An autonomous TCR signal-sensing switch influences CD4/CD8 lineage choice in mice

How multipotential cells initiate distinct gene expression programs in response to external cues to instruct cell fate choice remains a fundamental question in biology. Establishment of CD4 and CD8 T cell fates during thymocyte development is critically regulated by T cell receptor (TCR) signals, which in turn control expression of the CD4-determining transcription factor ThPOK. However, the mechanism whereby differential TCR signals are molecularly interpreted to promote or antagonize ThPOK expression, and thereby CD4 versus CD8 lineage fates remains unknown. Here we show, using reverse genetic and molecular approaches that an autonomous, position-independent TCR-sensing switch is embedded within the ThPOK locus. Further, using an in vivo mutagenesis approach, we demonstrate that differential TCR signals are interpreted during lineage commitment by relative binding of EGR, NFAT and Ebox factors to this bistable switch. Collectively our study reveals the central molecular mechanism whereby TCR signaling influences differential lineage choice. Ultimately, these findings may provide an important new tool for skewing T cell fate to treat cancer and autoimmune diseases.


Supplementary Figures.
Suppl. Figure 1. Alternate models of TCR-dependent control of ThPOK transcription. Alternate models of ThPOK transcriptional regulation in class II-restricted thymocytes (left panel). In model 1, TCR signal strength directly controls ThPOK silencer function, whereas in model 2, TCR signals control activity of other cis elements, which in turn regulate silencer activity. Model of ThPOK transcriptional regulation in class Irestricted thymocytes (right panel).
Suppl. Figure 2. Shared TF binding sites in Sil ThPOK elements, and mouse Sil CD4 : (a) TF consensus binding sites (as predicted by JASPAR algorithm [1], using default parameters) were mapped to Cd4 and ThPOK silencers for 3 indicated species (total number of predicted consensus sites is indicated in brackets for each species). b,c) Focusing among these 129 TFs that are expressed in SP thymocytes, reveals a high proportion that are differentially regulated between CD4 and CD8 thymocytes and thus may be important for lineage commitment/differentiation, including Runx3 and ThPOK. Gene expression data was downloaded from IMMGEN SKYLINE RNAseq database [2], and log2 (fold change) in CD4 versus CD8 thymocytes calculated. Fig. 3. TCRβ expression of gated DP and SDP CD8 thymocytes from wt and CD4 ThPOK.Sil/O mice, as indicated.

Suppl.
Suppl. Figure 4. CD4 ThPOK.Sil allele supports normal development of MHC class I-restricted thymocytes, but development of class II-restricted thymocytes is impaired. FACS analysis of CD4 and CD8a expression by indicated thymic or peripheral lymphocyte populations of wt or CD4 ThPOK.Sil/ThPOK.Sil mice back crossed a) AND TCR transgene on the H-2b/k background. b) MHCII -/background. Plots are a combination of 3 replicates per strain, and show % of indicated populations among total thymocytes, gated TCRβ+ thymocytes, or peripheral blood lymphocytes (PBLs) for mice of each genotype. Data are presented as mean values +/-SEM. A P value < 0.05 was considered significant. Statistical significance was determined by one-way Anova with post-hoc Tukey HSD, and indicated by asterisks (* p < 0.01; ** p < 0.005; *** p < 0.001). Statistical significance was calculated for each indicated mutant line relative to AND+ RAG-/-H- Suppl. Figure 5. Sil ThPOK elements responds to TCR signal: a) FACS analysis of CD4, CD8a, CD69 and TCRb expression by thymocytes of CD4 ThPOKsil/ThPOKsil MHCII -/-mice, either injected in vivo with anti-TCR antibody or untreated, as indicated. Note increase in CD69+ cells indicative of recent TCR stimulation. All FACS results are representative of at least 3 experiments. b) RT-PCR analysis showing relative expression of Cd4 mRNA in indicated sorted thymocyte subsets of CD4 ThPOKsil/ThPOKsil MHCII -/mice, either injected in vivo with anti-TCR antibody or untreated, and gated as in CD8 SP. Data were analyzed by applying unpaired two-tailed Student's t test, and one-way analysis of variance (ANOVA) with Bonferroni correction. A P value of less than 0.05 was considered significant. *P < 0.05, **P < 0.01, ***P < 0.001. All statistical analysis performed for sequencing data is mentioned elsewhere in the methods section. Data are presented as mean values +/-SEM. A P value < 0.05 was considered significant. N = 4 biological replicates. Statistical significance was calculated for each indicated stimulated cell population relative to equivalent unstimulated cell population.
Suppl. Fig. 7. Alignments of mammalian Sil CD4 elements, and comparison of predicted TF binding sites between Sil CD4 and Sil ThPOK elements. a) Alignments of Cd4 silencers from indicated mammalian species. Residues identical for all species are colored, according to length of homologous stretch, i.e. 1bp (green), 2bp (yellow), 3-4bp (purple), >5bp (red). Note that there is very limited homology between Cd4 silencers from different species. Comparison was carried out using Clustal Omega alignment program [3]. b) Venn diagram showing TF sites conserved in position and orientation between species for Cd4 silencers of different species. TF consensus binding sites (as predicted by JASPAR algorithm) were mapped to Cd4 silencers for 3 indicated species (total number of predicted consensus sites is indicated in brackets for each species). Using interspecies alignment, each consensus site was classified as conserved or nonconserved (in position and orientation), relative to the other species. 106 TF sites were found to be conserved in position/orientation across all 3 species for the Sil CD4 element.
Suppl. Fig. 8: Heat map showing number of binding sites for TFs predicted to recognize the Sil ThPOK and/or Sil CD4 . Number of sites for each TF is indicated by different colors: 0 = white; 1-2 = yellow; 3-4 = orange; 5-8 = Pink; 9-16 = red; >17 = dark maroon. TF sites are filtered for recognition by TFs that are expressed in T cells (according to IMMGEN Skyline RNAseq database), and for sites that are conserved between human and mouse homologs of either Sil ThPOK and Sil CD4 (or both). After applying these filters, consensus sites for 172 different TFs that remained were arranged according to relative abundance in ThPOK versus Cd4 silencer (ratio of # sites per silencer). a) Top 46 selectively enriched sites in Sil ThPOK . Note enrichment of NfkB, NFAT, Egr and E box sites (Tcf3 = E2A, Tcf4 = E2-2, Tcf12= HEB), representing putative TCR regulatory signature. b) Top 42 selectively enriched sites in Sil CD4 . (c) Positions of TF consensus motifs enriched for each silencer were mapped onto the indicated full-length mouse and human silencers. Height of the peak at each base pair position corresponds to the number of overlapping TF consensus motifs at that position. Note that factors with related consensus sites were combined into groups of same color for easier visualization. Color codes for left and right sides of the diagram are different.