Bcl11b dose restrains the CD8 T cell memory:naïve ratio

a –b. Flow cytometric characterisation of Bcl11b haploinsufficient and wildtype control splenic CD8 T cells. a. Representative flow cytometry plots of CD8 T cell frequencies (left) and quantification of total splenic cellularity and CD8 T cell frequencies (right). b. Representative flow cytometry plots of naïve (CD62L+ CD44-low, TN), central memory (CD62L+CD44+, TCM) and activated/effector memory (CD62L- CD44+, TEM/Act.) CD8 T cells (left), quantified right. c., as for b., comparing Bcl11bΔEnh/ΔEnh enhancer mutant and control splenic CD8 T cells. d. As for b. and c., comparing Bcl11bfl/+ Cd4-creTg conditional heterozygote and Cd4-creTg control splenic CD8 T cells. e. Comparison of Bcl11b expression between co-stained Bcl11b haploinsufficient, enhancer mutant and wildtype control splenic CD8 T cells. Upper, histograms of expression from representative samples, lower, quantification. f. Left, representative histograms of Bcl11b expression by conditionally haploinsufficient and Cre-expressing control CD8 T cells of naïve, TCM and activated/TCM phenotypes, quantified right. Significance tested using Student’s two-tailed t test (a) or one-way ANOVA with Šidák’s (b-d, f) or Tukey’s (e) correction for multiple comparisons.

Corepressor recruitment by Bcl11b restrains central memory CD8 T cell compartment size

a.-e. E17 fetal thymuses from Bcl11b+/+, Bcl11bR3S/+ and Bcl11bR3S/R3S embryos analyzed by flow cytometry. a. Total thymic cellularity. b. Frequency of lineage marker positive cells. c. Representative flow cytometry plots (left) and frequency (right) of DN subsets. d. Representative flow cytometry plots (left) and frequency (right) of CD4/CD8 expressing populations. e. Representative histograms of DN2 cells (left), and quantification among DN2-DN4 cells (right) of c-Kit expression measured by flow cytometry. f. Representative flow cytometry plots (left) and quantification (right) of major CD4/CD8 expressing populations among adult Bcl11bR3S/+ and littermate control thymuses. g. Bcl11b quantification by flow cytometry among splenic CD8 T cells. h.-i. TN, TCM and TEM frequencies among splenic CD8 T cell populations, h. 6-8 week old and i. 5-8 month old mice. Significance tested using one-way ANOVA with Tukey’s (a- e) or Šidák’s (f, h, i) correction for multiple comparisons, or Student’s two-tailed t test (g).

Bcl11b haploinsufficiency expands the virtual memory compartment

a.-d. Flow cytometric characterisation of the splenic CD8 T cell compartments of Bcl11b+/- and littermate control mice. a. Left, representative plots and right, quantification, of KLRG1 and CD127 (IL-7 receptor alpha chain) expression by TN, TCM and TEM CD8+ subsets. b. Upper, representative histograms and lower, quantification, comparing TCF1 and Eomes expression by TN, TCM and TEM CD8 T cells of each genotype. c. Left, representative plots and right, quantification of IFN-γ and TNF-α expression following stimulation with PMA and ionomycin with brefeldin. d. Left, representative flow cytometry plots and right quantification of expression of virtual memory T cell identity-associated proteins CD49d and NKG2D by TCM and TEM CD8 populations. e.-g. Bulk RNA-seq of splenic CD8 T cell populations in germline and conditional Bcl11b haploinsufficiency models. e. Z-score plot of all genes significantly differentially expressed between genotypes in naïve or central memory phenotype cells, select genes labelled. f. Principal component analysis of sequenced samples using 10% most variant genes. g. Summary of geneset enrichment analysis (GSEA) of select pathways significantly enriched between genotypes in naïve and TCM phenotype samples. Significance tested using one-way ANOVA with Šidák’s correction (a-c), or Student’s two-tailed t test (d).

Bcl11b dose determines TVM compartment size before T cells enter the periphery

The TVM generating potential of wildtype and Bcl11b-haploinsufficient cells were determined in transfers of the indicated cell populations to wildtype hosts and in the postnatal ontogenic timecourse of TVM production in mice with different Bcl11b genotypes. No recipient animals were depleted of pre-existing host CD8 T cells for these experiments. a. Representative flow cytometry plots of splenic CD8 T cells compartments of mixed bone marrow chimeras (left). Upper, from Bcl11b+/- or control donors, or lower, wildtype competitors. Right, quantification of TN, TCM and TEM frequencies among test compartments (upper) vs their wildtype competitors (lower). b. Naïve CD8 T cells from Bcl11b+/- or wild type control spleens were adoptively transferred into wild type hosts, and donor cells analysed two weeks later. Left, representative flow cytometry plots of host and donor CD8 T cells. Right, quantification of recovered cells and the frequency of TN, TCM and TEM phenotypes. c. 400,000 naïve CD8 T cells from Bcl11bfl/+Rosa26CreERT2/+ or Rosa26CreERT2/+ control mice were transferred into new hosts, which were subsequently administered tamoxifen. Right, representative flow cytometry plots of host and donor CD8 T cell phenotypes two weeks post transfer. Right, quantification of recovered cells and TN, TCM and TEM frequencies. d. Splenocytes from 7 day old Bcl11b+/- and littermate control pups were analysed by flow cytometry. left, representative flow plots and, right, quantification of TN, TCM and TEM cell frequencies among CD8 T cells. e. Bone marrow chimeras with either Bcl11bfl/+Cd4-creTg or Cre-control donor cells analysed four weeks post reconstitution. Left, representative flow plots of host and donor CD8 T cell phenotypes. Right, quantification of donor cell frequency among total (host + donor) CD8 T cells, and of of TN, TCM and TEM cell frequencies among donor cells by genotype. Significance tested using one-way ANOVA with Šidák’s correction (a-e), or Student’s two-tailed t test (b, c, e).

Bcl11b dose in the most mature CD8SP thymocytes determines TVM fate

a. Bcl11b protein expression determined by flow cytometry in indicated populations, among conditionally Bcl11b heterozygous or Cre-expressing control thymuses. b. Bcl11b protein expression in late thymic developmental stages, comparing co-stained Bcl11b haploinsufficient, enhancer mutant and wildtype control splenic CD8 T cells. Upper, representative histograms, lower, quantification. c. Bcl11b conditionally haploinsufficient and Cre-expressing control CD8SP thymocytes were analysed for expression of Ly6C and Eomes among Mature 1 (‘M1’, CD69+) and M2 stages. Upper, representative flow cytometry plots, lower, quantification of marker (co-)expression. d. Ly6C+ and Ly6C- M2 stage CD8SP thymocytes from Bcl11b conditionally haploinsufficient or control thymuses were isolated by flow cytometry and adoptively transferred into lymphoreplete hosts. Donor cell phenotype in the spleen analyzed one week later. Upper, representative flow cytometry plots of donor-derived cells’ TN, TCM and TEM frequencies, quantified below. Significance tested using one-way ANOVA with Šidák’s (a, c, d) or Tukey’s correction (b).

Bcl11b dose activates distinct gene sets found in scRNA-seq and regulates mitochondrial homeostasis in mature thymocytes

a. - c. Bulk RNA-seq of Mature 2 stage (‘M2’, CD69-), mature CD8SP (TCRβ+ CD4-CD8+) Bcl11b+/- and wildtype control thymocytes. a. Biplot of mean gene expression in each genotype. Significant genes colored by genotype bias, with select labelled significant genes highlighted in green. Dashed lines indicate a 2-fold change. b. - c. Geneset enrichment analysis for indicated gene sets. d. - i. Single cell RNA-seq of mature TCRβ+ thymocytes from the CD4+CD8int to CD8SP stage comparing Bcl11bfl/+Cd4-creTg and Cre-expressing control cells. d. Feature plots of expression of indicated genes, and cell cycle score, across a UMAP projection of all included cells. e. Upper, distribution of genotypes (left) and Leiden clusters (right) across the UMAP projected cells. Lower, frequencies of each sample’s cells within each cluster. f. - i. Pseudobulk differential expression calls using gene set enrichment analysis to compare the two genotypes within each cluster. Whole clusters are colored by normalized enrichment score (N.E.S.) for the indicated gene sets, with warm colors for clusters in which the geneset is enriched in the Bcl11b haploinsufficient cells and cold colors for clusters in which the gene set is enriched in the Bcl11b+/+ control cells. Non-significant clusters are colored grey and clusters insufficiently populated for testing are colored black. f. TVM, enforced Eomes expression, and Ly6C+ naïve CD8 T cell relevant gene sets. g. Bcl11b loss gene sets. h. Genesets of DP thymocytes positively selected with transgenic TCRs with a known hierarchy (OT-I>F5>TG6) of self-reactivity. i. Mitochondrial oxidative phosphorylation gene sets.

Bcl11b haploinsufficiency alters gene expression without altering chromatin accessibility.

a. - d. ATAC-seq was performed on thymic M1, Ly6C- M2 and Ly6C+ M2 thymocytes and naïve and central memory splenocytes from Bcl11bfl/+Cd4-creTg and Cre-expressing controls. a. Replicate-merged genome browser tracks from key loci. b. Normalised read counts within called open chromatin regions found to significantly differ between populations. c. As in a. d. Principal component plot of all 20 samples.

Bcl11b haploinsufficiency phenotype does not result from altered TCR signaling

a. Flow cytometric analysis of mitochondrial content and polarization. Representative histograms of Tom20 (left) and MitoTracker DeepRed FM (right) straining in indicated late thymic populations, quantified below. b. Representative histograms of Nur77-GFP expression by indicated thymic populations, quantified right . c. - d. Flow cytometric analysis of splenic CD8 T cells from OT-I TCR transgenic Bcl11b+/- and littermate Bcl11b+/+ controls. c. Representative flow cytometry plots (left) of, upper, TN, TCM and TEM frequencies, and lower, ‘true memory’ vs TVM frequencies, quantified right. d. Representative histogram of Ly6C expression by naïve CD8 T cells of each genotype (upper), quantified below. e. Model of a proposed intrathymic commitment pathway for TVM cells. Significance tested using one-way ANOVA with Šidák’s correction (a. – c,) or Student’s two-tailed t test (d.).

a. Reanalysis of differentially expressed genes between homozygous or heterozygous Bcl11b-deleted DN3 thymocytes and Cre-expressing controls (Hosokawa et al., 2018). Scaled venn diagrams indicate number of differentially expressed genes, with select genes labelled. b. Flow cytometric analysis of fetal liver chimeras reconstituted with homozygous or heterozygous Bcl11b-null cells. Right, representative flow cytometry plots of donor-derived cells in the spleen, quantified right. Significance tested using one-way ANOVA with Šidák’s test for multiple comparisons (b).

Summary of genetic models used.

a. The developmental stage in which indicated models display altered Bcl11b expression. b. Modifications to the Bcl11b locus in each genetic model. c. Map of the Bcl11b-null mCherry reporter used here.

a. Quantification of naïve CD8 T cell CD44 expression from Figure 1b. b.-c. CD4 T cell compartments of Bcl11b+/- and littermate control spleens analysed by flow cytometry for activation status and naïve cell CD44 expression. d.-e. Flow cytometric characterisation of Bcl11bΔEnh/ΔEnh enhancer mutant and control thymic and splenic populations. d. Numbers of major CD4/CD8 expressing populations. e. Splenic cellularity and CD8 T cell frequencies. f. - i. Flow cytometric characterization of Bcl11bfl/+Cd4-creTg and Cre-expressing control tissues. f. Representative flow plots (left) and quantification (right) of major CD4/CD8 expressing thymic populations. g. Splenic cellularity and CD8 T cells numbers of each genotype. h. CD8 and CD3 expression level among splenic CD8 T cells of each genotype. i. Bcl11b protein content data from Figure 1f normalized to control cells within TN and TCM populations. Significance tested using one-way ANOVA with Šidák’s test for multiple comparisons (b-d, f) or Student’s two-tailed t test (a-c, e, g-i)

a. As in Figure 3c. Representative flow cytometry plots of IFN-γ and TNF-α expression of TN, TCM and TEM CD8 T cell subsets of from Bcl11b+/- and control spleens. b. Representative flow cytometry plots (left) and quantification (right) of CD49d+ ‘true’ memory T cells among Bcl11bfl/+Cd4-creTg and Cre-control splenic CD8T cells. c. Representative flow cytometry plots of the distribution of CD44 and Ly6C expression among TN, TCM and TEM CD8 T cell subsets of from Bcl11b+/- and control spleens, and quantification (lower) of Ly6C expression. d. As in Figure 3g. Geneset enrichment analysis of signatures of Bcl11b deleted populations from published datasets, comparing Bcl11b+/- and control TN and TCM populations. e. Naïve Bcl11b+/- and control CD8 T cells were cultured overnight in the indicated concentrations of plate-bound anti-CD3. Representative flow cytometry plots of CD69 and CD44 expression (left), and quantification of CD69, CD44, CD25 and CD5 expression (right) after culture. Naïve Bcl11b+/- and control CD8 (f.) and CD4 (g.) T cells cultured in anti-CD3 and IL-2 for three days and proliferation assessed by flow cytometry. Representative histograms of division tracing dye dilution (left), and quantification of cell numbers (right). h. Naïve Bcl11b+/- and control CD8 T cells cultured in IL-7 for a week. Left, representative histograms of dilution of division tracing dye, right, quantification of cell numbers after culture. Significance tested using one-way ANOVA with Šidák’s correction for multiple comparisons (a, c, e), two-way ANOVA with Tukey’s test for multiple comparisons (f, g), or Student’s two-tailed t test (b, h).

a. Quantification of CD44 expression by donor-origin naïve CD8 T cells in Figure 4b. b. Quantification of CD44 expression by recovered donor-origin CD8 T cells in Figure 4c. c. As in Figure 4c, for the cells recovered from pooled brachial, axial and inguinal lymph nodes. d.-j. Bone marrow chimeras generated from Bcl11fl/+Cd4-creTg and Cre-expressing donors analysed at the indicated timepoints post transfer. d. CD44 expression by recovered naïve donor-origin cells at 4 weeks. e. Quantification (left) and representative histograms (right) of Ly6C expression by donor-origin naïve CD8 T cells at 4 weeks. f.-g. As for Figure 4e, at 5 weeks post reconstitution. h. and i., as for d. and e., at 5 weeks post reconstitution. j. Representative flow cytometry plots (left) and quantification (right) of CD73 expression by mature single positive thymocytes 4 weeks post reconstitution. Significance tested using one-way ANOVA with Šidák’s correction for multiple comparisons (g, j) or Student’s two-tailed t test (a-f, h, i).

a. Representative flow cytometry plots (left) and frequencies (right) of thymic regulatory T cell lineage cells among CD4 single positive thymocytes from Bcl11b+/- and littermate control animals. b. Representative flow cytometry plots (left) and quantification (right) of Ly6C expression by recovered naïve CD8 T cells in Figure 5d. Significance tested using one-way ANOVA with Šidák’s correction for multiple comparisons

Single cell RNA-seq of thymic CD4+CD8int and mature CD8SP cells, as in Figure 6.

a. Violin plots of expression values of hypothesis-relevant and control genes, within indicated clusters or pooled clusters. Note that the small cell number in TVM-like cluster 8 penalized the statistical significance of comparisons. b. Histogram of geneset score values for genes ever found significantly (pAdj < 0.05, fold change > 2) enriched in either control or conditionally Bcl1b haploinsufficient cells in any cluster or pool of clusters by pseudobulk differential expression calling. Significance tested by DESeq2 (a).

a.-b. Pseudobulk differential expression calls comparing genotypes in each cluster were treated to geneset enrichment analysis. Clusters are colored by normalized enrichment score (N.E.S.), with non-significant clusters grey and clusters insufficiently populated for testing colored black. a. Genesets of lymphocyte responses to cytokine stimulation. b. MSigDB genesets of TCR signaling. c. Geneset score for indicated gene sets was calculated per genotype and per sample. Left: featureplots showing enrichment of indicated gene set scores for each genotype in UMAP space. Right: Plots of mean gene set score values across UMAP2 bins. d. Venn diagram indicating common genes in the first four gene sets of panel c. Significance tested using a likelihood-ratio test comparing mixed-effects models of per-sample UMAP2-bin mean scores, with sample as a random intercept; the full model included a genotype x UMAP2-bin interaction and was compared to a reduced additive genotype + UMAP2-bin model (c),

Representative genome browser tracks for all ATAC-seq replicates for indicated loci, CPM normalized.

a. T cell memory relevant loci. b. Loci containing open chromatin regions statistically significantly (pAdj < 0.05) differentially accessible between genotypes (specific OCRs indicated by black boxes).

a. Flow cytometric analysis of mitochondrial content and polarization in splenic CD8 T cells. Representative histograms of Tom20 (left) and MitoTracker DeepRed FM (right) straining among Bcl11b+/- and littermate control spleens, quantified below. b - e. Flow cytometric analysis of Nur77-GFP reporter expression by Bcl11b+/- and littermate controls. b. Representative flow cytometry plot (upper) illustrating polygon gates used for panel c (upper) and representative histograms (lower) of Nur77-GFP expression within each. c. Representative histograms (left) of Nur77-GFP expression by Bcl11b heterozygote or control Ly6C and Eomes expressing CD8SP populations, quantified right. d. Representative histograms of Nur77-GFP expression among CD49d-negative TVM and CD49d-positive ‘true’ memory cells of each genotype (left), quantified right. e. Representative histograms of Nur77-GFP expression among Ly6C-postive and negative naïve CD8 T cells of each genotype (left), quantified right. Significance tested using Student’s two-tailed t test (a), or one-way ANOVA with Tukey’s (c) or Šidák’s (d, e) correction for multiple comparisons.