Fig 2: Overexpression of different NR4A proteins in single Nr4a-deficient T cells induces distinct exhaustion-associated phenotypes in vitro.(A) Schematic of in vitro CD8+ T cell transduction and stimulation timeline. Nr4a1fl/fl, Nr4a2fl/fl, or Nr4a3−/− cells were transduced to express Cre (MSCV-Cre-IRES-NGFR/GFP) and NR4A (MCSV-HA-Nr4a1- or Flag-Nr4a2- or V5-Nr4a3-IRES-NGFR/Thy1.1); Nr4a3−/− mice did not require Cre delivery to deplete endogenous NR4A3. Representative histograms of (B) TNF or (C) IFNg expression after 4-hour PMA/ionomycin stimulation in cells gated on Cre+ and either empty vector or NR4AOE. Quantification of cytokine expression shown to the right of each histogram (n=3). Representative histograms of PD-1 (left) or TIM3 (middle) expression from unstimulated cells gated on Cre+ and expressing (D) NR4A1OE (E) NR4A2, or (F) NR4A3. Quantification of surface receptor expression shown to the right of each histogram (PD-1, n=7; TIM3 n=3). Right, contour plots of PD-1 and TIM3 expression. (G) Illustration showing in vitro CD8+ T cells from Nr4aTKO mice (Nr4a1fl/fl, Nr4a2fl/fl, or Nr4a3−/−) transduced to express Cre and NR4A or an empty vector. (H) Representative histograms of PD-1 expression in unstimulated cells (Nr4aTKO and NR4A1/2/3OE) with quantification. Right, PD-1 expression across quartile bins compared to Nr4aTKO. (I) Representative histograms and quantification of TIM3 expression. Right, TIM3 expression across quartile bins compared to empty vector. Data is representative of at least two independent experiments. Statistical comparisons by paired T-test, one-way ANOVA with Tukey’s correction, or two-way ANOVA with Šídák’s correction: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. PMA, Phorbol 12-myristate 13-acetate; MFI, Mean Fluorescence Intensity.

Fig 3: Transcriptomic analysis reveals NR4A1 promotes progenitor-like exhaustion while NR4A2 and NR4A3 induce terminal exhaustion signatures.(A) Representative contour plots of PD-1 and TIM3 co-expression. (B) Frequency of PD-1hiTIM3+ population. (C) K-means clustering (K=10) of z-score normalized (row) RNAseq of Nr4aTKO CD8+ T cells transduced to express indicated NR4A and sorted by expression of PD-1 and TIM3; gene lists merged from differentially expressed genes as determined in all possible pairwise comparisons (adjusted p-value < 0.01 and log2 fold change ≥ 1 or ≤ −1). (D) Distribution of differentially expressed genes from pairwise comparisons within PD1hiTIM3+ groups across corresponding heatmap clusters from (C). (E) GSEA displaying enrichment between NR4As sorted on PD-1hiTIM3+ (i.e., NR4A2 vs NR4A1) for signatures of indicated in vivo T cell populations. NES, normalized enrichment score; MPEC, memory precursor effector cell; SLEC, short-lived effector cell; TRM, tissue resident memory; TCIRC, circulating T cell. Datasets from GSE122713, GSE84105, GSE8678, GSE107395, GSE123739, and GSE266287. n=3 biological replicates. (F) Enrichment of leading-edge genes in murine transgenic TCR T cells that were adoptively transferred to tumor-bearing mice and harvested at the indicated timepoints from tumors; leading edge genes were normalized by z-score across days. Line is mean +/− SEM; GSE89309. State 1 dysfunction is plastic, state 2 is fixed. OT-1 CD8+ T cells, transduced with an empty vector to express GFP were adoptively transferred to CD45.1 mice 12 days after B16-OVA tumor inoculation. Pre-transfer (Pre) T cells, or TILs isolated 3 (D3), or 8 (D8) days later were subjected to single-cell RNA-sequencing (G) normalized (Log10(CPM+1)) counts of Nr4a1, Nr4a2, and Nr4a3 in total cells (Pre) or cells filtered for 20k-50k total reads (D3, D8) and plotted as a violin; frequency at bottom represents the percent of cells with 0 detectable reads of each Nr4a; gray area marks the lowest possible expression value from a cell with 50k or 20k total read counts; blue arrows indicate pseudobulk calculation. (H) normalized psuedobulk (Log10(CPM+1)) counts of Nr4a1, Nr4a2, and Nr4a3 in CD8+ TILs based on their expression of Tcf7. (I) Experimental design. 1.5×106 OT-1 CD8+ T cells, modified by delivering RNPs targeting either the Rosa26 locus (Ctrl), Nr4a1, Nr4a2, or Nr4a3, were adoptively transferred into CD45.1 mice 12 days after B16-OVA tumor inoculation. TILs were isolated 5 days later (day 17) and assessed by flow cytometry. (J) Representative histograms of Ki-67 expression in OT-1 TILs; frequency quantification (right). (K) Representative histograms of TCF1 expression in PD-1hiTCF1+ OT-1 TILs; number is median fluorescence intensity (MFI), MFI quantification (right). (L) Quantification of PD-1hiTOX+ frequency in OT-1 TILs. Data in J-L are representative of at least two independent experiments; gCtrl (n=6), gNr4a1 (n=9), gNr4a2 (n=6), gNr4a3 (n=8). Statistical comparisons by one-way ANOVA with Tukey’s correction for multiple comparisons (B, J-L): *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Fig 4: Targeted degradation of endogenous NR4A1, NR4A2, and NR4A3 differentially modulates cytokine production and exhaustion markers in human CD8+ T cells.(A) gRNAs were designed such that the cut site would be within a 10 bp proximity to the transcription start site of each NR4A locus. CD8+ T cells were isolated from healthy donor blood and gRNAs screened for indel frequency determined by (B) Inference of CRISPR Edits (ICE) analysis of each gRNA. (C) Schematic of HDR template delivered by AAV to achieve unique fluorescent reporter (FP) and in-frame knock-in of V5-FKBP12F36V to each NR4A locus. (D) Experimental scheme to edit endogenous NR4As in human CD8+ T cells. (E) NR4A gene expression (TPM) in CD4 or CD8 T cells from healthy donors in the DICE Database; T cells purified based on CD3+ CD45RA+ CD127+ CCR7+ and CD4+ or CD8+, activation performed with anti-CD3/anti-CD28 beads for 4 hours. (F) Representative histograms of fluorescent protein expression in T cells receiving AAV-delivered HDR templates, measured by flow cytometry 4 hours after ionomycin activation. (G) CD8+ human T cells from two donors were electroporated with gRNAs targeting NR4A2 locus or a control locus (AAVS1) and subsequently delivered HDRT to insert a mCherry reporter and V5-FKBP12F36V. T cells were purified by mCherry+ reporter expression and genomic DNA isolated. PCR amplicons of NR4A2 target locus were run on a 1% agarose gel to estimate bi-allelic integration frequency based on band density (ImageJ); quantification (right). All flow cytometry data gated on live CD8+ events. Experiments performed in at least two independent donors.

Fig 5: Acute depletion of individual NR4A proteins diminishes several features of T cell exhaustion.FKBP12F36V-V5-NR4A3 protein abundance in wildtype CD8+ T cells treated with (A) increasing concentration of dTAG-13 or (B) 500nM dTAG-13 for the indicated time. (C) Experimental scheme for reconstitution of cells lacking a single NR4A protein with the corresponding FKBP12F36V-NR4A fusion proteins. CD8+ T cells were isolated from wildtype (C57BL/6), Nr4a1fl/fl, Nr4a2fl/fl, or Nr4a3−/− mice and retrovirally transduced with Cre and an empty vector or an expression plasmid for FKBP12F36V-NR4A fusion proteins; KO = Nr4a-deficient; FP = FKBP12F36V-NR4A; TX = FKBP12F36V-NR4A-expressing cells treated with dTAG-13. (D) MFI of single Nr4a-deficient CD8+ T cells transduced to express the indicated NR4A protein, treated (filled symbols) or untreated (open symbols) with dTAG-13, then stimulated with PMA and ionomycin for 4 hr. NR4A1, orange; NR4A2, green; NR4A3, purple. (E) Experiment graphic for reexpression of FKBP12F36V-NR4A fusion proteins in Nr4aTKO cells lacking all three NR4A proteins. CD8+ T cells were isolated from Nr4a1fl/fl Nr4a2fl/fl Nr4a3−/− (Nr4aTKO) mice and retrovirally transduced with Cre and either an empty vector or an expression plasmid for FKBP12F36V-NR4A fusion proteins, then treated with vehicle or dTAG-13 for 48 hours. (F) Reporter levels for each NR4A fusion protein expression plasmid were binned into quartiles (~25% each) and TNF/IFNg MFI was calculated within each bin; representative histograms are shown. Reporter plasmids encoded FKBP12F36V-HA-NR4A1-IRES-Thy1.1 for NR4A1, FKBP12F36V-Flag-NR4A2-IRES-Thy1.1 for NR4A2, FKBP12F36V-V5-NR4A3-IRES-Thy1.1 for NR4A3. (G) Frequencies of TNF+ IFNg+ CD8+ Nr4aTKO T cells transduced to express NR4A fusion proteins and treated or untreated with dTAG-13 for 48 hours, then stimulated with PMA and ionomycin for 4 hr. (H) Baseline frequencies of PD-1hiTIM3+ CD8+ cells transduced to express NR4A fusion proteins and untreated or treated with dTAG-13 for 48 hours. (I) Representative histograms of TIM3 expression. (J) Quantification of the TIM3+ T cells. Data are representative of at least two experiments. Statistical significance determined by paired t-test, one-way ANOVA with Tukey’s correction for multiple comparisons, two-way ANOVA with Šídák’s correction for multiple comparisons, or paired t-test; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.