Volume 19, Issue 2, Pages 281-294 (August 2003) CD8 T Cell-Specific Downregulation of Histone Hyperacetylation and Gene Activation of the IL-4 Gene Locus by ROG, Repressor of GATA  Miyuki Omori, Masakatsu Yamashita, Masamichi Inami, Maki Ukai-Tadenuma, Motoko Kimura, Yukiko Nigo, Hiroyuki Hosokawa, Akihiro Hasegawa, Masaru Taniguchi, Toshinori Nakayama  Immunity  Volume 19, Issue 2, Pages 281-294 (August 2003) DOI: 10.1016/S1074-7613(03)00210-3

Figure 1 Cytokine Production Profiles and Transcript Levels in CD4 and CD8 T Cells Differentiated under Type 1- and Type 2-Skewed Conditions (A) Freshly prepared splenic CD4 and CD8 T cells from C57BL/6 or BALB/c mice were cultured under type 1- and type 2-skewed conditions for 5 days. Representative IFNγ/IL-4 profiles are shown. The numbers represent the percentages of cells in each quadrant. Four independent experiments including sorted CD44low naive CD4 and CD8 T cells were done with similar results. (B) C57BL/6 and STAT6-deficient T cells cultured as in (A) were restimulated with immobilized anti-TCR mAb for 24 hr, and the production of IL-4, IL-13, IL-5, IFNγ, and IL-2 in the culture supernatant was measured by ELISA. Three independent experiments were done with similar results. (C) Total RNAs were prepared from the cultured cells as in (A). The levels of mRNA of IL-5, IL-13, IL-4, and β-actin were determined by semiquantitative RT-PCR analysis with 3-fold serial dilution of template cDNA. Three experiments were done with similar results. Immunity 2003 19, 281-294DOI: (10.1016/S1074-7613(03)00210-3)

Figure 2 Limited Generation of IL-4-Producing CD8 T Cells Was Not due to IFNγ Production, Difference in Cell Cycle, or Low-Level Activation of TCR-Mediated Signaling (A) Freshly prepared splenic CD4 or CD8 T cells from C57BL/6 or IFNγ-deficient mice with C57BL/6 background were cultured under type 2-skewed conditions for 5 days. Representative IFNγ/IL-4 profiles after 6 hr restimulation (left panel) and mean cytokine concentrations in the culture supernatant after 24 hr restimulation are shown (right panel). The experiments were done twice and with similar results. (B) CD4 or CD8 T cells were labeled with CFSE and stimulated under type 2-skewed conditions. After 2 day cultivation, cells were restimulated and subjected to the intracellular staining with APC-conjugated anti-IL-4 mAb. The percentages of the cells gated representing numbers of cell division (#0 to #7) in total cells and of IL-4-producing cells in each gate are shown in the right panels. (C) Freshly prepared splenic CD4 or CD8 T cells were stimulated under the conditions as indicated and infected on day 2 with a retrovirus encoding an active form of raf (Raf-CAAX) or an active form of calcineurin (CN Δ-CAM) bicistronically with EGFP. Three days after infection, the cells were restimulated, and intracellular IFNγ/IL-4 profiles of electronically gated GFP+ populations were determined. The GFP (control) represents infection with an EGFP-containing retrovirus vector. Immunity 2003 19, 281-294DOI: (10.1016/S1074-7613(03)00210-3)

Figure 3 Chromatin Remodeling of the IL-4-Related Gene Loci in Tc2 Cells (A) Decreased histone hyperacetylation in the IL-4-related gene loci. C57BL/6 CD4 and CD8 T cells were cultured under type 1- or type 2-skewed conditions for 5 days. After overnight culture without stimulus, ChIP assays were performed. PCR was performed with 3-fold serial dilution of template cDNA. Shown are the PCR product bands for each primer pair (left) and the Tc2/Th2 ratio of the band intensities (right). Three independent experiments with different T cell preparations were done with similar results. (B) Failure of detecting intergenic transcripts of CNS1 site and VA enhancer site. Intergenic transcripts of CNS1 and VA enhancer sites in the indicated cells were determined by a semiquantitative RT-PCR analysis with 3-fold serial dilution of template cDNAs. (C) Histone hyperacetylation within the IL-4 and IL-13 loci in Th2 and Tc2 cells. Shown are the PCR product bands for each primer pair (bottom) and the Tc2/Th2 ratio of the band intensities (middle). Three independent experiments were performed with similar results. (D) Decreased hyperacetylation in the downstream region of the IL-13 gene exon 4 in Tc2 cells. Schematic representation of the IL-13 gene locus, and the locations of PCR primer pairs used in the ChIP assay are shown (upper panel). The amounts of input DNA from Th2 and Tc2 for PCR reactions were confirmed to be similar by direct PCR. Three independent experiments were performed and similar results were obtained. Shown are the PCR product bands for each primer pair (left) and the Tc2/Th2 ratio of the band intensities (right). (E) Decreased histone H3 hyperacetylation in the IL-4-related gene locus in Tc2 cells. Developing Th2 and Tc2 cells were prepared as in (A). After staining of intracellular IL-4 and IFNγ, IL-4-positive and IL-4-negative cells in IFNγ-negative fraction were sorted by flowcytometry to a purity >95%. Shown are the PCR product bands for each primer pair (left) and the arbitrary band intensities (right). Two independent experiments with different T cell preparations were done with similar results. Immunity 2003 19, 281-294DOI: (10.1016/S1074-7613(03)00210-3)

Figure 4 Ectopic Expression of GATA3 in Developing Tc2 Cells Did Not Induce Sufficient Generation of IL-4-Producing Cells (A) The amounts of GATA3, c-Maf, NFAT1 (NFATc2), NFAT2 (NFATc1), and Tubulin-α as a control in the developing Th2 and Tc2 cells were determined by immunoblotting with specific mAbs. The lysates from 1 × 106 (left lane) and 0.3 × 106 (right lane) cells were loaded per lane. The results are representative of three independent experiments. Arbitrary densitometric units are depicted under each band. (B) Freshly prepared splenic CD4 or CD8 T cells were stimulated under the conditions as indicated and were infected on day 2 with retrovirus encoding GATA3 bicistronically with EGFP (pMx-GATA3-IRES-GFP). Three days after the infection, the cells were restimulated, and intracellular IFNγ/IL-4 profiles of electronically gated GFP+ populations were determined. (C) The GFP+ cells prepared as in (B) were sorted on day 5 by a cell sorter, and the amount of GATA3 protein was determined by immunoblotting. The lysates from 1 × 106 GFP+ cells were loaded per lane. Immunity 2003 19, 281-294DOI: (10.1016/S1074-7613(03)00210-3)

Figure 5 A Conserved ROG Response Element (cROGRE) Located in the IL-13 Gene Exon 4 Binds to ROG, HDAC1, and HDAC2 In Vitro and In Vivo (A) mRNA levels of ROG, GATA3, T-bet, and β-actin in the 5-day cultured developing Th2 and Tc2 cells were determined by semiquantitative RT-PCR analysis. Shown are the PCR product bands. Arbitrary densitometric units are shown. Three independent experiments including different culture times were done with similar results. (B) ROG and Tubulin-α protein in the developing Th2 and Tc2 cells prepared as in (A) were detected by immunoblotting. Three experiments were done with similar results. (C) The DNA sequence of ROG (Tzfp binding site, tbs) motif on Aile1, the first ROG motif existing in exon 4 of the IL-13 gene (ROG motif in IL-13 exon4-#1) and the second ROG motif in exon 4 of the IL-13 gene (ROG motif in IL-13 exon4-#2) are shown in the top, and the location of the ROG#1 and ROG#2 sites in the IL-13 gene locus exon 4 is shown in the bottom. The asterisks represent critical nucleic acid residues for binding. (D) Cos 7 cells were transfected with a control pFLAG-CMV2 (Cont.) or pFLAG-CMV2-ROG (Flag-ROG), and nuclear extracts prepared 2 days later. Biotinylated ROG#1WT, ROG#2WT, and ROG#1 and ROG#2 oligonucleotides with adenine mutations were absorbed by streptavidin-agarose beads and then incubated with the nuclear extracts. Then the amounts of ROG protein in the precipitates were assessed by immunoblotting with anti-Flag mAb. Total nuclear extracts (1 × 106 equivalent/lane) were also run as controls. Two independent experiments were done with similar results. (E) Developing Th2 and Tc2 cells were prepared as in (A), and nuclear extracts prepared. Biotinylated ROG#1WT and the ROG#1 mutant used in (D) were used for precipitation. The amounts of ROG, HDAC1, and HDAC2 in the precipitates were assessed by immunoblotting with anti-ROG antibody, anti-HDAC1, and anti-HDAC2 mAb, respectively. Two independent experiments were done with similar results. (F) ChIP assay using normal rabbit IgG (Control Ig), anti-ROG, anti-HDAC1, and anti-HDAC2 antibodies, and a primer pair to amplify the ROG#1 site was done with developing Th2 and Tc2 cells. As a control, direct PCR products from input DNA are shown. The arbitrary densitometric units relative to each Tc2 band (1.0) are depicted under each band. For control Ig, the units are expressed relative to intensity of the band of ROG in Tc2. (G) Th2 and Tc2 cells (1 × 107) were subjected to immunoprecipitation with anti-ROG antibody and following immunoblotting with anti-HDAC2 antibody. Total cell lysates (1 × 106 equivalent) were used for detecting whole HDAC2 molecules in Th2 and Tc2 cells. Immunity 2003 19, 281-294DOI: (10.1016/S1074-7613(03)00210-3)

Figure 6 Repressive Function of cROGRE (A) M12 B cells were transfected with 10 μg of the promoter reporter constructs, pGL2-IL-4p, pGL2-IL-4p-cROGRE/372bp, or pGL2-IL4p-ΔcROGRE/265bp. M12 B cells express ROG protein (Miaw et al., 2000) but not GATA3 protein (Yamashita et al., 2002). Luciferase assay was done as described in the Experimental Procedures. (B) Repressive activity of ROG in the GATA3-dependent generation of IL-4-producing cells. A pMx-IRES retrovirus vector coexpressing ROG and GFP was used to infect CD4 T cells stimulated under type 1-skewed conditions on day 2. Cells were allowed to develop until day 5 in the presence of IL-2, IL-12, and anti-IL-4 mAb. Then, the infected GFP+ cells were purified by cell sorting to a purity >95%. After an additional 2 day culture in the presence of IL-2, the cells were stimulated again with anti-TCR mAb under type 1-skewed conditions for 2 days. Then (on day 9 after the start of the original culture) cells were infected again with pMx-IRES-CAR-GATA3 or control pMx-IRES-CAR retrovirus. Another 4 day culture in the presence of IL-2, IL-12, and anti-IL-4 mAb allowed for GATA3-dependent generation of IL-4-producing cells. Then, CAR+ GATA3-infected cells were electronically gated, and the levels of IL-4-producing cells were compared between GFPlow (expressing low levels of ROG) and GFPhigh (expressing high levels of ROG) populations. The percentages of IL-4-producing cells are shown in each panel. The expression profiles of CAR on the gated cells (G1 and G2) are also shown. Immunity 2003 19, 281-294DOI: (10.1016/S1074-7613(03)00210-3)

Figure 7 A Schematic Representation of Possible Molecular Events Underlying Type 2-Specific Histone Hyperacetylation within the IL-13 and IL-4 Gene Loci in Developing Th2 and Tc2 Cells Details are described in the Discussion. Immunity 2003 19, 281-294DOI: (10.1016/S1074-7613(03)00210-3)