Temporal and context-dependent requirements for the transcription factor Foxp3 expression in regulatory T cells

Abstract

Regulatory T (T_reg) cells, expressing the transcription factor Foxp3, are obligatory gatekeepers of immune responsiveness, yet the mechanisms by which Foxp3 governs the T_reg transcriptional network remain incompletely understood. Using a novel chemogenetic system of inducible Foxp3 protein degradation in vivo, we found that while Foxp3 was indispensable for the establishment of transcriptional and functional programs of newly generated T_reg cells, Foxp3 loss in mature T_reg cells resulted in minimal functional and transcriptional changes under steady state. This resilience of the Foxp3-dependent program in mature T_reg cells was acquired over an unexpectedly long timescale; however, in settings of severe inflammation, Foxp3 loss led to a pronounced perturbation of T_reg cell transcriptome and fitness. Furthermore, tumoral T_reg cells were uniquely sensitive to Foxp3 degradation, which led to impairment in their suppressive function and tumor shrinkage in the absence of pronounced adverse effects. These studies demonstrate a context-dependent differential requirement for Foxp3 for T_reg transcriptional and functional programs.

Fig. 1. Foxp3 degradation causes minimal immune activation in adult lymphoreplete mice.

a, Schematic of the inducible Foxp3 protein degradation model. SCF, Skp-1-Cullin-F-box complex. b, Schematic of the Foxp3^AID and R26^TIR alleles. c, Flow cytometry plot showing 5-ph-IAA-induced Foxp3 protein degradation after 24 h (left). Scatter-plot of Foxp3 protein expression median fluorescence intensity (MFI) as assessed by flow cytometric analysis in Foxp3^AID mice after 7 days of daily 5-ph-IAA injection (right). d, Experimental design. e, Size of spleen and lymph nodes after 14 days of T_reg cell ablation or Foxp3 degradation. f, Activation, proliferation and cytokine production of CD4⁺ (top) and CD8⁺ (bottom) T cells following T_reg cell ablation or Foxp3 degradation. g, Number of eosinophils, neutrophils and monocytes and CD86 levels on dendritic cells following T_reg cell ablation or Foxp3 degradation. h, Serum antibody levels following T_reg cell ablation or Foxp3 degradation. i,j, Representative hematoxylin and eosin (H&E) stain (i) and histology scores (j) of the liver following T_reg cell ablation or Foxp3 degradation. k, Liver damage measured by serum alanine aminotransferase, albumin and albumin:globulin ratio. Scatter-plots represent mean ± s.e.m. Each point represents a unique mouse. Data are pooled from two independent experiments. Statistical analysis was conducted by one-way analysis of variance (ANOVA). Source data

Fig. 2. Foxp3 degradation induces minimal gene expression and functional changes in mature T_reg cells.

a, Experimental design of scRNA-seq and functional assays. Each genotype and time point consisted of four independent biological replicates. b, UMAP visualization of scRNA-seq data from Foxp3^AIDR26^WT and Foxp3^AIDR26^TIR1(F74G) T_reg cells before and 3 or 7 days after 5-ph-IAA-induced Foxp3 degradation. c, UMAP visualization of the same scRNA-seq data, colored by identified clusters. d, Fraction of each cluster within the total pool of Foxp3^AIDR26^WT or Foxp3^AIDR26^TIR1(F74G) T_reg cells separated by time point. Each point represents a unique mouse. e, In vitro suppression assay of T_reg cells sorted from Foxp3^AIDR26^WT and Foxp3^AIDR26^TIR1(F74G) mice after 7 days of in vivo 5-ph-IAA treatment. 5-ph-IAA was included in culture to sustain Foxp3 degradation. Line graph represents mean ± s.e.m. Data are pooled from two independent experiments and analyzed using two-sided multiple t-tests. NS, not significant. f, Experimental design of the bulk RNA-seq analysis. Each genotype and time point consisted of three independent biological replicates. g, Gating strategy for sorting resting and activated T_reg cells. h, Schematic comparison of the Foxp3^GFPKO reporter-null allele and the functional Foxp3^GFP allele. i, Scatter-plots and bar graphs showing the number of DEGs in resting or activated T_reg cells caused by Foxp3 protein degradation or genetic Foxp3 deficiency. Source data

Fig. 3. Foxp3 degradation in mature T_reg cells induces expression changes in a small set of genes.

a, T_reg cells from the scRNA-seq dataset were classified as resting or activated based on exceeding the threshold for resting or activated gene signature scores and were subsequently analyzed. b, Scatter-plot showing the correlation of gene expression changes induced by Foxp3 degradation at day 3 and day 7 in resting and activated T_reg cells. FC, fold change. c, UMAP visualization of resting and activated T_reg cells colored by gene signature scores for the ‘TIR1-up’ and ‘TIR1-down’ gene sets, up- and downregulated upon Foxp3 degradation, respectively. d, Dot plot summarizing statistically significant DEGs in resting or activated T_reg cells following 3 or 7 days of 5-ph-IAA-induced Foxp3 degradation. The color represents the log₂ fold change of R26^TIR1(F74G) versus R26^WT and the size represents the Benjamini–Hochberg adjusted P value of the differential expression test. e, Flow cytometry analysis of Foxp3 protein and mRNA levels (reported by ZsGreen) in Foxp3^AID T_reg cells from heterozygous Foxp3^AID/WTR26^WT and Foxp3^AID/WTR26^TIR1(F74G) females after 7 days of Foxp3 degradation. Scatter-plots represent mean ± s.e.m. Data are pooled from two independent experiments and analyzed using a two-way ANOVA. pLN, peripheral lymph nodes. f, Flow cytometry analysis of CD25, CD122, OX40, GITR and FR4 protein levels in Foxp3^AID T_reg cells from heterozygous Foxp3^AID/WTR26^WT and Foxp3^AID/WTR26^TIR1(F74G) females after 7 days of Foxp3 degradation. g, Flow cytometry analysis of CD127 and TCF1 protein levels in Foxp3^AID T_reg cells from heterozygous Foxp3^AID/WTR26^WT and Foxp3^AID/WTR26^TIR1(F74G) females after 7 days of Foxp3 degradation. Scatter-plots represent mean ± s.e.m. (f,g). Each point represents a unique mouse. Data are pooled from two independent experiments and analyzed with two-sided multiple t-tests. Source data

Fig. 4. Foxp3 degradation-sensitive genes in mature T_reg cells are enriched for Foxp3 binding.

a, Bar graphs showing the proportion of ATAC-seq peaks near Foxp3 degradation-induced DEGs bound by Foxp3 (ref. ). Genes are stratified by statistical significance (P values) in resting and activated T_reg cells. Bars represent mean ± s.e.m. Data were analyzed using a Mann–Whitney U-test. b, Bar graphs showing the proportion of ATAC-seq peaks near Foxp3-dependent DEGs bound by Foxp3. Genes are stratified by P values in resting and activated T_reg cells. Bars represent mean ± s.e.m. Data were analyzed using a Mann–Whitney U-test. c, H3K27Ac and H3K27me3 ChIP-seq signals at Foxp3-bound ATAC-seq peaks near Foxp3 degradation-induced DEGs. Line graph represents mean ± s.e.m. Data were analyzed using a Mann–Whitney U-test. d, Dot plot showing transcription factor motif enrichment within Foxp3-bound regions near Foxp3 degradation-induced DEGs. e, Schematic diagram illustrating the ‘on’ and ‘off’ states of the reversible reporter-null Foxp3^LSL allele. f, Experimental design of the gain-of-function experiment to induce Foxp3 expression in T_reg ‘wannabe’ cells. Each genotype and time point consisted of two independent biological replicates. g, Line graph depicting gene expression changes of TIR1-up or TIR1-down genes below a specific P value cutoff across different time points following Foxp3 induction in T_reg ‘wannabe’ cells. The line and shading represent the mean ± s.e.m. Data were analyzed using a Mann–Whitney U-test.

Fig. 5. Foxp3 is preferentially required for regulation of gene expression during early T_reg cell differentiation.

a, Experimental design for transcriptional profiling of developing thymic T_reg cells. Each genotype consisted of three independent biological replicates. b, Gating strategy used to sort CD73⁻ nascent thymic T_reg cells. c, Bar graph comparing the number of Foxp3 degradation-induced DEGs in thymic, resting and activated T_reg cells from Foxp3^AID/WT mice. d, Pearson correlation between Foxp3 degradation-induced and Foxp3-dependent DEGs in thymic, resting and activated T_reg cells. e, Scatter-plot and cumulative distribution function (CDF) plots comparing Foxp3 degradation-induced and Foxp3-dependent DEGs across the three T_reg populations. Data were analyzed using a Mann–Whitney U-test. f, Metacell analysis of thymocyte scRNA-seq data correlating UMI-normalized Foxp3 expression levels in each metacell with the expression of TIR1-up and TIR1-down gene signatures identified in a–c. UMAP plots are colored by scaled expression levels of TIR1-up, TIR1-down and UMI-normalized counts of Foxp3. Dashed red line depicts line of best fit. Correlations and corresponding P values were calculated with Pearson correlation over all genes. g, Experimental design for in vivo Foxp3 degradation in 1-day-old neonatal Foxp3^AID mice and adult Foxp3^AID mice. h, CD4⁺ and CD8⁺ T cell activation in adult and neonatal Foxp3^AID mice following Foxp3 degradation. i, Expansion of eosinophils and neutrophils in adult and neonatal Foxp3^AID mice after Foxp3 degradation. j,k, Representative H&E staining (j) and histology scores of liver inflammation (k) in neonatal Foxp3^AID mice following Foxp3 degradation. l, In vitro suppression assay of T_reg cells sorted from Foxp3^AIDR26^WT and Foxp3^AIDR26^TIR1(F74G) neonatal mice after 7 days of in vivo 5-ph-IAA administration. 5-ph-IAA was also included in culture to maintain Foxp3 degradation. Each point represents a unique mouse (h–l). Data are pooled from two independent experiments. Scatter-plots represent mean ± s.e.m. Data were analyzed using a one-way ANOVA. m, Bar graphs summarizing the number of Foxp3 degradation-induced DEGs in T_reg cells from neonatal and adult Foxp3^AID/y mice after 14 days of Foxp3 degradation. n, Scatter-plot correlating gene expression changes induced by Foxp3 degradation and Foxp3 gene deficiency in neonatal mice. o, Scatter-plots comparing gene expression changes induced by 7 days of Foxp3 degradation in neonatal T_reg cells to those in adult thymic, resting and activated T_reg cells from Foxp3^AID/WT mice. Source data

Fig. 6. Proliferation and inflammation increase T_reg cell sensitivity to Foxp3 degradation.

a, Experimental design of proliferating T_reg analysis in vivo. b, Flow cytometry analysis of CD25, GITR and CTLA4 protein levels in dividing versus nondividing T_reg cells following 7 days of in vivo Foxp3 degradation, in comparison to Foxp3-deficient T_reg ‘wannabe’ cells. Scatter-plots represent mean ± s.e.m. Each point represents a unique mouse. Data are representative of two independent experiments and were analyzed using a one-way ANOVA. c, Experimental design of proliferating T_reg cell analysis in vitro. d,e, Combined data (d) and representative plots (e) showing CD153, GARP, CD4 and Foxp3 protein levels in lowly and highly divided T_reg cells. Similarly treated naive CD4 T cells serve as Foxp3⁻ controls. Bar graphs represent mean ± s.e.m. Each point represents cells from a unique mouse. Data are representative of two independent experiments and were analyzed using a two-way ANOVA. f, IL-2, IL-4 and IL-13 concentrations in the supernatant of in vitro proliferating T_reg assay. Bar graphs represent mean ± s.e.m. Each point represents cells from a unique mouse. Data are pooled from two independent experiments and were analyzed using a two-tailed t-test. g, Experimental design of proliferating T_reg cell analysis in vitro. Experiment consisted of five technical replicates per genotype. h, Volcano plots of DEGs from g separated by their presence among all up- or downregulated genes in identified in Fig. 3a. Numbers of up- or downregulated genes are labeled in each plot. i, Experimental design of inflammatory T_reg cell analysis in vivo. Experiment consisted of four biological replicates per condition (Tcrbd double knockout (dKO) or Foxp3^DTR). j, Flow cytometric analysis of the transferred T_reg cell mixture in i. k, Recovery of cells in each condition stratified by genotype (mCherry⁺ for Foxp3^AIDR26^TIR1(F74G) or mCherry⁻ for Foxp3^AIDR26^WT) as identified in the scRNA-seq data. Each point represents a unique mouse. Bar graphs represent mean ± s.d. Data were analyzed using a paired two-tailed t-test. l, Density contour plots of R26^WT and R26^TIR1(F74G) overlayed on UMAP embeddings of the scRNA-seq data. m, Leiden clustering of gene expression data visualized on UMAP embedding for T_reg cells from each condition. Clustering was performed independently for each condition using the same resolution value. Fraction of each cluster within the total pool of R26^WT and R26^TIR1(F74G) T_reg cells separated by condition. Within each genotype, each point represents a unique mouse. Data were analyzed using a paired two-tailed t-test. n, Number of DEGs in each condition, colored by up- or downregulation. o, Scatter-plot of log₂ fold changes of DEGs between Tcrbd KO and Foxp3^DTR conditions. Points are colored by their direction and populations in which they are altered. Source data

Fig. 7. Foxp3 degradation leads to tumor shrinkage with minimal adverse effects.

a, Schematic of the tumor experiment design. s.c., subcutaneous. b, Tumor burden over time, shown as average (left) and individual (right) tumor growth curves. Line graph represents mean ± s.e.m. (left). Each line represents a unique mouse (right). Data are pooled from two independent experiments and were analyzed using a two-way ANOVA (mixed-effects model) with Geisser-Greenhouse correction. c, Representative tumor images on day 20. d, Representative flow cytometry plots (left) and combined data (right) of IFNγ production by tumor-infiltrating CD8⁺ T cells. Each point represents a unique mouse. Scatter-plot shows mean ± s.e.m. Data are pooled from two independent experiments and were analyzed using a two-tailed t-test. e, Representative flow cytometry plots (left) and quantification (right) of IL-4 production by tumor-infiltrating ZsGreen⁻ CD4⁺ T cells. Each point represents a unique mouse. Scatter-plot shows mean ± s.e.m. Data are pooled from two independent experiments and were analyzed using a two-tailed t-test. f, Body weight monitoring throughout the experiment. Line graph shows mean ± s.e.m. Data are pooled from two independent experiments. g, H&E staining of liver and intestine on day 20. Images are representative of two independent experiments. h, Expression levels of Foxp3, GITR, CD39, ZsGreen, CTLA4 and TCF1 in ZsGreen⁺ CD4 T cells from the dLN, ndLN and tumor on day 20. Each point represents a unique mouse. Scatter-plots show mean ± s.e.m. Data are pooled from two independent experiments and were analyzed using multiple t-tests. i,j, UMAP visualization of scRNA-seq analysis of tumor T_reg cells from Foxp3^AIDR26^WT and Foxp3^AIDR26^TIR1(F74G) mice on day 14 after tumor implantation, colored by genotype (i) or cluster (j). k, Heatmap showing scaled mean UMI-normalized expression values for each cluster in j. l, Proportional distribution of Foxp3^AIDR26^WT and Foxp3^AIDR26^TIR1(F74G) T_reg cells within each cluster in j. Each point represents a unique mouse. Scatter-plot represents mean ± s.e.m. Data were analyzed using multiple log-normal t-tests. m, Number of DEGs between Foxp3^AIDR26^WT and Foxp3^AIDR26^TIR1(F74G) T_reg cells in each cluster shown in j. Source data

Extended Data Fig. 1. Generation of Foxp3^AID mice.

(a) Gene targeting strategy. The auxin-inducible degron (AID) sequence was fused to the N-terminus of Foxp3 via a seven-amino-acid flexible linker. An IRES-ZsGreen-T2A-iCre-Frt-neo-Frt cassette was inserted into the 3’ UTR. Arrows indicate the locations of PCR primers used to distinguish Foxp3^WT and Foxp3^AID alleles. IRES: internal ribosome entry sequence. iCre: codon improved Cre recombinase. PGK: PGK promoter. Neo: neomycin-resistant gene. pA: bGH polyA sequence. (b) PCR validation of the knock-in allele. Data is representative of over 200 mice from each genotype. (c) Expression pattern of the ZsGreen reporter in Foxp3^AID/WT heterozygous females. ZsGreen expression was restricted to CD25⁺ CD4 T cells, consistent with Foxp3 expression. (d) ZsGreen⁺ Foxp3^AID T_reg cells suppressed naïve CD4⁺ T cell proliferation comparably to Foxp3^GFP T_reg cells in vitro. Line graph represents mean ± SEM of two biological replicates. (e) Naïve CD4⁺ T cells from Foxp3^AID mice were cultured under T_reg inducing conditions and transduced with either a TIR1-encoding retrovirus or the empty vector control. AID-tagged Foxp3 protein was selectively degraded in TIR1-transduced induced T_reg (iT_reg) cells upon indole acetic acid (IAA) treatment. Scatter plots represents mean ± SEM. Each point represents cells from a unique mouse. Source data

Extended Data Fig. 2. Generation of ROSA26^TIR1 and ROSA26^TIR1(F74G) mice.

(a) Gene targeting strategy. WPRE: Woodchuck hepatitis virus post-transcriptional regulatory element; DTA: Diphtheria toxin fragment A). (b) Southern blot validation of heterozygous ROSA26^TIR1/+ mice using the hybridization probe shown in (a). Data is representative of two independent F1 mice. (c) PCR validation of ROSA26^TIR1/+ mice. Data are representative of over 100 mice from each genoype. (d–f) Naïve CD4⁺ T cells from ROSA26^TIR1/+ mice were co-transduced with retroviruses expressing Cre and AID-Foxp3 (d). TIR1 expression was induced in a Cre-dependent manner (e), resulting in AID-Foxp3 degradation upon IAA treatment (f). Data are representative of two independent experiments. (g) Guide RNA (gRNA) design for CRISPR-mediated F74-to-G mutation in TIR1. The gRNA seed sequence is shown in gray; the PAM sequence is in pink. The F74G mutation creates a KasI restriction site. (h–i) Validation of the F74G mutation by KasI digestion (h) and Sanger sequencing (i). (j–k) The TIR1 F74G mutation enables in vivo protein degradation in response to 5-ph-IAA. Source data

Extended Data Fig. 3. Foxp3 protein degradation in adult lymphoreplete mice induces minimal immune activation.

(a) Size of the spleen and lymph nodes after 28 days of Foxp3 degradation. (b) Serum antibody levels following T_reg ablation or Foxp3 degradation. (c,d) Representative H&E stain (c) and histology scores (d) of the skin following T_reg ablation or Foxp3 degradation. Each point represents a unique mouse. Data are pooled from two independent experiments. Scatter blots represent mean ± SEM. One-way ANOVA. Source data

Extended Data Fig. 4. Inducible Foxp3 gene knockout causes minimal immune activation in adult lymphoreplete mice.

(a) Experimental design. (b-c) Representative plot (b) and combined data (c) showing the efficiency of Foxp3 gene knockout. (d) Size of the spleen and lymph nodes after 14 days of inducible Foxp3 gene knockout. (e-g) T cell activation (e), serum antibody levels (f), and myeloid cell expansion (g) following 14 days of inducible Foxp3 gene knockout. (h) Representative H&E stain (left) and histology scores (right) of the skin, liver, and lung following 14 days of T_reg ablation or Foxp3 degradation. Each point represents a unique mouse. Data are pooled from two independent experiments. Scatter blots represent mean ± SEM. One-way ANOVA. Source data

Extended Data Fig. 5. Foxp3 degradation in developed T_reg cells induces minimal gene expression and chromatin accessibility changes.

(a) UMAP visualization of scRNA-seq data from Foxp3^AIDR26^WT and Foxp3^AIDR26^TIR1(F74G) T_reg cells on days 0, 3 and 7 of 5-ph-IAA-induced Foxp3 degradation. (b) UMAP visualization of representative genes from the single-cell RNA-seq dataset colored by expression level. (c) MA plot showing differentially accessible ATAC-seq peaks induced by Foxp3 degradation (left) and Foxp3 gene deficiency (right). Red points represent differentially accessible regions.

Extended Data Fig. 6. Foxp3 degradation in developed T_reg cells altered the expression of a small group of genes.

(a) The composition of Foxp3^AIDR26^WT and Foxp3^AIDR26^WT(F74G) T_reg cells within resting and activated subsets from day 0, day 3, and day 7 of Foxp3 degradation. (b) Number of resting and activated T_reg cells utilized for the differential gene expression analysis in Fig. 3. (c) Flow cytometry analysis of Foxp3 protein and mRNA levels (reported by ZsGreen) in Foxp3^WT T_reg cells from heterozygous Foxp3^AID/WTR26^WT and Foxp3^AID/WTR26^TIR1(F74G) females after 7 days of Foxp3 degradation. Scatter plots represent mean ± SEM. Data are pooled from two independent experiments. Two-way ANOVA. (d) Bulk RNA-seq read counts of indicated genes in Foxp3^AID T_reg cells from heterozygous Foxp3^AID/WTR26^WT and Foxp3^AID/WTR26^TIR1(F74G) females after 7 days of Foxp3 degradation. Each point represents a unique mouse. Bar graphs represent mean ± SEM. (e) Flow cytometry analysis of CD25, CD122, OX40, GITR, and FR4 protein levels in Foxp3^WT T_reg cells from heterozygous Foxp3^AID/WTR26^WT and Foxp3^AID/WTR26^TIR1(F74G) females after 7 days of Foxp3 degradation. (f) Flow cytometry analysis of CD127 and TCF1 protein levels in Foxp3^AID T_reg cells from heterozygous Foxp3^AID/WTR26^WT and Foxp3^AID/WTR26^TIR1(F74G) females after 7 days of Foxp3 degradation. (e-f) Scatter plots represent mean ± SEM. Each point represents a unique mouse. Data are pooled from two independent experiments. Multiple t-tests. Source data

Extended Data Fig. 7. Genes sensitive to Foxp3 degradation in mature T_regs are potentially enriched for direct Foxp3 targets.

(a–b) Representative tracks showing Foxp3 ChIP, T_reg H3K27Ac, T_reg H3K27me3, and RNA-seq profiles in activated, resting, and nascent thymic T_reg cells for candidate Foxp3-activated (a) and Foxp3-repressed (b) genes. (c) Representative flow cytometry plots and gating strategy for RNA-seq analysis of T_reg “wannabe” cells with or without Foxp3 induction at the indicated time points following 4-OHT administration.

Extended Data Fig. 8. Conservation of Foxp3 degradation effects across different tissues.

(a) Experimental design of single-cell RNA-seq experiment. Each genotype consisted of four independent biological replicates. (b) Density contour plots of Foxp3^AIDR26^WT and Foxp3^AIDR26^TIR1(F74G) T_reg cells from SLO, liver, lung and LILP overlayed on UMAP embeddings. (c) UMAP visualization of T_reg cells from depicted tissues colored by “resting” and “activated” gene signature scores based on the scoring strategy from Fig. 3a. (d) Bar plots showing the number of DEGs between R26^WT and R26^TIR1(F74G) in resting and activated T_reg cells from indicated tissues, colored by up- or downregulation. (e) Pairwise Pearson correlation coefficients for log₂ fold changes of DEGs between activated and resting T_reg cells from each tissue. Rows and columns are hierarchically clustered. (f) Heatmap of log₂ fold changes in Foxp3 degradation DEGs separated by T_reg cell activation status and tissue of origin. Genes are clustered by k-means clustering and cluster numbers are indicated by the color bars on the left. (g) Heatmap of average log₂ fold changes for each k-means cluster in T_reg cells separated by tissue and activation status. Rows and columns are hierarchically clustered. (h) Leiden clustering results visualized by UMAP for LILP T_reg cells. Clusters are labeled as peripheral T_reg cell (pT_reg), dividing T_reg cell, lymphoid tissue thymic T_reg cell (LT-tT_reg) and non-lymphoid tissue thymic T_reg cell (NLT-tT_reg) based on differences in their gene expression profile. (i) Heatmap of scaled mean UMI-normalized expression values of indicated genes in each cluster. (j) Number of DEGs between R26^WT and R26^TIR1(F74G) T_reg cells in pT_reg, LT-tT_reg and NLT-tT_reg cells, colored by up- or downregulation. (k) Log₂ fold changes of “Foxp3-repressed” and “Foxp3-activated” genes in pT_reg, LT-tT_reg and NLT-tT_reg cells. Each point represents a gene from the Foxp3-repressed or Foxp3-activated gene set identified in Fig. 2. Horizontal line marks the median and box marks the interquartile range. Mann-Whitney U test. (l) Scatter-plot of log₂ fold changes of DEGs between pT_reg and NLT-tT_reg cell clusters (left-top), pT_reg and LT-tT_reg cell clusters (left-bottom) and LT-tT_reg and NLT-tT_reg cell clusters (right-top). Points are colored by their direction and populations in which they are altered. Correlations were calculated with Pearson correlation over all DEGs.

Extended Data Fig. 9. Foxp3 is preferentially required during T_reg cell development.

(a) Pearson correlation between Foxp3 degradation-induced and Foxp3-dependent DEGs in thymic, resting, and activated T_reg cells, limited to Foxp3-bound genes. (b) Meta-cell analysis of resting and activated T_reg scRNA-seq data from secondary lymphoid organs following Foxp3 degradation, correlating Foxp3 expression levels with “TIR1-up” and “TIR1-down” gene signatures identified in Fig. 3. Dashed red line depicts line of best fit. Correlations and corresponding p-values were calculated with Pearson correlation over all genes. (c) Cytokine production by CD4⁺ T cells from neonatal Foxp3^AID mice after 14 days of Foxp3 degradation, in comparison to age-matched Foxp3^WT and Foxp3^GFPKO mice. (d) Neutrophil expansion in adult and neonatal Foxp3^AID mice following 14 days of Foxp3 degradation. Age-matched Foxp3^WT and Foxp3^GFPKO mice serve as controls for neonatal Foxp3^AID mice. (e) Representative H&E staining and histology scores of skin inflammation in neonatal Foxp3^AID mice after 14 days of Foxp3 degradation, in comparison to age-matched Foxp3^WT and Foxp3^GFPKO mice. (c–e) Scatter plots represent mean ± SEM. Each point represents a unique mouse. Data are pooled from two independent experiments. One-way ANOVA. (f) Scatter plots of DESeq2 normalized counts of select DEGs from the neonatal Foxp3 degradation RNA-seq experiment (Fig. 5m). Each point represents a unique mouse. Two-sided t-test. (g-h) Representative plots (g) and combined data (h) showing T_reg proliferation in 8-week-old adult versus 7-day-old neonatal mice measured by Ki67 positivity. Scatter-plot represents mean ± SEM. Data are combined from two independent experiments. Two-tailed t-test. Source data

Extended Data Fig. 10. Foxp3 degradation leads to tumor shrinkage with minimal adverse effects.

(a–b) Representative flow cytometry plots (a) and combined data (b) of Ki67 expression and EdU incorporation in T_reg cells from the tumor and tumor-draining lymph node (dLN). Scatter-plot shows mean ± SEM. Each point represents a unique mouse. Data are representative of two independent experiments. Two-tailed t-test. (c) Representative flow cytometry plots (left) and combined data (right) of IFN-γ production by tumor-infiltrating NK cells. Scatter-plot shows mean ± SEM. Each point represents a unique mouse. Data are pooled from two independent experiments. Two-tailed t-test. (d) Representative flow cytometry plots (left) and quantification (right) of IFN-γ production by tumor-infiltrating ZsGreen⁻ CD4 T cells. Scatter-plot shows mean ± SEM. Each point represents a unique mouse. Data are pooled from two independent experiments. Two-tailed t-test. (e) Volcano plts showing the number of genes up- and downregulated in Foxp3^AIDR26^TIR1(F74G) tumor T_reg cells within each clustered defined in Fig. 7(l). Source data