated beads at a beads-to-cell ratio of 1:1 using Dynabeads mouse T-cell activator CD3/CD28 kit (Life Technology), and cultured for 5 days in Th17- or Th1-polarizing conditions as described (32).

Staining for intracellular proteins and flow cytometry

For intracellular staining of PLZF and RORγt, anti-human/mouse PLZF or anti-human/mouse RORγt antibody (eBioscience) was used with the supplier's Foxp3/Transcription Factor Staining Buffer Set. For staining cytokines, cells were stimulated with Leukocyte Activation Cocktail, with GolgiPlus™ (BD Pharmingen) for 6 h at 37°C before being stained with anti-IL17A (eBioscience) or anti-IL-22 or anti-CCL20 (R&D Systems) by using Cytofix/CytoPerm Plus kit (BD Pharmingen). For some experiments, cells were stained with anti-CCR6 for 30 min at room temperature before activation. Other than for cell sorting, all flow cytometry was done using an LSR II System flow cytometer (BD Biosciences), and the data were subsequently analyzed and presented using FlowJO software (TreeStar).

In vitro activation of naïve CD4⁺ T lymphocytes from cord blood and adult peripheral blood

T cells were cultured at 1 × 10⁶ cells/ml in 24 well plates in RPMI 1640 medium supplemented with 10% FBS. Stimulation was done using anti-CD2/CD3/CD28 coated beads (1 bead/cell) from T Cell Activation/Expansion Kit (Miltenyi Biotec) in nonpolarizing conditions, which included IL-2 (200 units/ml), anti-IL-4 (0.4 μg/ml), anti-IL-12 (2 μg/ml), anti-IFN-γ (8 μg/ml), and TGF-β1.2 (10 ng/ml), or in Th17 polarizing conditions, which included IL-6 (10 ng/ml), IL-23 (20 ng/ml), IL-1β (10 ng/ml), anti-IL-4 (0.4 μg/ml), anti-IL-12 (2 μg/ml), and anti-IFN-γ (8 μg/ml).

Statistical analysis

Statistical analysis was done using Student's t-test.

Results

Expression of ZBTB16 correlates positively with the Th17 phenotype, and negatively with Th1, Th2 and T_FH phenotypes in effector/memory CD4⁺ T cells

In our initial observations on the relationship between expression of IL-17A and CCR6 in human T cells, we noted that IL-17A was produced preferentially by cells with the highest surface expression of CCR6 (11). We hypothesized that CCR6 surface staining could be used in a general way to identify cells with differing levels of expression of genes (and proteins) characteristic of the Th17 phenotype, and that we might be able to identify factors regulating the Th17 phenotype by comparing global gene expression profiles among cells displaying various levels of CCR6. By sorting cells and activating them ex vivo, we found, as shown in Fig. 1A, that cells with increasing expression of CCR6, designated +1, +2, and +3, showed progressively higher expression of IL17A and RORC, and, conversely, CCR6 expression showed negative or no correlations with expression of genes characteristic of Th1, Th2, and T_FH cells.

FIGURE 1.

Expression of ZBTB16 correlates positively with the Th17 phenotype, and negatively with Th1, Th2 and T_FH phenotypes in effector/memory CD4⁺ T cells. (A) Gating strategy for sorting CD45RO⁻, CD45RO⁺, CCR6⁻, CCR6⁺, and, within the CCR6⁺ cells, CCR6+1, CCR6+2 and CCR6+3 subsets from CD4⁺ T cells purified from elutriated lymphocytes from healthy adult donors (top). Relative levels of mRNAs of indicated genes in cells sorted into subsets based on expression of CD45RO, and, within the CD45RO⁺ cells, based on expression of CCR6 (bottom). (B) ZBTB16 mRNA levels in cells sorted into subsets as shown in (A). (C) Intracellular staining with anti-PLZF antibody (left) or anti-RORγt antibody (right) in cells sorted into subsets as shown in (A). Results are from a single donor, representative of three donors/experiments. (D) ZBTB16 mRNA levels in CD4⁺ T cells sorted into subsets based on expression of CD45RO, and, within the CD45RO⁺ cells, based on expression of CXCR3, CCR4, and CXCR5. For (A), (B), and (D), after purification by cell sorting, cells were stimulated with PMA and ionomycin for 4 h and mRNAs were measured using real-time RT-PCR. Values were normalized to results for GAPDH and then to the results for CD45RO⁻ (naïve) cells. Bars show means ± SEM from three-five independent experiments, each with cells isolated from a different donor. For comparisons indicated by the horizontal lines above the bars, *, ** and ns denote p < 0.05, p < 0.01, and not significant, respectively.

A comprehensive analysis of gene expression profiles using microarrays will be published separately. However, as shown in Fig. 1B, among the genes showing a positive correlation with expression of CCR6 was ZBTB16, which encodes PLZF. Because of the high expression of PLZF in mouse and human NKT cells (29, 31, 33), we verified that the CD4⁺CD45RO⁺CCR6⁺ cells were not contaminated with any cells expressing the Vα24Jα18 TCR (data not shown). As shown in Fig. 1C, we also analyzed CD4⁺CD45RO⁺ cells for expression of PLZF by intracellular staining, and were able to detect some expression on the CCR6+2, and significantly more on the CCR6+3 cells, where staining of iNKT cells served as a positive control. The progressive increase in RORγt in these subsets is also shown. In order to determine if the correlation between ZBTB16 and CCR6 was specific, analogous to our approach using CCR6, we also sorted cells based on levels of expression of CXCR3, CCR4, and CXCR5, the signature chemokine receptors of the Th1, Th2, and T_FH pathways, respectively. In contrast to the data for CCR6, expression of ZBTB16 showed negative correlations with the other chemokine receptors (Fig. 1D). As shown in Supplemental Fig. 1, gene expression profiles support the validity of using CXCR3, CCR4, and CXCR5 to identify populations of increasing Th1 or Th2 or T_FH character, respectively. Of note, although ex vivo activation of cells was used whenever it was needed to reveal the cells’ cytokine profiles, activation did not affect the expression of the transcription factor genes ZBTB16, RORC, TBX21, and GATA3 (data not shown).

PLZF binds at enhancer-like elements in the 5’ flanking region of CCR6 in CCR6⁺ cells

In order to determine if PLZF has a direct role in regulating CCR6, we performed ChIP assays using primers spaced generally at 1 kb intervals along regions of CCR6. As shown diagrammatically in Supplemental Fig. 2A, CCR6 is predicted to have two transcription start sites (TSS's). Consistent with these assignments, ChIP using antibodies to RNA polymerase II (Pol II) showed enrichment in DNA fragments from the predicted TSS's in samples from the CD4⁺CD45RO⁺CCR6⁺, but not the CD4⁺CD45RO⁺CCR6⁻ cells. ChIP using antibodies to PLZF showed no enrichment in DNA from the TSS's, but did show enrichment in fragments at −9/−10 kb and −12 to −14 kb upstream of the 5’ TSS, again specifically in DNA from the CCR6⁺ cells.

Given the sites of PLZF binding that were revealed in the initial ChIP experiments, subsequent analysis was limited to a smaller region, extending from 18 kb upstream to 4 kb downstream of the 5’ TSS. We used antibodies to Pol II, histone H3 dimethyl lysine 4 (H3K4me2), and histone H3 trimethyl lysine 4 (H3K4me3), as well as PLZF, to immunoprecipitate DNA from CD4⁺CD45RO⁺ cells that were CCR6⁻ or within the CCR6+1, CCR6+2, or CCR6+3 subsets. As shown in Fig. 2A, the anti-Pol II precipitate was progressively enriched in DNA at the 5’ TSS in samples from the +1 to +3 subsets. H3K4me2, a modified histone associated with promoters and enhancers (34), was found preferentially near the TSS and at the regions −8 to −10 kb, −12 to −14 kb, and −16 to −18 kb, again with increasing signal from the CCR6+1 to CCR6+3 subsets. We found progressive enrichment in H3K4me3, as expected, at the TSS, but H3K4me3 was also detected at low levels in the CCR6+3 subset at −9/−10 kb and −13/−14 kb (see inset). For PLZF, we found increasing enrichment again at −9/−10 kb and −13/−14 kb (Fig. 2B). As shown in Supplemental Fig. 2B, the −9/−10 kb and −13/−14 kb regions have areas of conservation between human and other species but, notably, not with rodents. Using a combination of computational tools, we identified possible PLZF binding sites in the −9/−10 kb and −13/−14 kb regions, as well as elsewhere along CCR6 (Supplemental Fig. 1C).

FIGURE 2.

PLZF binds at enhancer-like elements in the 5’ flanking region of CCR6 in CCR6⁺ cells. (A) CCR6⁻, CCR6+1, CCR6+2 and CCR6+3 CD45RO⁺ CD4⁺ T cells were purified as in Fig. 1A and DNAs were analyzed by ChIP using antibodies against Pol II, H3K4me2, and H3K4me3. Precipitants were analyzed using a tilling array spanning the region from 18 kb upstream (−18) to 4 kb downstream (+4) of the CCR6 5’ TSS. CCR6 primers were positioned approximately midway within 1-kb intervals extending upstream and downstream of the 5’ TSS, which is marked by a right-facing arrow and was designated position 0. IGX1A primers are from an intergenic region lacking promoters and used as a negative control, and results for these primers are shown to the immediate left of the “IGX1A” label. For CCR6, data from a single primer pair are displayed between the tics bordering the one-kb region covered by those primers. Results are presented as % input enrichment from a single analysis, representative of three donors/experiments and an inset shows the H3K4me3 data using an expanded scale on the Y-axis for the region from −4 to −15 kb. (B) Samples as in (A), were analyzed using anti-PLZF antibody. Bars show means ± SEM from three independent donors/experiments. For comparisons indicated by the horizontal lines above the bars, *, **, and ns denote p < 0.05, p < 0.01, and not significant, respectively. (C) CD4⁺ T cells from healthy donors were sorted into CCR6⁻ and CCR6⁺ CD45RO⁺ CD4⁺ T cells and DNAs were analyzed by ChIP using antibodies against p300 or H3K27ac. Precipitants were analyzed and results presented as in (A). Data are from a single analysis, representative of three donors/experiments. (D) Cells were sorted as in (C) and transfected with Control or ZBTB16 siRNAs and after 72 h DNAs were analyzed by ChIP using anti-PLZF antibody. Precipitants were analyzed and results presented as in (A). Data are from a single analysis, representative of three donors/experiments.

Next, we used antibodies against p300/histone acetyltransferase KAT2B and histone H3 acetylated at lysine 27 (H3K27ac), which are associated with active enhancers (35, 36), and found evidence of both at the −9/−10 kb and −13/−14 kb regions specifically in the CCR6⁺ cells (Fig. 2C). As shown in Fig. 2D, we also performed ChIP using CCR6⁺ cells that had been transfected 3 days earlier with control siRNAs or siRNAs for ZBTB16. We found a decrease in PLZF binding at the −9/−10 kb and −13/−14 kb regions in cells transfected with ZBTB16 siRNAs as compared to siRNA controls.

Knockdown of ZBTB16 mRNA diminishes the expression of CCR6/CCR6 and other Th17-cell specific genes and proteins

We next investigated the role of PLZF in CCR6/CCR6 expression. As shown in Fig. 3A, at three days after transfection with ZBTB16 siRNAs there were significant decreases in the ZBTB16 mRNA and PLZF in the CCR6⁺ cells. As shown in Fig. 3B, knockdown of ZBTB16/PLZF also resulted in decreased expression of both CCR6 mRNA and surface CCR6 on the CCR6⁺ cells. We extended the functional analysis of PLZF to the expression of other genes characteristic of the Th17 phenotype in CCR6⁺ effector/memory cells. As shown in Fig. 3C, knockdown of ZBTB16 led to decreases in mRNAs for IL17A, IL17F, RORC, KLRB1 (which encodes CD161), IL23R and IL22. Knockdown of ZBTB16 had no effect on CCL20, or on IFNG and IL4. Knockdown of ZBTB16 also had no effect on expression of IL21, whose expression is enriched in the CCR6⁺ cells but is not limited to these cells or the Th17 pathway (37) (and see Fig. 1A and Supplemental Fig. 1). These data and those in Fig. 2D were obtained using a pool of four ZBTB16 siRNAs. As shown in Supplemental Fig. 3, we also tested four individual ZBTB16 siRNAs separately, only one of which was in the original siRNA pool, and found that these also knocked down expression of CCR6 and IL17A. (Expression of the additional genes shown in Fig. 3C were not analyzed.)

FIGURE 3.

Knockdown of ZBTB16 mRNA diminishes the expression of CCR6 and other Th17-cell specific genes. (A) CCR6⁻ and CCR6⁺ CD45RO⁺ CD4⁺ T cells were sorted from peripheral blood and transfected with Control or ZBTB16 siRNAs as indicated on the X axes. After 72 h mRNA levels for ZBTB16 were quantified using real-time RT-PCR (left panel). Values were normalized to GAPDH and then to the results for CCR6⁺ (Control) cells, which were set at 100. Bars show means ± SEM from five donors/experiments. Other panels show results for intracellular staining with anti-PLZF antibody in cells purified and treated similarly. Middle panel shows results from a single donor, representative of three donors/experiments and right panel shows pooled data of mean fluorescent intensities (MFI) from three donors/experiments. (B) Level of CCR6 mRNA (left panel) and CCR6 (right panel) in/on cells sorted into subsets and transfected with control or ZBTB16 siRNAs as in (A). Bars for mRNA expression show means ± SEM from five donors/experiments. Bars for MFI's show means ± SEM from three donors/experiments. (C) CCR6⁻ and CCR6⁺ CD45RO⁺ CD4⁺ T cells transfected as in (A), were stimulated with PMA and ionomycin and levels of mRNAs for the genes indicated on each bar graph were quantified using real-time RT-PCR. Values were normalized to GAPDH and then to the results for CCR6⁺ (Control) cells, which were set at 100. Bars show means ± SEM from 3-10 donors/experiments. (D) CCR6⁺ CD45RO⁺ CD4⁺ T transfected as in (A), were stimulated with the leukocyte activation cocktail for 6 h, fixed, permeabilized and stained for IL-17A, IL-22, or CCL20. For determining percentages of cells staining positive for these cytokines, as noted, quadrants were drawn based on staining with isotype-matched control antibodies (data not shown). Dot plots show results from a single donor, representative of three donors/experiments, and bar graphs show means ± SEM of % positive cells from three donors/experiments. (E) CD4⁺ T cells from healthy donors were sorted into CCR6⁻ and CCR6⁺ CD45RO⁺ CD4⁺ T cells and DNAs were analyzed by ChIP using anti-PLZF antibody. Precipitants were analyzed using a tilling array spanning the region from 26 kb upstream (−26) to 10 kb downstream (+10) of the RORC 5’ TSS, and results presented as in Fig. 2A. Bars show means ± SEM from three independent donors/experiments (top). *, **, and ***, denote p < 0.05, p < 0.01, and p < 0.001, respectively, for comparisons with CCR6⁻ cells. Below is a schematic representation of the region of RORC on chromosome 6, with transcriptional start sites indicated by left-facing arrows and positions of computationally predicted binding sites for PLZF identified with down-going bars. (F) CCR6⁻ and CCR6⁺ CD45RO⁺ CD4⁺ T cells were sorted from peripheral blood and transfected with Control or RORC siRNAs as indicated on the X axes. After 72 h, cells were stimulated with PMA and ionomycin and levels of mRNAs for the genes indicated on each bar graph were quantified using real-time RT-PCR. Values were normalized to GAPDH and then to the results for CCR6⁺ (Control) cells, which were set at 100. Bars show means ± SEM from 3-5 independent donors/experiments. For (A-D) and (F), *, **, ***, and ns denote p < 0.05, p < 0.01, p < 0.001, and not significant, respectively, for comparison with CCR6⁺ cells transfected with control siRNAs.

Using intracellular staining of CCR6⁺ cells activated ex vivo, we show in Fig. 3D that, consistent with the data on gene expression, knockdown of ZBTB16 mRNA resulted in decreased production of IL-17A and IL-22, but no change in CCL20. Since these results suggested that expression of RORC is partially dependent on ZBTB16, we investigated if PLZF might regulate RORC directly. As shown in Fig. 3E, ChIP analysis showed binding of PLZF near the downstream transcription start site and in the 5’ flanking region of RORC, suggesting a direct role for PLZF in RORC expression. We also considered whether, on the other hand, RORγt might regulate ZBTB16. As shown in Fig. 3F, knockdown of RORC did in fact decrease expression of ZBTB16, as well as IL17A, CCR6, and IL23R – although not CCL20.

It has been reported recently that PLZF can associate with the E3 ligase cullin 3, which, like PLZF, was required for NKT cell development (38). In data not shown, we investigated a role for CUL3 in the Th CCR6⁺ subsets and found no increase in CUL3 expression concomitant with the increase in expression of ZBTB16 from the CCR6+1 to +3 cells, and no effect of siRNA-mediated knockdown of CUL3 on the expression of ZBTB16, RORC, CCR6, or IL17A.

ZBTB16/PLZF supports the differentiation of human Th17 cells

We next investigated a possible role for PLZF in the differentiation of Th17 cells from naïve CD4⁺ T cells isolated from the blood of adults. As shown in Fig. 4A, activation of naïve cells ex vivo under Th17 conditions up-regulated ZBTB16, and, similar to the data using effector/memory cells, knockdown of ZBTB16 during culturing diminished expression of CCR6, IL17A, RORC, and KLRB1, but not CCL20. Along with the decrease in CCR6 mRNA, knockdown of ZBTB16 also led to loss of CCR6 surface expression (Fig. 4B).

FIGURE 4.

PLZF/ZBTB16 supports the differentiation of human Th17 cells. (A) Naïve CD4⁺ T cells were sorted from peripheral blood of adults, transfected with Control or ZBTB16 siRNAs as indicated on the X axes, polarized under Th17 conditions for 5 days and then stimulated with PMA and ionomycin before mRNAs for the genes indicated on each bar graph were quantified using real-time RT-PCR. Input cells (Naïve) were activated and harvested without culturing for purposes of comparison. Values were normalized to GAPDH and then to the results for naive cells, which were set at 1. Results are from a single donor, representative of three donors/experiments. (B) Naïve human CD4⁺ T cells were sorted from peripheral blood of adults, transfected with Control or ZBTB16 siRNAs as indicated, polarized under Th17 conditions for 5 days, stained for CCR6, and analyzed by FACS. Cells transfected with Control siRNAs were also stained with an isotype-matched control antibody. Results are from a single donor, representative of three donors/experiments.

Our data suggesting cross-regulation by PLZF and RORγt led us to investigate the effects of knocking down both factors on Th17 differentiation. Naïve CD4⁺ T cells from adult blood were transfected with siRNAs for ZBTB16 and RORC alone and in combination, and the cells were activated and polarized under Th17 conditions for 5 days. Given the rather modest, although still significant effects of knocking down ZBTB16 on expression of RORC and visa versa, knocking down ZBTB16 plus RORC did not result in lower expression of ZBTB16 than knocking down ZBTB16 alone (Fig. 5, leftmost panel), nor did it result in lower expression of RORC than knocking down RORC alone (Fig. 5, second panel). In decreasing the expression of CCR6, knocking down both ZBTB16 and RORC was no more effective than knocking down ZBTB16 or RORC alone (Fig. 5, third panel). These results are consistent with PLZF and RORγt acting cooperatively rather than independently on CCR6. On the other hand, in decreasing the expression of IL17A, knocking down both ZBTB16 and RORC was more effective than knocking down ZBTB16 or RORC alone (Fig. 5, right-most panel). These results are consistent with PLZF and RORγt acting independently rather than cooperatively than at IL17A.

FIGURE 5.

PLZF and RORγt show both interdependent and independent activities in regulating Th17-associated genes. Naïve CD4⁺ T cells were sorted from the peripheral blood of adults, transfected with Control, ZBTB16, RORC or a combination of both ZBTB16 and RORC siRNAs as indicated on the X axes, polarized under Th17 conditions for 5 days and then stimulated with PMA and ionomycin before mRNAs for the genes indicated above each graph were quantified using real-time RT-PCR. Input cells (Naïve) were activated and harvested without culturing for purposes of comparison. Values were normalized to GAPDH and then to the results for naive cells, which were set at 1. Bars show means ± SEM from three independent donors/experiments. * and ** on top of the vertical bars denote p < 0.05 and < 0.01, respectively, for comparison with cells transfected with control siRNAs. For comparisons indicated by the horizontal lines above the bars, * and ns denote p < 0.05 and not significant, respectively.

Although PLZF was reportedly absent from conventional T cells in mice (29, 31), we nonetheless analyzed a possible role for PLZF in mice for Th17 differentiation. As shown in Supplemental Fig. 4A, we could not detect induction of Zbtb16 in naïve mouse CD4⁺ T cells cultured under Th1 or Th17 conditions, although Zbtb16 mRNA could be readily detected in mouse NKT cells. Moreover, as shown in Supplemental Fig. 4B, naïve CD4⁺ T cells from mice homozygous for a nonsense (luxoid) mutation in Zbtb16 (39) could be differentiated into Th17 cells to the same extent as cells from wild-type mice.

ZBTB16/PLZF supports Th17 differentiation in CD161⁺ cells from cord blood

To investigate PLZF further in human Th17 differentiation, we used umbilical cord blood as the source of naïve CD4⁺ T cells. A 5-day time course of gene expression in cord blood CD4⁺ T cells after activation ex vivo under Th17 conditions showed early induction of RORC with a peak at Day 3, followed by induction of IL17A, which was continuing to rise as of Day 5 (Fig. 6A). Surprisingly, ZBTB16 mRNA was at relatively high levels in non-activated naïve cells, with levels falling initially after activation and then rising on Days 4 and 5. Cord blood has been reported to contain a population of naïve, CD4⁺CD161⁺ T cells that are the precursors of Th17 cells (40), but which are not found in the naïve population in adults (40). We sorted the cord blood cells into CD161⁻ and CD161⁺ subsets, and found, as shown in Fig. 6B, that expression of ZBTB16 was limited to the CD161⁺ cells. As shown in Fig. 6C, consistent with the published data (40), IL17A could be induced under Th17 culture conditions only in the CD161⁺ cells. In the CD161⁺ cells, just as in the cultures of the unfractionated naïve cells, knockdown of ZBTB16 significantly decreased the induction of CCR6 and IL17A.

FIGURE 6.

PLZF/ZBTB16 supports Th17 differentiation in CD161⁺ cells from cord blood. (A) Naïve CD4⁺ T cells were sorted from cord blood and cultured under nonpolarizing (NP) or Th17 conditions for 5 days. Cells were harvested at indicated time points and stimulated with PMA and ionomycin before mRNAs for IL17A, RORC, and ZBTB16 were quantified using real-time RT-PCR. Input cells (Naïve) were activated and harvested without culturing for purposes of comparison. Values were normalized to GAPDH and then, for IL17A and RORC to the results for Naive cells, which were set at 1. For ZBTB16, results were normalized to the lowest value, which was the 12-h sample for non-polarized cells, which were set at 1. Bars show means ± SEM for data from three donors/experiments. (B) Naïve CD4⁺CD161⁺ and CD4⁺CD161⁻ T cells were sorted from cord blood and ZBTB16 mRNA was quantified using real-time RT-PCR. Values were normalized to GAPDH and then to the results for CD161⁻ cells, which were set at 1. Bars show means ± SEM for data from three donors/experiments. *** denotes p < 0.001. (C) Naïve CD4⁺CD161⁺ and CD4⁺CD161⁻ T cells were sorted from cord blood, transfected with Control or ZBTB16 siRNAs as indicated on the X axes, polarized under Th17 conditions for 5 days and then stimulated with PMA and ionomycin before mRNAs for ZBTB16, CCR6, and IL17A were quantified using real-time RT-PCR. Values were normalized to GAPDH and then to the results for CD161⁻ (Control) cells, which were set at 1. Bars show means ± SEM for data from three donors/experiments. ** denotes p < 0.01. (D) Gating strategy for sorting CD4⁺ T cells purified from elutriated lymphocytes from healthy adult donors into CD45RO⁻ and CD45RO⁺ cells and the four CD45RO⁺ subsets based on expression of CCR6 and CD161 (left). Purified subsets were activated and stained for IL-17A and IL-22 (right). For determining percentages of cells staining positive for the cytokines, as noted, quadrants were drawn based on staining with isotype-matched control antibodies (data not shown). Data shown are from one donor, representative of more than three donors/experiments. (E) Cells were purified as in (D) and ZBTB16 mRNA was quantified using real-time RT-PCR. Values were normalized to GAPDH and then to the results for CD45RO⁻ cells, which were set at 1. Bars show means ± SEM from three donors/experiments. For comparisons indicated by the horizontal lines above the bars, * and ns denote p < 0.05 and not significant, respectively.

These data, and the reports that CD161 could be used as a marker for human Th17 cells (Cosmi et al., 2008), led us to examine more closely the relationships among expression of ZBTB16, CD161 and CCR6 in cells from adult blood. As shown for a representative donor in Fig. 6D, although among memory cells sorted on the basis of CCR6 and CD161 the CCR6⁺CD161⁺ cells contained the highest percentage of cells with inducible expression of IL-17A, the CCR6⁺CD161⁻ cells also contained a significant number of IL-17A producers. Numbers of IL-22-expressing cells did not differ between the CD161⁺ and CD161⁻ subsets. In analyzing the expression of ZBTB16, as shown in Fig. 6E, there were significant, independent positive correlations between CCR6 and CD161 and the expression of ZBTB16, so that ZBTB16 was up-regulated in both the CCR6⁺CD161⁻ and CCR6⁻CD161⁺ cells versus the CCR6⁻CD161⁻ cells.

Discussion

The starting point for our studies was CCR6, both in defining the cells of interest and in investigating mechanisms of gene regulation in these cells. In addition to data on roles for CCR6 in mouse models of immune-mediated diseases (15-19), genome-wide association studies have implicated CCR6 in Crohn's disease, rheumatoid arthritis, generalized vitiligo, and Graves’ disease (41-45). It may be relevant in this regard that the CCR6⁺ subset of Th cells encompasses not only the cells that can make IL-17A and/or IL-17F, but also cells expressing RORC, IL22, CCL20, and IL23R (15, 22), and S.P.S. and J.M.F., unpublished data, and this report) – all of which are associated with the Th17 phenotype, but which, importantly, are not invariably co-expressed (22, 46). This pattern suggests that CCR6 expression uniquely defines a major branch of the Th effector/memory population that includes cells with both Th17 and Th17-associated phenotypes. It follows that studies of CCR6⁺ cells and the regulation of CCR6 might reveal factors of importance for establishing and/or maintaining this more inclusive population.

We found that ZBTB16 expression was closely associated with the Th17 phenotype and that PLZF contributed to expression of not only IL-17A/IL-17F, but also many other Th17-associated genes. Outside the hematopoietic system, PLZF plays roles in limb patterning and in maintaining stem cells for spermatogonia (39, 47). Within the hematopoietic system, PLZF is important in the differentiation of iNKT cells (29, 33, 47, 48), and, in mice, a subset of innate-like γ/δ T cells and innate lymphoid cells (28-31). In humans, ZBTB16/PLZF has been described in MAIT cells (31), fetal CD4⁺ T cells selected on MHC class II⁺ thymocytes (49), and more recently in γ/δ T cells, NK cells, NKT cells, and subsets of CD4⁺ and CD8⁺ T cells (33).

Because we found both that knockdown of ZBTB16 diminished expression of CCR6 and RORC, and that PLZF bound within the upstream region of these genes, we conclude that PLZF is a direct, positive regulator of CCR6 and RORC in human Th cells. From our analysis of a broad region upstream of the CCR6 transcription start sites, we found PLZF binding to sites at −9/−10 kb and −13/−14 kb that also contained H3K4me2, H3K27ac, p300/histone acetyltransferase KAT2B, and low levels of H3K4me3. The presence of p300, along with this pattern of histone modification, are typically found at active enhancers (34-36, 50-53), and transcription factors often localize at enhancers identified in this way (34, 51). In analysis of the upstream region of RORC, we also found binding of PLZF, preferentially in the CCR6⁺ cells, generally but not uniformly corresponding to segments containing predicted PLZF binding sites.

The BTB-ZF transcription factors have generally been described as transcriptional repressors. PLZF, through its BTB domain and other regions, can interact with corepressor complexes (38, 54-57). Nonetheless, PLZF has also been described as a direct activator of gene expression (58). PLZF is a positive regulator of the gene encoding the p85α subunit of phosphoinositide 3-kinase (PIK3R1) as well as genes induced by type 1 interferons (59, 60). With the exception of CCR6 and RORC, whether or not PLZF is a direct activator of other Th17-associated genes remains to be established.

Studies in mice have revealed a number of transcription factors that are essential components of the Th17 genetic program, including factors that are relatively lineage-specific, such as RORγt and RORα, and those with more global activity, such as STAT3, Rel/NF-κB, BATF, IRF4, Runx1, HIF-1α, Ahr, and IκBζ (3, 4, 8, 61). Some of these latter factors (4, 62) serve as transcriptional activators of Rorc, as well as direct activators of downstream genes, such as Il17a (63). Although studies of the molecular pathways regulating the phenotype of human Th17 cells have been more limited, regulators that have been implicated include STAT3, RORγt, AHR, and HIF-1α (3). Our studies have now added PLZF to the list of factors important for supporting the Th17 phenotype in human Th cells. Moreover, we found that PLZF is among those factors that directly regulates not only “downstream” genes, such as CCR6, but also at least one “upstream” transcription factor, RORC/RORγt.

Because we found that knockdown of RORC also decreased expression of ZBTB16, we investigated their possible functional interactions by analyzing the effects on expression of CCR6 and IL17A after knocking down RORC and ZBTB16 in combination as well as singly. For CCR6, the results are consistent with the activities of PLZF and RORγt being interdependent. Mutual cross-regulation of expression by ZBTB16 and RORC is one possible mechanism of interdependence. However, such cross-regulation would not seem adequate to explain our findings. Firstly, knockdown of ZBTB16 has an effect on CCR6 expression equal to the effect of direct and significant knockdown of RORC, while the decrease in expression of RORC after knockdown of ZBTB16 is modest. By the same reasoning, it is unlikely that the effect of knockdown of RORC on expression of CCR6 was be due primarily to the modest decrease in ZBTB16 expression after knockdown of RORC. More importantly, our ChIP data show binding of PLZF to CCR6, and published data showed that RORγt binds to Ccr6 in mouse Th17 cells (61), providing strong evidence of direct regulation of CCR6 by these transcription factors. Together, the data suggest that PLZF and RORγt interact cooperatively at CCR6, at least functionally if not physically. By contrast, our data that knocking down both ZBTB16 and RORC was more effective at decreasing expression of IL17A than knocking down ZBTB16 or RORC alone are consistent with PLZF and RORγt acting independently in supporting expression of IL17A. In any case, our data suggest that both through cross-regulation that forms a positive feedback loop and functional cooperation ZBTB16 and RORC contribute to producing and maintaining the Th17 phenotype.

In addition to efforts to neutralize Th17-associated cytokines, inhibiting transcription factors is also being pursued with a view towards therapy (6, 7, 64). Factors that contribute to maintaining the expression of critical Th17-associated genes are of particular interest with respect to targeted therapies, since such therapies will always be used in circumstances of established disease mediated by differentiated effector/memory cells. Our data show that ZBTB16/PLZF is important not only for the differentiation of Th17 cells from naïve cells activated ex vivo, but also in the maintenance of the Th17 phenotype in effector/memory cells isolated from blood. Few transcription factors have been investigated in bona fide effector/memory human Th17 cells. As far as we are aware, only RORγt has also been examined in this context (7, 20). It is of interest that during the differentiation of naïve Th cells from cord blood, ZBTB16 is initially down-regulated in the CD161⁺ population before going up later, with a time course that matches that of IL17A but differs from that of RORC. This is consistent with a role for ZBTB16/PLZF in the stabilization of the Th17 program.

Similar to published data, we did not find expression of Zbtb16 in naive mouse CD4⁺ T cells from spleen, or in mouse Th17 (or Th1) cells activated and polarized ex vivo, nor was PLZF required for producing mouse Th17 cells, and conserved sequences corresponding to the PLZF binding regions in CCR6 were not present in mouse Ccr6. It will be of interest to investigate a possible analogous role for one of the other BTB-ZF factors in Th17 cells in mice.

In mice and/or humans, PLZF is expressed, and in some cases established to be functionally important, in ILC precursors (28) and in a number of lymphocyte subsets that exit the thymus either as effector cells or as cells pre-committed to develop down defined effector/memory pathways (29, 31, 49, 65). Of particular interest, after this manuscript was submitted, it was reported that PLZF is expressed in a small subset of “nTh17” cells, which leave the thymus displaying a memory phenotype and are able to produce IL-17A immediately after exposure to IL-23 plus IL-1β (66). Although an analogous population of cells has not been identified in humans, we cannot rule out that some of the PLZF-expressing memory cells may be such nTh17 cells. However, our ex vivo experiments clearly demonstrate that PLZF is induced during, and is important for, the differentiation of Th17 cells from adult, peripheral naïve CD4⁺ T cells, which is not the case in mice ((66) and this report). Therefore, whatever the contribution of nTh17 cells to the peripheral Th17 population in humans, our findings suggest that PLZF is also important for the differentiation of human induced Th17 (“iTh17”) cells.

Among the pre-committed lymphocytes that express ZBTB16 are naïve CD8⁺CD161⁺ T cells in human umbilical cord blood that are believed to contain precursors of MAIT cells, which display CCR6 and express other type-17 genes, such as RORC (65). Similarly, we found that ZBTB16/PLZF is expressed in the CD4⁺CD161⁺ cord blood T cells that, although unable to produce IL-17, give rise to Th17 cells in culture (40). It has been suggested that CD161 can serve as a marker for Th17 cells (40), although our data show that there are CD161⁻ Th17 cells. However, we found significant, independent associations of expression of ZBTB16 with CD161, as well as with CCR6, and knockdown of ZBTB16 had dramatic effects on the expression of KLRB1 (CD161) in CD4⁺ T cells. Taken together, these data on expression and function of ZBTB16/PLZF provide further support for the importance of common molecular pathways underlying the ontogeny and phenotypes of Th17 and MAIT cells, and suggest additional connections between Th17 cells and subsets of innate-like lymphocytes and innate lymphoid cells that warrant further investigation.

Abbreviations used in this article

APC

allophycocyanin

ChIP

chromatin immunoprecipitation

TSS

transcription start site

PLZF

promyelocytic leukemia zinc finger protein

Pol II

RNA polymerase II