Repression of the genome organizer SATB1 in regulatory T cells is required for suppressive function and inhibition of effector differentiation

Abstract

Regulatory T cells (T(reg) cells) are essential for self-tolerance and immune homeostasis. Lack of effector T cell (T(eff) cell) function and gain of suppressive activity by T(reg) cells are dependent on the transcriptional program induced by Foxp3. Here we report that repression of SATB1, a genome organizer that regulates chromatin structure and gene expression, was crucial for the phenotype and function of T(reg) cells. Foxp3, acting as a transcriptional repressor, directly suppressed the SATB1 locus and indirectly suppressed it through the induction of microRNAs that bound the SATB1 3' untranslated region. Release of SATB1 from the control of Foxp3 in T(reg) cells caused loss of suppressive function, establishment of transcriptional T(eff) cell programs and induction of T(eff) cell cytokines. Our data support the proposal that inhibition of SATB1-mediated modulation of global chromatin remodeling is pivotal for maintaining T(reg) cell functionality.

Figure 1. Foxp3-dependent repression of SATB1 expression in human T_reg cells

(a) Microarray analysis of SATB1 mRNA expression in human T_conv and T_reg (Rest=resting, Act=activated, TGF=TGF-β treated, Exp=expanded). (b) Relative SATB1 mRNA expression in human T_reg and T_conv. (c) Immunoblotting (left) for SATB1 (top) and β-actin (bottom) in human T_reg and T_conv, and densitometric analysis to quantify ratio of SATB1 to β-actin (right). (d) SATB1 expression was determined by flow cytometry (left) in human T_reg and T_conv with quantification shown in the graph at right. (e) SATB1 expression was determined by flow cytometry after stimulation of human T_reg and T_conv for 2 d (left) with quantification shown normalized to resting T_conv in the graph at right. (f) Relative SATB1 mRNA expression in human iT_reg, stimulated T cells (stim T) and unstimulated T cells (unstim T) on day 5. (g) SATB1 expression was determined by flow cytometry in human iT_reg, stim T and unstim T (left) with quantification shown in the graph at right. (h) Cytometric bead assay for IL-4 and IFN-γ secretion in the supernatants of human iT_reg, stim T and unstim T. (i) Relative Foxp3 (left) and SATB1 mRNA expression (right) in T_conv lentivirally transduced with Foxp3 vector (Foxp3 vector) or control vector (Ctrl vector) rested for 3 days. (j) Relative IL-5 (left) and IFN-γ mRNA expression (right), assessed and presented as in i. *P < 0.05 (Student’s t-test). Numbers in plots indicate mean fluorescence intensity (d,e). Data are representative of five experiments (b,e,i,j; mean and s.d.), six experiments (c,f; mean and s.d.), three experiments (g; mean and s.d.), eleven experiments (d; mean and s.d.), or three independent experiments (h; mean ± s.d. of triplicate wells), each with cells derived from a different donor.

Figure 2. Rescue of SATB1 expression after silencing of Foxp3 in T_reg cells

(a) Foxp3 expression was determined by flow cytometry (left) in human T_reg transfected with siRNA targeting Foxp3 (Foxp3 siRNA) or nontargeting (control) siRNA (Ctrl siRNA) 48 h post knockdown with quantification shown in the graph at right. *P < 0.05 (Student’s t-test). (b) Suppression of allogeneic CD4⁺ T cells, labeled with the cytosolic dye CFSE (responding T cells (T_resp)) by human T_reg transfected with siRNA targeting Foxp3 (Foxp3 siRNA) or nontargeting (control) siRNA (Ctrl siRNA), presented as CFSE dilution in responding T cells cultured with CD3+CD28-coated beads and T_reg at a ratio of 1:1 or without T_reg (T_resp only). (c) Relative SATB1 mRNA expression in human T_reg after silencing of Foxp3. T_reg were transfected and cultivated for 48 hours without stimulation (Rest) or in the presence of CD3 and IL-2 (α-CD3+IL-2) or CD3 and CD28 (α-CD3+α-CD-28). *P < 0.05 (Student’s t-test). (d) Relative IL-5 (left) and IFN-γ mRNA expression (right) in siRNA-treated T_reg stimulated for 48 hours in the presence CD3 and IL-2. *P < 0.05 (Student’s t-test). (e) Cytometric bead assay for IL-4 and IFN-γ secretion in the supernatants of siRNA-treated T_reg stimulated as in d. *P < 0.05 (Student’s t-test). (f) Relative IL-5 (left) and IFN-γ mRNA expression (right) in T_reg transfected with nontargeting (control) siRNA (Ctrl siRNA), Foxp3-targeting siRNA (Foxp3 siRNA) or Foxp3- and SATB1-targeting siRNA (Foxp3 + SATB1 siRNA) followed by stimulation with CD3+CD28-coated beads for 48 hours. *P < 0.05 (one-way ANOVA with LSD). Data are representative of six experiments (a,c; mean and s.d.), four experiments (d,f; mean and s.d.), three independent experiments (b), or three independent experiments (e; mean ± s.d. of triplicate wells), each with cells derived from a different donor.

Figure 3. Foxp3-dependent repression of SATB1 expression in mouse T_reg cells

(a) SATB1 expression was determined by flow cytometry (left) in freshly isolated T_reg and T_conv from spleen of male DEREG mice with quantification shown in the graph at right. *P < 0.05 (Student’s t-test). (b) Immunoblotting for SATB1 (top) and Erk1, Erk2 (bottom) in mouse T_reg and T_conv. (c) Z-projection of immunofluorescence for SATB1 (red) and GFP (green) in thymocytes from male DEREG mice; nuclei are stained with the DNA-intercalating dye DAPI. White arrow depicts T_reg. (d) Relative SATB1 mRNA expression in CD4⁺GFP⁺ (T_reg) and CD4⁺GFP⁻ (T_conv) T cells from male DEREG or Foxp3-deficient DEREG × scurfy mice. *P < 0.05 (Student’s t-test). (e) Relative IL-6 (left) and IFN-γ mRNA expression (right) in CD4⁺GFP⁺ T_reg derived from male DEREG or Foxp3-deficient DEREG × scurfy mice. *P < 0.05 (Student’s t-test). (f) Confocal microscopy for SATB1 (red) and Foxp3 (green) in thymic T_reg from female DEREG mice heterozygous for the scurfy mutation counterstained with DAPI (blue). CD4⁺GFP⁺FOXP3⁻ Foxp3-deficient T_reg (top, Foxp3^sf) or CD4⁺GFP⁺FOXP3⁺ Foxp3-competent T_reg (bottom, Foxp3^wt). (g) SATB1 expression was determined by flow cytometry (left) in thymic single positive CD4⁺GFP⁺FOXP3⁺ (left, grey) and CD4⁺GFP⁺FOXP3⁻ T_reg (right, black) from female DEREG mice heterozygous for the scurfy mutation with quantification shown in the graph at right. Isotype control shown as solid line. Numbers in plots indicate mean fluorescence intensity (a,g). Data are representative of three independent experiments (a; mean and s.d.), two independent experiments (g; mean and s.d.), two independent experiments (c,f,b), or two independent experiments (d,e; mean ± s.d. of triplicate wells).

Figure 4. Direct suppression of SATB1 transcription by Foxp3

(a) Foxp3 ChIP tiling array data (blue) from human expanded cord-blood T_reg. Data were analyzed with MAT and overlayed to the SATB1 locus to identify binding regions (1–13, p < 10⁻⁵ and FDR < 0.5%). Data are representative of two independent experiments with cells derived from different donors. (b) Schematic representation of Foxp3 binding regions (BR) at the human genomic SATB1 locus identified by in silico prediction within the regions identified in a. (c) ChIP analysis of human expanded cord-blood T_reg cells with a Foxp3-specific antibody and PCR primers specific for Foxp3 binding regions. Relative enrichment of Foxp3 ChIP over input normalized to IgG was calculated. A region −15 kb upstream (−15kb) was used as a negative control. *P < 0.05 (Student’s t-test). Data are representative of three independent experiments (mean and s.d.) with cells derived from different donors. (d) Luciferase assay of FOXP3 binding to the respective binding regions at the SATB1 locus in HEK293T cells, transfected with luciferase constructs containing wild-type (WT) or mutated (Mut) binding regions BR9–14 of the SATB1 locus (with mutation of the Foxp3-binding motifs), together with a FOXP3 encoding expression vector or control vector (Ctrl). *P < 0.05 (Student’s t-test). Data is from one representative experiment of two (mean and s.d. of triplicate wells). Numbers indicate Foxp3-binding motifs within each region. (e) Determination of the K_D-values of Foxp3 binding to a wild-type (WT) or mutated (Mut) Foxp3 binding motif in BR9 and BR10 at the SATB1 locus by filter retention analysis. Data are representative of three independent experiments (mean and s.d. of triplicates).

Figure 5. SATB1 expression in human T_reg reprograms T_reg into effector T cells

(a) Suppression of CD8⁺ T cells, labeled with the cytosolic dye CFSE (CD8⁺ T cells) by human T_reg lentivirally transduced with SATB1 (T_reg cells (SATB1), blue) or control vector (T_reg cells (Ctrl), red), presented as CFSE dilution in responding T cells cultured with CD3-coated beads and T_reg at a ratio of 1:1 or without T_reg (CD8⁺ T cells only; left), and as percent proliferating CD8⁺ T cells versus the T_reg cell/CD8⁺ T cell ratio (right). Data are representative of three independent experiments (mean and s.d.) with cells derived from different donors. (b) Cytometric bead assay for IL-4 and IFN-γ secretion in the supernatants of human T_reg lentivirally transduced with SATB1 (T_reg cells (SATB1), blue) or control vector (T_reg cells (Ctrl), red) assessed 4 and 16 h after stimulation with CD3+CD28-coated beads. *P < 0.05 (Student’s t-test). Data is from one representative experiment of two (mean and s.d. of triplicate wells) with cells derived from different donors. (c–e) Relative IL-5 (c), IFN-γ (d), and IL-17A (e) mRNA expression (right) in human T_conv (grey) and T_reg lentivirally transduced with SATB1 (T_reg cells (SATB1), blue) or control vector (T_reg cells (Ctrl), red) activated for 16 h with CD3+CD28-coated beads. *P < 0.05 (Student’s t-test). Data is from one representative experiment of two (mean and s.d. of triplicate wells) with cells derived from different donors.

Figure 6. Transcriptional T_eff programs are induced in SATB1 expressing T_reg

(a) Microarray analysis of human T_reg lentivirally transduced with SATB1 (SATB1) or control vector (Ctrl) after 16 hours stimulation with CD3+CD28-coated beads, presented as a heat map of differentially expressed genes. Data were z-score normalized. (b) Cross annotation analysis using 4 classes: genes associated with T_conv but not T_reg cells (T_conv cell-dependent, yellow), with T cell activation (activation-dependent, green), common T-cell genes (common, orange), and SATB1-induced genes (SATB1-induced, black). Numbers indicate genes within each group. (c) Visualization of gene expression levels of genes previously associated with T_H1, T_H2, or T_H17 differentiation, presented as a heat map. Data were z-score normalized. P values for T_H associated genes determined by χ² test in comparison to the complete data set (T_H specific gene enrichment): T_H1, P = 3.24×10⁻⁶; T_H2, P = 9.03×10⁻¹⁵; T_H17, P = 1.16×10⁻⁶. (d) Changes in genes associated with the human T_reg signature as assessed in T_reg lentivirally transduced with SATB1 (SATB1) in comparison to control vector (Ctrl). The mean log₂ fold-changes of the comparison T_reg to T_conv (red dots) and SATB1 to control-transduced T_reg (blue dots) were plotted and both comparisons were ranked by fold change in the T_reg vs. T_conv comparison. Data are derived from three independent experiments with cells derived from different donors.

Figure 7. SATB1 expression is repressed by miRNA in T_reg

(a) Representation of the human genomic SATB1 3′ UTR and the conserved miRNA binding sites. (b) Relative miR-155, miR-21, miR-7, miR-34a, and miR-18a expression human T_reg and T_conv. *P < 0.05 (Student’s t-test). Data are representative of five experiments (mean and s.d.) with cells derived from different donors. (c) Correlation of miRNA expression with SATB1 mRNA expression is plotted against miRNA fold change (T_reg vs. T_conv) for all 735 miRNA assessed by microarrays. Highlighted are miR-155, miR-21, miR-7, miR-34a, and miR-18a. Data are representative of three experiments with cells derived from different donors. (d) Foxp3 ChIP tiling array data (blue) for miR-155, miR-21, and miR-7 from human expanded cord-blood T_reg. Data were analyzed with MAT and overlayed to the respective locus to identify binding regions (P < 10⁻⁵ and FDR < 0.5%). Data are representative of two independent experiments with cells derived from different donors. (e) ChIP analysis of human expanded cord-blood T_reg cells with a Foxp3-specific antibody and PCR primers specific for miR-155, miR-21, and miR-7. Relative enrichment of Foxp3 ChIP over input normalized to IgG was calculated. The AFM locus was used as a negative control. *P < 0.05 (Student’s t-test). Data are representative of three independent experiments (mean and s.d.) with cells derived from different donors. (f) Dual-luciferase assay of HEK-293T cells transfected with luciferase constructs containing wild-type (WT) 3′ UTR or mutated (Mut) 3′ UTR of SATB1 (with mutation of the miRNA-responsive elements), together with synthetic mature miRNAs or a synthetic control miRNA (Ctrl). *P < 0.05 (Student’s t-test). Data are representative of three independent experiments (mean and s.d.). (g) Immunoblotting for SATB1 (top) and Erk1, Erk2 (bottom) in T_reg from mice with a T_reg-specific complete DICER loss (Dicer1^fl/fl) in comparison to Dicer1^+/fl T_reg. Data are representative of two independent experiments.

Figure 8. Ectopic expression of SATB1 in T_reg results in reduced suppressive function in vivo

(a) Suppression of CD4⁺ T cells, labeled with the cytosolic dye eFluor 670 (T_conv cells) cultured with CD3+CD28-coated beads by expanded mouse T_reg lentivirally transduced with SATB1 (SATB1, blue) or control vector (Ctrl, red), presented as percent proliferating T_conv cells versus the T_reg cell/T_conv cell ratio (right). (b) Relative IL-5 (left) and IFN-γ mRNA expression (right) in murine T_reg lentivirally transduced with SATB1 (SATB1, blue) or control vector (Ctrl, red). (c) Hematoxylin and eosin staining of colon sections from Rag2^−/− mice at 9 weeks after the transfer of CD4⁺CD45RB^hi naive T cells (naïve T) alone or in combination with control-transduced T_reg (T_reg cells (Ctrl)) or SATB1-transduced T_reg (T_reg cells (SATB1)). Original magnification, ×200; scale bars, 100 µm. (d) Histology scores of colon sections of Rag2^−/− mice at 9 weeks after the transfer in c. (e) Body weight of Rag2^−/− mice at 9 weeks after the transfer in c, presented relative to initial body weight. (f,g) Recovery of T_conv (f) and T_reg (g) from spleens, mesenteric, and peripheral lymph nodes of Rag2^−/− mice at 9 weeks after the transfer in c. *P < 0.05 (Student’s t-test). NS, not significant. Data are representative of three independent experiments (b; mean and s.d.), two independent experiments (a; mean and s.e.m. of triplicate cultures), or are pooled from two independent experiments (c–g; mean and s.d. of four resp. five recipient mice).